DOCUMENT RESUME ED 194 060 IR 008 912 AUTHOR Harris, Diana, Ed.: Ccllison, Beth, Ed. TITLE Proceedings of NECC/2 National Educational Computing Conference 1980 (Norfolk, Virginia, June 23-25, 1980). INSTITUTION Iowa Univ., Iowa City. Computer Center. PEPORT NO IsBN-0-937114-00-6 PUB DATE Jun 80 NOTE 306p. EDRS PRICE MFOI Plus Postage. PC Not Available from EDRS. DESCRIPTORS *Computer Assisted Instruction: Computer Assisted Testing: Computer Oriented Programs: Ccnputer Programs: *Computers: *Computer Science: Elementary Secondary Education: Games: Higher Education: Humanistic Education; Learning Laboratcries: Material Development: Mathematics Instruction: *Microcomputers: Science Education IDENTIFIERS *Computer Literacy ABSTRACT This proceedings, which includes 52 papers and abstracts of 13 invited and nine tutorial sessions, provides an overview cf the current status of computer usage in education and offers substantive forecasts for academic computing. Papers are presented under the following headings: Business -- Economics, Tools and Techniques for Instruction, Computerd in Humanistic Studies, Computer Literacy, Science and Engineering, structured Programming, ACM Elementary and Secondary Schools Subcommittee, Computer Science Education, Integrating Computing into K-12 Curriculum, Mathematics, Testing-Placement, Pre - College' Instructional Materials, Minority Institutions--ECM/, Computer Laboratories in Education, Computer Gt.mes in Instruction, and Computing Curricula. Abstracts are provided for 13 invited sessions dealing with such topics as microcomputers in education: research in microcomputer uses: personal computing: educational ccmputing: past, present, and future: the Open University: CAUSE program and projects: funding academic computing programs: computer-based resource sharing: computers and instruction: improving utilization cf 2-year college computer centers: teaching computer ethics: data sets available from the federal govenment: and MIS education, as well as nine tutorial sessions designed to provide attendees with the opportunity to expand their appreciation of and involvement in educational computing. (CHC) *********************************************************************** * Reproductions supplied by EDRS are the best that can be made * * ircm the original document. * ***********************************************************************
Transcript
DOCUMENT RESUME
ED 194 060 IR 008 912
AUTHOR Harris, Diana, Ed.: Ccllison, Beth, Ed.TITLE Proceedings of NECC/2 National Educational Computing
Conference 1980 (Norfolk, Virginia, June 23-25,1980).
INSTITUTION Iowa Univ., Iowa City. Computer Center.PEPORT NO IsBN-0-937114-00-6PUB DATE Jun 80NOTE 306p.
EDRS PRICE MFOI Plus Postage. PC Not Available from EDRS.DESCRIPTORS *Computer Assisted Instruction: Computer Assisted
ABSTRACTThis proceedings, which includes 52 papers and
abstracts of 13 invited and nine tutorial sessions, provides anoverview cf the current status of computer usage in education andoffers substantive forecasts for academic computing. Papers arepresented under the following headings: Business -- Economics, Toolsand Techniques for Instruction, Computerd in Humanistic Studies,Computer Literacy, Science and Engineering, structured Programming,ACM Elementary and Secondary Schools Subcommittee, Computer ScienceEducation, Integrating Computing into K-12 Curriculum, Mathematics,Testing-Placement, Pre - College' Instructional Materials, MinorityInstitutions--ECM/, Computer Laboratories in Education, ComputerGt.mes in Instruction, and Computing Curricula. Abstracts are providedfor 13 invited sessions dealing with such topics as microcomputers ineducation: research in microcomputer uses: personal computing:educational ccmputing: past, present, and future: the OpenUniversity: CAUSE program and projects: funding academic computingprograms: computer-based resource sharing: computers and instruction:improving utilization cf 2-year college computer centers: teachingcomputer ethics: data sets available from the federal govenment: andMIS education, as well as nine tutorial sessions designed to provideattendees with the opportunity to expand their appreciation of andinvolvement in educational computing. (CHC)
************************************************************************ Reproductions supplied by EDRS are the best that can be made *
* ircm the original document. ************************************************************************
U S DEPARTMENT OF NEACTN.EDUCATIONE *waitsNATIONACINSTITUTE OF
EDUCATION
THIS DOCUMENT NAS owl REPRO-ovceo exam., AS RECEIVED FROMTHE PERSON OR ORGANIZATION ORIGIN.MING IT POINTS of view OR OPINIONSSTATED 00 NOT NECESSARILY 'TERREset:TORSION. NATIONAL INSTITUT ECFEDUCATION POSIT/04 OR POLICY
,Proceedings of NECC/2National Educational Computing Conference 1980
Edited byDiana HarrisBeth Collison
Hosted byChristopher Newport CollegeNewport News, Virginia
Held etHoliday Inn/ScopeNorfolk, Virginia
23, 24, 25 June 1980
..
"PERMISSION TO REPRODUCE THISMATERIAL IN MICROFICHE ONLYHAS BEEN'ORANTED BY
Ted Sjoerdeme
TO THE EDUCATIONAL RESOURCESINFORMATION CENTER (ERIC).'
CopyrightInternational Standard Book Humbert 0-937114-00-d
Published by19Sgs SZCC
The University of IowaWag Computing Center
forIowa City. Iowa 52242
guns 1980NAtIonal aluestIonal Computing Conference II
Cover dealgn by Madeline WIn4auer
1
NECC 1980 Steering Committee
Alfred BolaUniversity of California-Irvine
Bobby BrownUniversity of Iowa
William S. DornUniversity of Denver
Karen DuncanGerald L. Engel
Christopher Newport Co.iegeNorman Gibbs
College of William and MaryJohn W. Hamblen
University of Missouri-RollaPaul Hazen
Johns Hopkins UniversityHarry Hodges
Michigan State UniversityLawrence A. John
University of DaytonSister Mary Kenneth Keller
Clark* CollegeWilliam E. Knabe
University of IowaDavid R. Kniefel
New Jersey Computer NetworkDoris K. Lidtke
Towson State UniversityElisabeth Little
EDUCOM/EDUNETSister Patricia Marshall
Xavier University of LouisianaRichard W. Pogue
Medical College of GeorgiaJames Poirot
North Texas State UniversityNancy Roberts
Lesley CollegeTheodore J. OPerdeme
University of IowaDavid L. Stonehill
University of Rochester'David B. Thomas
University of Iowa
General Chairman: Gerald EngelProgram Co- chairmen: Richard H. Meting,
Dori:: K. LidtkeLocal Arrangements Committee: Robert Mathis,
Michael StamenPublications Committee: Diana Harris. Theodore
Sjoerdema, Madeline Windauer
The National Educational Computing Conference wishes to thank the followingpeople for their contribution of effort, time, and knowledge as referees forthe papers submitted for presentation.
Robert Aiken, University of TennesseeWilliam Atchison, University of MarylandEmile Attala, Polytechnic State UniversityR. P. Banaugh, University of MontanaBruce Barnes. National Science FoundationJoyce Damao, Southern UniversityAlfred Bork, University of CaliforniaO. R. Boynton. Indiana UniversityMary Jane Brannon, Huntingdon CollegeHans &trey, Tennessee Tech. UniversityTom Carroll, Michigan State UniversityN. John Castellano Indiana UniversitySylvia Charp, School District of PhiladelphiaRonald Collins, Eastern Michigan UniversityFrank W. Connolly, Montgomery CollegeMichael J. D'Amore, New Jersey Educational
Computing NetworkCharles Davidson, University of WisconsinHerbert Dershem,\Hope CollegeDavid Evans, VIMSSelby Evans, Texas Christian UniversityStefan Yeyoch, College of William and MaryF. T. Fink, Michigan State UniversityFred Gage, Texas Christian UniversityNorman Gibbs, College of William and MarySheldon'P. Gordon, Suffolk County Community
CollegeKeith Hall, Ohio State UniversityJohn W. Hamblen. University of MissouriMarry Hodges, Michigan State UniversityDarlene Heinrich, Florida State UniversityJews A. Higgins. Digital Equipment Corp.A. AA J. Hoffman. Texas Christian UniversityLawrence Jehn, University of DaytonVincent H. Jones. Baton Roque, LouisianaBasheer Khumawala, University of North CarolinaJoyce C. Little. Community College of BaltimoreRichard W. Lott, Bentley College
David Maharry, Wabash CollegeSister Patricia Marshall, Xavier UniversityDonald H. McClain, University of IowaDonald E. McLaughlin. Augustana CollegeEdward D. Meyers Jr.. Center for the Study of
Health DevelopmentPercy L. Milligan, Southern UniversityRobert D. Montgomery. North Carolina Central
UniversityCatherine Morgan, Kensington. MarylandMike Moshell, University of TennesseeRobert Noonan, College of William and MaryBruce E. Norcron. SUETY at BinghamtonLinda Petty, Hampton InstituteCharles Pfleteger, University of TennesseeRichard Pogue, Medical College of GeorgiaW. W. Porterfield, Hampden-Sydney CollegeJames Powell, Burroughs-Wellcome Co.C. A. Quarles, Texas Christian UniversityJoe Rabin. Queen College of BURYOttis Richard, University of DenverDavid Rine, Western Illinois UniversityPeter Riese. Control Data Corp.Leroy Roquemore. Southern UniversityR. C. Rosenberg. Michigan State UniversityTed Sjoerdsma, University of IowaDavid Smith, Duke UniversityPhilip F. Spelt. Wabash CollegeElliot A. Tanis. Hope CollegeRobert Tannenbaum, Ancram, New YorkDavid Thomas. University of IowaRobert Thompson, University of DaytonJohn Van Iwaarden. Hops CollegeRita Wagstaff, Temple UniversityT. C. Willoughby, Ball State UniversityGary Wittlich, Indiana UniversityAllen Disbur, SUNK at Binghamton
And special thanks to the authors whose cooperation kept this publication onschedule. .,.-Diana Harris
5
Foreword
The Second National Educational Comput-ing Conference (NECC/2) builds on thesuccess of last year's conference at theUniversity of Iowa. In organizing theprogram for NECC /2 we have attempted toimplement many of the suggestions we havereceived during the year. Of specialsignificance are the tutorial sessionswhich are designed to provide attendeeswith the opportunity to expand theirappreciation of and involvement ineducational computing.NECC/2 is a broadly based conference'
bringing together, in the common inter-ests of computers in education, a greatnumber of individuals with diverse back-grounds and a great number of cooperatingsocieties. It is hoped that in thisforum the diversity can be focused toimprove our common interest.This volume presents the papers pre-
sented at the conference and summaries ofmost of the special sessions. The coordi-nation of the contributed papers was inthe able hands of Richard H. /mating ofthe University of Maryland, and the coordi-nation of the special sessions was in the
CV
equally able hands of Doris K. Lidtke ofTowson State University. We give them oursincerest thanks for the many hours spenton these tasks. We also thank the authorsfor submitting their works and working withus to maintain our schedule.Thanks are also due to the entire NECC/2
Steering Committee for their excellentguidance in preparing for the Conference.Special acknowledgements go to Ted
Sjoerdsma of the University of Iowa whocoordinated publicity and Diana Barrisalso of the University of Iowa who editedthese proceedings.Finally we thank all those individuals
who came to NECC/2 and helped insure thatthe concept of this series of conferencesis a success.
Gerald L. EngelChairman, NECC/2Christopher Newport CollegeNewport News, Virginia 23606
Table of Contents
TUTORIALS1 CAI-An Introduction
Michael Aronson and Harold Rabmlow2 How to Choose a Microcomputer for Educational Use
Kevin Hausmann3 Turning Students on to the Computer:
The Introductory CourseGary Shelly
4 Instructional DesignGary Stokes
INVITED SESSION5 . Microcomputers in Education
Chaired by Murali R. Varanasi
TUTORIAL6 Program Development Techniques
A. J. Turner
BUSINESS/ECONOMICS7 Computer Science and MIS College Students:
An Investigation of Career-Related CharacteristicsEleanor W. Jordan
12 Force- Feeding SPS8 in Market Research and Analysisat Hampton Institute
Howard F. Wehrle III16 Short-Run Forecasting of the U.S. Economy
Minims R. Bowman
TOOLS MD TECHNIQUES FOR INSTRUCTION19 Hypertext: A General-Purpose Educational Computer Tool
Darrell L. Ward and Steve Bush25 A Dynamic Process in Teaching Techniques
Jamil E. Effarah31 Considerations and Guidelines for Developing Basic Skills
Curriculum for Use with Microcomputer TechnologyRobert M. Caldwell
INVITED SESSION37 'Research in Microcomputer Uses in Education
Chaired by David Xdiefel
CCMPOTERS IN HUMANISTIC STUDIES39 Creativity through the Microcomputer
George M. Bass, Jr.42 Giving Advice with a Computer
James W. Carson46 ?roma Theory of Reading to Practice via the Computer
Dale M. Johnson and R. Scott Baldwin54 Non - Harmony: A Vital Element of Bar-Training in usic CAI
Joan C. Groom-Thornton and Antoinette Tracy Corbst
7vi
COMPUTER LITERACY58 A Case for Information Literacy
Bruce B. Schimming62 A Byte of BASIC.
Judith Am Hopper65 A Computer Workshop for Elementary and Secondary Teachers
Herbert L. LW:rah:to and John T. Whittle68 Microcomputers and Computer Literacy: A Case Study'
Robert J. Wilson
INVITED MISSION73 Personal Computing: An Adventure of the Mind--
Paul Hazen
INVITED SESSIONS74 Educational Computing: Past, Present, and Future
Ronald W. Collins75 The Open University
Prank Lovis and William Dorn
SCIENCE AND ENGINEERING76 Demographic Techniques in Ecology:
Computer-Enhanced-LearningA. John Gets, Jr.
81 Microcomputers as Laboratory Instruments:Two Applications in Neurobiology
Richard F. OlivaClassical Mechanics with Computer Assistance
A. Douglas Davis90 Computer-.Augmented Video Education in Electrical
Engineering at the U.S. Naval MalawiTian S. Lim, Michael W. Magee, and Richard A. Pollak
STRUCTURED PROGRAMMING96 The Use of Programming Methodology in Introductory
Computer Science CoursesElisabeth Alpert
103 FORTRAN 77: Impact on Introductory Courses inProgramming Using FORTRAN
Prank L. Friedman112 Using Model -Based Instruction to Teach PASCAL
Bodgan Csejdo119 Structured Machine-Languages An Introduction to Both
Low- and High -Level ProgrammingDavid G. Hannay
ACM ELEMENTARY AND SECONDARY SCHOOLS SUBCOMMITTEE125 ACM Elementary and Secondary Schools Subcommittee
Progress ReportDavid MOursund
130 Computing Competencies for School TeachersRobert P. Taylor, James L. Poirot, and James D. Powell
INVITED SESSION137 CAUSE Program and Projects
Chaired by Lawrence Oliver
TUTORIAL138 Databases - What Are They?
Arlan DeKock
COMPUTER SCIENCE EDUCATION139 Required Freshman Computer Education in a
Liberal Arts CollegeDavid E. Wetmore
143 Development of Communications Skills inSoftware Engineering
John A. Seidler and John G. Meinke147 Systematic Assessment of Programming Assignments
Judy M. Bishop152 Data Structures at the Associate Degree Level
Richard F. Dempsey
INTEGRATING COMPUTING INTO K-12 CURRICULUM155 The Scarsdale Project: Integrating Computing
into the K-12 CurriculmThomas Sobol and Robert P. Taylor
Tutorial Session is designed for personswho have had little or no experienceworking with computer-based instruction(CBI). This session will give basicinformation that will help them getstarted in CBI. The topics for the ses-sion include:
(1) An examination of terms related toCBI and the differences between com-puter assisted instruction, computer-managed instruction, and computertest generation
(2) Characteristics of today's educationalenvironment and the place of CBI in it
(3) Alternate CBI systems and character-istics of each
(6) Examples of successful CBI(5) Questions one should consider before
getting into CBI(6) Where to get additional information
ABSTRACT--Milypeople are realizing the tremendouspotential that microcomputers have foreducation; however, it is becomming increas-ingly difficult to stay abreast of all thevarieties of microcomputers currentlyavailable. One solution to this problemis to define the components that make upa minimal educational microcomputer systemand then consider only those systems whichmeet minimal criteria.The minimal system, as defined by
Minnesota educators, consists of thefollowing:- An input device must be a typewriter
keyboard and output a multilinemonitor or printer.
- a permanent file storage must be someform of disk storage.
- The BASIC language must be supported.- At least 12X of user memory must be
availalbe, excluding operatingsystem and language processor(s).
Since BASIC is the most often usedlanguage, an evaluation of BASIC languagefeatures and capabilities should be made.BASIC features typically consideredinclude sequential file handling, randomaccess file handling, chaining, specialfunctions, matrix operations, formattedOutput, and good graphics commands.
Once the range of systems has beennarrowed, ways to provide or acquiresupport should be considered. Thefollowing points were deemed importantin the Minnesota plan for microcomputersupport:- One specific microcomputer (chosen
through a bid process) should beavailable to educational agenciesby a state contract.
- Instructional service support forthe selected microcomputer should bedefined end increased to the levelcurrently available for timesharing.
- Continuous analysis and evaluationof hardware and software must keepup with the changing technology.
The technical evaluations and theinvitation for bid used by the MinnesotaEducational Computing Consortium areavailable from XECC Publications, 2520Broadway Drive, St. Paul, Minnesota 55113Ask for the 79 -80 Microcomputer Report.
13
TURNING STUDENTS ON TO THE COMPUTER:THE INTRODUCTORY COURSE
Gary 8. ShellyAnaheim Publishing Company
1120 East AshFullerton, CA 92631
(714) 879-7922
ABSTRACT--7Flai long as computing has been taughtin high schools, colleges, and universities,there hii been controversy concerning thefirst course in the curriculum. Some haveadvocated a high-level course dealing withalgorithms and programming, while othershave taken a computersn-society approach.With the increased enrollment in thiscourse, notonly in colleges and univer-sities but in high schools and even juniorhigh schools, guidance must be given toteach a worthwhile course which can satisfythe large number of students from alldisciplines who will be taking it.This tutorial will present a suggested
course content and methods for implement-ing this course at all levels of education.It will include justification for thesubject matter which the author views ascritical for the course and unique andeffective ways in which this subjectmatter can be taught effectively.
14
Tutorial 3
4 NECC 1980
INSTRUCTIONAL DESIGN AND COMPUTER EDUCATION
Gordon StokesComputer Science DepartmentBrigham Young University
Provo, Utah 84602(801) 374-1211 Ext. 3027
ABSTRACTmace of performance objectives,course organization, instructionalstrategies, and evaluation procedureswill be discussed in this session. Theinstructional design of an introductoryFORTRAN class will be the case historythat illustrates the instructionaldesign process in a classroom.The presentation will give the partici-
pants enough information and examples tohelp them organize their own classes.The evaluation procedures will cover bothformative and summative exams.Some recent research results on indi-
vidualized instruction in an introductoryclass will be presented, and theirApplications for instructional designwill be discussed.
15
Invited Session
. MICROCOMPUTERS IN EDUCATION
Chaired by Murali R. VaranasiDept. of Electrical Engineering
Old Dominion UniversityNorfolk, VA 23508(804) 440-3742
ABSTRACT7-the introduction of the first
microprocessor in 1972, advances in themicroprocessor field have so acceleratedthat universities have been faced with aaerious educational challenge. To takeadvantage of the cost and flexibilityoffered by the programmable LSI devices,the digital designers have to be educatedin software engineering; computer scien-tists have to learn digital systems atthe gate and subsystem level. Eventhough microprocessors, they must be dealtwith as any other computer--i.e., inter-faces have to be built, buses have to bedesigned, and programs written. There-fore the following challenges arise:
1. How are design skills to be taught?,2. What sort of courses are needed?3. What sort of laboratories are needed?4. How does one introduce a Computer
Science/Engineering student toapplication areas?
This panel will address the above ques-tions from the perspective of experiencededucation, industrial, and governmentexperts knowledgeable in computer educa-tion. Future applications of micro-processors will also be discussed.
15 -
Tutorial
PROGRAM DEVELOPMENT TECHNIQUES
A. J. TurnerClemson University
ABSTRACTComputer programs are often developed by
users without considering the effort thatwill be required for someone other than theauthor(s) of a program to understand itmodify it, or correct an error. However,several techniques are available to assistin the development of computer programsthat are easier to road, understand, debug,and modify. Since tee techniques alsofacilitate the initial implementation ofmost programs, they are valuable even if aprogram is intended to be used withoutmodification.
Techniques for three aspects of programdevelopment are considered in this tuto-rial: design, programming, and implemen-tation. Topdown design is discussed asthe basic approach to program design.Module independence, module function,information hiding, and the HIPO tech-nique are considered as modularizationcriteria and paradigms to -id in thedevelopment of the design. The two pro-gramming techniques discussed that facil-itate the development of program code arethe use of pseudo-code and stepwiserefinement. Also included are techniquesfor improving the readability of programcode, such as structured coding andprogram formatting and commenting conven-tions. The use of iterative. enhancementand module stubs are discussed as implemen-tation techniques that complement thedesigr and programming techniques andfacilitate the use of a topdown approachto program development.Emphasis is placed on the use of these
techniques by a small implementhtionteam or a single individual. Examplesin the BASIC and PASCAL programming lan-guages are included.
6
17
Business/Economics
COMPUTER SCIENCE AND MIS COLLEGE STUDENTS:
IS THE STAMPEDE FOR THE MONEY?
Eleanor W. JordanDepartment of General BusinessUniversity of Texas at Austin
Austin, Texas 78712512-471-5322
INTRODUCTIONCollege programs in computer science
and management information systems (MIS)are generally experiencing increasingenrollments in spite of declining orstable university-wide enrollments.Since the demand for DP professionalsis currently very high, enrollmentincreases in these two areas may be anindication of the practical orienta-tion of the '70s college students whosupposedly left the liberal arts anddemonstrations of the '60s to pursuesafer career paths in a declining jobmarket. if this is true, will collegesbe producing more computer graduates inthe '80s who are money-oriented ratherthan computer-oriented?
industry is unlikely to be dis-mayed at the prospect of more practicalcomputer science graduates. BruceGilchrist (5) predicts that the short-age of OP professionals will continuethrough the *80s mainly because educa-tional institutions are not producing
7
graduates with an adequate educationfor business applications. He suggeststhat computer companies spend lesstime raiding each other's DP employ-ees and more time interacting witheducational institutions in order toimprove the DP personnel situation.
Several-educators have joinedindustry recently in complaining thatcomputer science education is oftenirrelevant to industry needs (2).Resulting recommendations have oftenbeen to add particular courses orstress particular languages (2), butthere has been an increasing interestin new programs designed for thebroadly defined fields of softwareengineering (7) or information systems(4).
Before too many new programs aredeveloped, it might be valuable toconsider whether it is possible torelieve industry's frequently expressedfrustration with computer sciencegraduates by making curriculum changes.
.18
8 NECC 1980
Paul Anagnostopoulos (1), a participantin the 1979 Brown University conferenceon the difference between softwaretheory and practice, decided that themuch talked-about software crisis isdue mainly to programming personalitydefects. His suggestion that moreattention be paid to the psychologicalaspects of programming and systemsdesign is similar to Daniel Cougarand Robert Zawacki'e (3) suggestionthat DP professionals consider behav-ioral aspects of personnel responsibi-lities as well as the required technicalskills. In their investigation ofmore than 600 DP professionals (analysts,programmer/analysts, and programmers),Couger and Zawacki found that thestereotype of programmers as lonershad considerable basis in fact: DPprofessionals reported significantlyless need for interaction with othersthan all six of the other professionalgroups studied.
Are people who choose DP careers markedlydifferent from people in general? Cougerand Zawacki found that practicing DP pro-fessionals have lower social needs thanothers and concluded that this trait hadimportant implications for how DP workenvironments and job requirements should bedesigned. If computer science and MISstudents differ from other students injob-related attitudes and achievement moti-vations, it could have implications forprogram design, course design, and studentadvising in the computer science and MISareas.
STUDENT SURVEY PROJECTIn spring 1977 I was involved in the
development of an undergraduate MIS programin the College of Business Administrationat The University of Texas at Austin. Astrong computer science program alreadyexisted at the graduate and the undergrad-uate level, but odr business school committeethought that the orientation of the computer
TABLE 1. Mean Ratings of Importance of Possible ConsiderationsIn Career Choice for Student Sub-Groups
Career Considerationsin RInk Order ofImportance
Business
(N=326)
MIS
(N=21)
ComputerScience
(N2r20)
LiberalArts(1440)
NaturalSciences(N=45)
Engineering
(N=15)
1. Personal Interest* 3.4 3.3 3.0 3.6 3.5 3.4
2. Personal Abilities 3.1 3.2 3.2 3.3 3.4 3.1
3. Career Potential 3.1 3.3 3.3 2.8 2.9 3.5
4. Job Security 3.0 2.6 3.1 2.4 2.7 2.7
5. Salary** 2.7 2.4 2.9 2.2 2.3 2.5
6. Amount of TimeAllowed Outside Work 2.6 2.4 2.4 2.5 2.3 2.9
7, Flexibility 2.5 2.5 2.5 2.5 2.4 2.4
8. Variety of JobRequirements 2.4 2.3 2.3 2.4 2.2 1.9
9. Opportunity for**Community Service 1.8 1.5 1.3 2.2 2.5 1.5
10. Prestige 2.0 1.5 1.9 1.9 1.6 1.7
Items are measured on a five-point scale with Osindicating "not at all important* az.d4 indicating "extremely important."
Asterisks indicate that an analysis of variance resulted in a statistically significantdifference among sub-group means at the following levels: *--p4.05, **--14.01
19'
Business/Economics g
TABLE 2. Mean Scores for Student Sub-Groups OnWork and Pamily Orientation ?actors
Factor Business MIS ComputerScience
LiberalArts
NaturalSciences
Engineering
(A=326) (N=21) (N=20) (N=40) (N=45) (N=15)
Mastery 20.1 20.3 19.0 19.2 19.6 19.7
Work 19.7 20.7 18.4 20.4 20.5 19.9
Competi-tiveness** 13.7 12.1 11.2 12.1 12.3 12.9
PersonalUnconcern 10.3 10.3 11.2 10.9 10.7 10.1
The overall group means for these factors were very close to the Helmreich and Spencenormative data for 1455 college students.
**An analysis of variance resulted in a statistically significant difference amongsub-group means at the .01 level.
science undergraduate program was moresuitable for aspiring graduate students,systems programmers, or scientific pro-grammers than business applications sys-tems analysts. The MIS program was de-signed to provide the education for aknowledgeable business user who can ef-fectively define his information needsas well as analysts who can design therequested' software without the usualconflict between user and DP personnel.
Pall 1979 was the first semesterfor the proposed MIS program. Since itis a computer program in the businessschool I was interested in whether thestudents attracted by the program wouldhave work-related attitudes and motiva-tions similar to business students orCS students. I therefore included stu-dents in two of the required MIS classesin a Pall 1979 survey project that mea-sured attitudes from a large sample ofbusiness students at the undergraduateand masters level. A smaller sample ofliberal arts, natural sciences (includ-ing computer science), and engineeringstudents was also included for a compar-ison with the MIS group. The resultingsample included 467 undergraduates) thelarge majority were business majors inaccounting, management, or marketing.
RESULTS
Determinants of Career ChoiceMost of the students participating
in the survey reported that they haddecided on a career goal, even thoughthe sample included about as many fresh-man and sophomores as it did juniors andseniors. Approximately half of the de-
cided group had specific career goals,while the other half indicated that theyhad decided on a general area but were notsure about what their specialization mightbe.
On the survey students were asked torate the importance of ten elements intheir considerations for choosing a careerfield on a scale of 0 to 4. The ten ca-reer choice items included on the surveyare listed in Table 1 in the order of aver-age importance to the entire sample.
Personal interest in a career fieldhad the highest mean rating for all stu-
..J'aftt 'sub-groups except the computer sciencestudents and engineering students who tendedto rate the importance of career potentialmore highly than any of the other itemslisted.
The generally high ratings given -topersonal interests in career choice deliber-ations are not consistent with a popularview of '70s students as obsessed with fi-nancial security, but the rest of the itemsin the top half all fit this view. Matchinga career with personal abilities ranks sec-ond for most groups and career potential,job security, and salary follow.
To determine what reliable inter-groupdifferences might exist in deliberationsabout career choice, I performed a separateanalysis of variance for each of the tenitems. Statistically significant differ-ences among means were found for three ofthe items: personal interest, salary, andopportunity for community service. For allthree of these items the mean for computerscience students was at one end of therange of sub-group means and the
20
10 NECC 1980
means for liberal arts students andnatural science students were at the otherextreme. The mean for MIS students wascloser to the other sub-group means forall three of these career choice items.For these three considerations in makinga decision about a career, the MIS stu-dents appear to be more like other stu-dents than the computer science students.For the other considerations includedin the survey, the computer sciencestudents, like the MIS students, appearedto rate each item of similar importanceor lack of importance to the other sub-groups.
Achievement MotivationThe su=vey also included four
measures of achievement motivation.These measures were taken from RobertHe]areich and Janet Spence's (6) Workand Family Orientation Questionnairesince considerable evidence exists forthe statistical reliability and valid-ity of this instrument. The four fac-tors identified in Helmreich andSpence's analyses are designated aswork, mastery, competitiveness, andpersonal unconcern. The first twowould seem to be highly relevant torealistic motivations of aspiringDP professionals in the rapidly chang-ing computer environments work isintended to measure the desire to workhard and mastery measures desire forintellectual challenge. The itemscomposing the competitiveness factorare related to a desire to succeed incompetitive, interpersonal situations.A high score on the personal unconcernfactor indicates a relative lack ofconcern about the possible negativeinterpersonal consequences of achieve-ment. In validation studies these fac-tors have been found to be appropri-ately related to measures of scienti-fic achievement, college grades, andincome.
Computer science and MIS students,according to my results, are similarto business, liberal arts, naturalsciences, and engineering students interms of work, mastery, and personalunconcern (Table 2). In a comparisonof the six sub -group means the resultsof an analysis of variance test was notstatistically significant for any ofthese three factors. However a Statis-tically significant difference amongmeans was found for competitiveness.The highest mean factor score was forthe business students and the lowestwas for the computer science students,
the mean score on competitiveness for MISstudents was similar to the middle scoresof liberal arts and natural science stu-dents.
SUMMARY AND CONCLUSIONSThe sample of computer science and
MIS students participating in this sur-vey was quite small so any conclusionsthat can be drawn from the results mustbe tentative. Where comparisexe couldbe made to other mtudies the data didappear to be representative of collegestudents and consistent with the Cougerand Zawacki (3) study of DP professionals'personal needs. The additional implica-tions of my study may then be worth consi-dering at least for purposes of discus-sion.
For the most part computer 'scienceand MIS students appear to make careerchoices in a manner similar to otherstudents and have similar achievementmotivations. However, where statisti-cally significant differences do exist,the computer science students were foundconsistently to have an extreme posi-tion, while the MIS students consistentlyindicated choices and motivations simi-lar to the other student sub - groups.
Implications for IndustryIf the results of this study in an
academic setting are combined with theCouger and Zawacki (3) study, the impres-sion of the DP professional is that ofsomeone who has low needs for socialinteraction and little desire for inter-personal competition relative to otherprofessionals or aspiring businessexecutives. Computer science studentsare just as interested in an intellec-tual challenge ae other students, accord-ing to my study, and have a greater needfor personal growth than other profes-sionals, according to the Couger andZawacki study. But challenge and growthare apparently not defined in terms ofsocial interaction or interpersonalcompetition. If the business execu-tive understands this then it seems highlyplausible that it would be possible tocapitalize on the task orientation ofthe traditional DP professional andbenefit.from the lack of possibly des-tructive interpersonal competition.Some of the closed-door complaints thatare sometimes heard about DP shops maystem from an assumption that the sourceof the isolation policy is a know-it-all egocentricity, when in fact it islikely to be a difference in task orien-tation.
21
The results of this survey relatedto MIS students seem very hopeful for afuture reduction of the frequency ofconflicts between computer expertsand business managers. For all ana-lyses where differences in motivationsor importance placed on career choiceconsiderations were found to be statis-tically significant, the MIS studentstended to take a middle position.Hiring a combination of MIS and compu-ter science graduates may very gall resultin a DP shop that has better intro-company relationships and still deliversthe technical expertise that may re-quire an almost exclusive orientationtoward the task at hand.
Another hopeful note is that theprogrammer's big ego desoribed insome of the stories of Gerald Weinberg(8) in The Psychology of Computer?rIgraw,mEg doesn't seem o pre-va ent among either MIS or CS students.Weinberg's observations may be morebased on egocentricity than egotism.
Implications for EducatorsGordon Davis (4) has discussed
the conceptual differences in the func-tions of information analysts and systemanalysts as a basis for designingcurriculum in MIS and computer science.The results of my study provide someadditional support for the distinc-tion between broadly defined programsin information systems and the tradi-tional computer science program.The middle position occupied by theMIS students on a number of careerorientation and achievement moti-vation measures may mean that theywill generally be successful in theliaison and managerial roles they areoften placed in. Also MIS programsapparently attract a different typestudent than the computer scienceprograms. As the number of MIS.pro-grams increase the number of gradu-ates seeking DP positions may increaseat a faster pace than expected.
The high salaries and eager on-campus industry recruiters would seemto be the most obvious reasons forthe continuing increase in MIS andcomputer science enrollments, butthat's not what these students arereporting. Personal interests aremore important than salaries for allthese late '70s college students.With a continuing increase in collegegraduates and 'a variety of computereducational programs, perhaps the soft-ware race Will catch up with the hard-ware advances faster than the latest
predictions indicate.
REFERENCES
1.
Business/Economics 11
Anagnostopoulos, P. "Software Crisis:Method or Psychology?" Computerworld,October 29, 1979, pp. 23, 28.
2. Cook, J. R., Gallagher, M. C., Johnston,M. A. 'An Analytical Study of Industry'sComputer Education Needs." Interface,Vol. 1, No. 1, Winter 1979, pp:37.
3. Conger, J. D. and Eawacki, R. A. "WhatMotivates DP Professionals?" Datamation,Vol. 24, No. 9, September 197 8,pp. 116-123.
4. Davis, G. B. "Information SystemCurricula in the Business School."Interface, Vol. 1, No. 1, Winter 1979,113727:177-
5. Gilchrist, B. "Wanted: DP Professionalsor the '80's: Computerworld, January7, 1980, pp. 6 -T.
6. Helmreich, R. L. and Spence, J. T."The Work and Pamily Orientation Ques-tionnaire: An Objective Instrument toAssess Components of AchievementMotivation and Attitudes Toward Familyand Career." JSAS Catalog of SelectedDocuments in Psychology, Vol. 8, 1978,pp. 35-55.
7. Jensen, R. W. and Tonics, C. C. "Soft-ware Engineering Education - -A Construc-tive Criticism." Proceedings of theSixth Texas Conference on ComputingSystems, November 1977, pp. 5B-7 - 511-13.
Weinberg, G. The Psychology of ComputerProgramming. New York: D. Van Nostrand
iVil.
a.
12 NEM 1900
FORCE-FEEDING SPSS INMARKET RESEARCH AND ANALYSIS AT
HAMPTON INSTITUTE
Howard F. Wehrle, IIISchool of BusinessHampton Institute
Hampton, Virginia 23668(804) 727-5362
INTRODUCTIONHampton Institute (HI) is a small,
!West/stove, four-year historicallyblack college that grants graduate degreesin education, nursing, and (in conjunctionwith George Washington University) engi-neering. As of September 1979, the Schoolof Business included 15 faculty membersand 792 students (the largest enrollmentof any element of the Institute) who re-presented 36 states, the District ofColumbia, and the Virgin Islands. TheSchool of Business is actively pursuingpreliminary actions toward accreditationof the undergraduate program by theAmerican Assembly of Collegiate Schoolsof Business (AACSB), with the mid-rangegoal of an accredited Master of BusinessAdministration (MBA) program.
COMPUTER EDUCATION FOR BUSINESS ATHAMPTON INSTITUTE
The focal point for computer educationat HI is in the Division of ComputerScience, Department of Mathematics, whichoffers a major in computer science. COBOLis the principle vehicle for teaching, al-though two APL terminals are available.
Xn the School of Business, a singlecomputer course, "Computer Concepts inBusiness," has used FORTRAN. Until the1978-79 school year the course war. taughtby a full-time visiting professor, anIBM employee. Upon his departure, theauthor volunteered to teach both sectionsof the course, continuing the soiewhatrestrictive program-writing orientationusing FORTRAN, which was 4stablished byhis predeceesor.
Beginning in September 1979, however,the main thrust of the course was sub-stantially redirected from a rather narrowtechnical orientation to a broader mana-gerial orientation more appropriate topotential junior managers who are under-graduate students in business. This re-direction included a new text, O'Brien's
Computers in Business Management (7),which emphasises business applicationsand problems with substantially lessattention to detailed programming tech-niques. In addition, preliminary guidancefrom the Chairman of the Department ofManagement and Marketing had emphasizedthat FORTRAN is not necessarily the pre-ferred language. Accordingly, minorpreliminary study has addressed the possi-ble introduction of the C language (2)since software compatible with the HarrisPDP-11 (now biking installed at HI) isavailable. This guidance recognizes thefact that the goal of the School ofBusiness is not to train computer pro-grammers or potential data processingmanagers, but to train competent juniormanagers with a breed overview of busi-ness.
Hardware available to the School ofBusiness as of September 1979 includedfive keypunches, one terminal interactivewith the IBM 370/168 computer at TheCollege of William and Mary in Virginiaat Williamsburg, and an IBM System 3 atHampton Institute. The- System 3 allowsbatch processing of Statistical Packagefor the Social Sciences (SPSS) programs.
THE CHALLENGE OP BUSINESS 428-01On August 30, 1979, the author, primarily
management rather than marketing oriented,was offered the opportunity to teach, asan overload, Business 428-01, "Market.Research and Analysis." This opportunitywas eagerly grasped, after only briefcone/devotion. A conventional syllabuscomplemented the text end supported theconcept of team determination, team lead-er selection, and research topic recomeen-dation by the students. For the class of59 students (34 juniors and 25 seniors)specific guidance was given orally at thesame meeting: "This course is team-oriented. Accordingly, by the next class
23
meeting please arrange yourselves in teamsof not less than five nor more than sevenmembers, at least one of whom must havecompleted 'Computer Concepts in Business'(to ensure a basic familiarity with themechanics of computer programming, key-punch operation, and program debugging).Select your teas leader, and provide theinstructor a written list of the teammembers, designating the team leader, theindividual with basic familiarity withcomputer procedures, and a list of threetopics you propose for research. One ofthese topics, if deemed suitable, will beapproved by instructor. if none of thefirst three is considered suitable, theteam will be required to submit a secondslate of three topics." (This requirementwas never invoked, although the instructorwas asked by members of several teams, "Whydid you approve that topic? lt wasn't theone we really wanted; we just added it onto meet your requirement for three topics."The response: "lt was the only one onyour slate which had any 'meat', the otherswere lightweights which would not hey*presented any challenge to you.")
By the second class meeting, eight teamswere organized and moving out smartly; theremaining seven students wasted their firstfour class meetings fighting the problembut none dropped the course.
A sample milestone chart and work planwere distributed the first day of class;teams were advised the next requirementsafter approval of their research topicwere successive submission of an hypothesis,work plan, milestone chart, and question-naire(s) to elicit data appropriate todetermine whether the hypothesis could besupported or rejected.
RESPONSE: CONDUCT OF THE COURSEOne may wonder, at this point, what need
is there for computer support or what com-puter application exists? After topic ap-proval specified questionnaire(s) would beadministered to a random sample of not lessthan one hundred nor more than five hun-dred respondents.
Each team was issued (and acknowledgementmade by signature of s student team member)a folder containing selected extracts fromthe SPSS manual describing the optionsand statistics available for SPSS proce-dures CONDESCRUT1VE, CROSSTABS, FRE='Imam PEARSON CORR, and SCATTERGRAWand a data binder containing an unexplodedprintout of a program operating on asample of five with such data as last fournumbers of social security account number,age, height, weight, shoe site, color ofhair, and color of eyes. These data items,admittedly remote form most business appli-cation, were selected primarily both to
Business/Economics 13
demonstrate sample SPSS capabilities sadto challenge the students to define ana-logous application to their project.
For further reference, two copies ofthe 4PSS manual (5) were available: onechained to the counter in the computercenter where card decks are turned in forbatch processing; the other, the instrue-.torsa parsonel copy. (A third copy,recommended for purchase by the Institutelibrary as a reserve reference,was not yet availeble at this writing.)The instructor (along with his SPSSmanual) was also available to team membersas a resource person/advisor.
The projects moved forward. Hostclasses were split between rather cursorylecture coverage of the assigned text andsubstantial time for team work and coa-gulation with the instructor as required.Meny questions arose, most of which wereanswered in the Socratic manner or by:oscine reference to assigned coveragein the text. "Be of good cheer and readahead." Needless to say, these techniquesroused el:beta:nisi complaint from some ofthe class.
Sampling of the approved student topicsis included in the appendix co this paper.One common complaint concerning thesetopics took the general form,"What doesthis have to do with market research andanalysis?" The reply was standard: "Thethrust of this course is precisely thatof market research and analysis; whilewe are not attempting to ace the Nielsenratings, or ascertain the marketabilityod Old Tennis Shoe Bourbon, you are exer-cising the techniques that would be ap-plied in either of those two or a multi-tude of other examples."
in an otherwise orderly progression oflearning by doing was the delayed dis-tribution of the sample SPSS program bythe instructor to each team. (He had hisown problems in debugging what should havebeen a relatively simple, straight-forwardprogram, owing to his lack of familiaritywith job control language (JCL) and theevolutionary progression of SPSS fromVersion 6.0, which he had last used sixyears ago, to the current Version R.S.0)
As the semester moved forward, and themilestone for submission of completedreports approached, the instructor receivedsuccessive feelers and ultimately almostfrantic pleas to delay submission of thereport. Although time was not availableto do other than edhere to the schedule,as a small encouragement the teams wereadvised that chioce of date and order ofpresentation would be available to the
74i4w
14 NECC 1980
teams in order of their report submission:the first team would choose first, second,third, position on any of the three days;the last team would have no choice.
Several teams complained, with apparentjustification,about a severe imbalanceof work; some team members were obviouslynot pulling their share of the workload.Since it had been emphasized that the samereport grade would accrue to each teammember, concern was expressed about theapparent injustice of the workers sub-sidizing the drones. So, peer ratingswere required of all members of each team.These were compiled, reviewed, and, afterconsultation with a colleague more ex-perienced than the author in personneltesting and mensuration, appropriatelyweighted beforetkacorporation in the finalcourse grade. (In addition to the report,four graded exercises highlighting appli-cation of text material were administeredduring the term.)
Immediately after submitting each re-port, it was comprehensively and severelyreviewed-by the instructor, as if it werethe first draft of a professional report.The following day, the report was re-turned to the team leader, end after hispersual, was discussed at length with himand selected members of his teas. It wasemphasized to each team that their oralpresentation would provide an opportunityto recoup some of the cuts suffered on thewritten report.
Several teams wanted to revise and resubmit the written report, but this com-mendable response wss rejected because thelessons learned from the severely criticizedoriginal report might be dulled. Thepreparation of oral reports would be in-hibited, and such a requirement wouldplace an inordinate burden upon the studentsat a time when preparation for final ex-aminations should make major demands ontheir available time.
FINAL ASSESSMENT AND LESSONS LEARNEDAlthough the author considers that busi-
ness 428-01, "Market Reasearch and Analysis ",was successfully (albeit rather painfully,for some students) completed, he learnedsome lessons which should significantlyimprove his second and subsequent conductof the course.
First, completion of "Computer Conceptsin Business" should be a prerequisite toBusiness 428. This requirement would tendto even the student workload, since eachteem member would have some experience inkeypunching program/data cards; and wouldbe able to participate more fully in pre-paration of basic data for machine compu-tation.
Second, although Lehmann (2) provides
a comprehensive review of elementary sta-tistics as an appendix to one chapter,the statistics course required in allundergraduate sequences in business atNI should be a prerequisite to business428. This requirement would eliminatethe tendency of some students to call upunnecessary statistical routines and failto call up some useful routines.
Third, SPSS is a valuable research toolwhich all undergraduate students in busi-ness should learn to use. This limitedapplication in "Market Research and Ana-lysis" at Eamptoa Institute is the firststep in that direction: it is a requiredcourse for ell. undergraduate marketingmajors.
0
APPENDIX: APPROVED RESEARCH TOPICS FORBUSINESS 428
Team
1 What effect does the Hampton Institutepopulation have on gross sales ofstores in the Coliseum Mall?
Subject
2 Declaration of majors: Why we choosethe majors we do
3 Food additives and preservatives
4 Salt II Treaty
5 Gas rationing program in the area
6 Trade school is a good alternativeto college
7 Hampton Institute's intended growth
8 How future computerized purchasewill affect consumers and industry
9 Social Security
Business/Economics 15
REFERENCES
1. Alexander, Daniel E. and Messer, AndrewC. PORTRAN IV Pocket Handbook. New York:McGraw -Hill, 1972.
2. Kernighan, Brian V. and Ritchie, DennisM. The C Programmina Language. Engle-wood Cliffs, New Jersey: Prentice-Hall Inc, 1978.
3. Lehmann, Donald R. Market Research andAnalysis. Homewood, IL* Richard D.Irwin, Inc, 1979.
4. May, Phillip T. Programming BusinessApplicatkons 4,12 FORTRAN. BOStOkiaoughton-Mifflin Co., 1973.
5. Hie, Norman H., et.al, Statistical,Package for the Social Sciences, 2nded. New York: McGraw -Hill Book Co.,1975.
6. and Hull, C. Hadlai.SPSS Batch Release 7.0 Update Manual,March 1977. Williamsburg, VA:Computer Center, The College of Williamand Mary in Virginia, 1977.
7. O'Brian, James A. Computers in BusinessManagement: An Introduction, reviseded. Homewood, IL: Richard D. Irwin,Inc, 1979.
William R. BowmanEconomics DepartmentU.S. Naval AcademyAnnapolis, Maryland
(301-267-3156)
OBJECTIVEThe work of professional economists
most often used by government officialsand private businessmen lies within therealm of macro-economic forecasting.This art, however: is rarely acquired bystudents of economics due to the expenseof developing large-scale econometricmodels of the economy and the high levelof statistical knowledge presumed to benecessary in macro-economic modeling.
In response to this void, a researchseminar in economics at the :*.S. NavalAcademy has been designed to offer under-graduate students a chance to buildinexpensive, simplified forecastingmodels. These models are based uponeconomic theories, or extensions oftheories, learned in previous courses.They are derived with limited know-ledge of few, but powerful, statisticaltechniques rarely encountered by under-graduates. As piurt of a class exercise,each student develops one sector of amacro - economic model. These sectors thenbecome integrated into a simplified modelof Gross National Product (GNP) andinflation (as measured by the GNPimplicit price deflator).
BACKGROUNDThe art of forecasting the economy
combines intuitive judgments withcomputer-based statistical models ofaggregate spending behavior. The goal ofthese models is. quite heroic, to say theleast; one must model the decisions ofgroups of millions of individuals in emarket place. These individual decisions.when aggregated, determine the level. of
........._._production,--as well as the aggregateprice 3evel of all goods and servicesproduced.
The uses of macro-sccnomic forecastshave been widespread. They are used bygovernment policymakers when evaluating
alternative fiscal and monetary proposals,and by private individuals when evalua-ting expected economic growth rates andinflation within defined sub-sectors ofthe economy. The number of the more wellknown large -scale models is limited to ahandful, however, due to the enormouscost of building and maintaining thecomputer-based models. Some models, likethe Data Resources Inc. (DRI) model, haveas many as 200 stochastic equations andover 350 endogenods variables.
As indicated above, these cost consid-erations also limit how undergraduateschools teach this important aspect ofeconomics. If macro- economic forecasting.is taught, it is often dons by having thestudent merely manipulate previouslydeveloped models. That is, they may plugin a hypothetical value for a chosenpolicy variable (say the corporationincome tax rats) and observe the model'spredicted effects on the aggregate econ-omy. While this approach teaches stu-dents the sensitivity of economic sectorsto public policy, it tells them nothingabout the assumptions behind the model ofthe structure of the equations used inthe model.
METHODOLOGYRe ression Analysis
Tito methodology used for introducingthe student to the art of forecasting isbasically one of learning-by-doing. Thestudent is first giypp an introduction toregression analysis.( The theory behindthe statistical procedure is discussedonly with regard to the assumptions onemust make when using regression analysis.(Formulae are not proven-since the stu-dent is considered a user.)
Computers make it easy -- and fun --for the student to do regression analysisfrom the very beginning of the class.Simple time-series practice data files
27
are created and stored. These files are"accessed and analyzed with the regressionpackage Time Series Processor (TSP) thatwas originally written at Harvard andlater modified at Dartmouth to run on atime-sharing basis.
The control statements used to executethe program are quickly and easilylearned. Students May print out theirdata base in easily readable format, com-pute correlation matrices, and run pre-liminary regressions with simple controlstatements. (See Table1 of Appendix A.)More advanced techniques made necessaryby the problem of autocorrelated errorterms can also be taught to the studentwith the TSP package. Date transforma-tions such as discrete lags, logarithms,and first differences are easily com-puted. Their effects on the predictivepower and autocorrelation can be observedby comparing the model results Usingtransforms& data with tho3e of the pre-liminary regressions.
More advanced transformations,,includ-ing the Cochrane-Orcutt procedure(21 andthe Almon Polynomial Distributed Lag(MOM, are then presented and used.The latter transformations are usuallyconsidered to be beyond the realm of under-graduate students; however, thelearning-by-doing approach (made avail-able by the TSP package) makes their useboth possible and highly instructive.(See Table 2 of Appendix A for the con-trol statements used for some of thesedata transformations.)Model Design
With the statistical knowledgeacquired through doing regressions on thepractice data files, the student buildsa simple regression model (Ordinary LeastSquares (OLS) or Two Stage Least Squares(TSLS)] for a selected part of thelarger macro model. The choice ofexplanatory variables is based uponeconomic theories discussed earlier inthe research paper. The quarterly database used to analyze each student's modelis the Department of Commerce's data bankused in its Bureau of Economic Analysis(SEA) quarterly model. This data baseconsists of nearly 750 variables andtransformations of these variables thatmay be used as endogneous and exogenousvariables. Thus, it provides (at thetime of this writing) the student with anexhaustive source of-quarterly timeseries information for the period of1949:1 through 1979:3.
The student selects independent vari-ables that are deemed appropriate for thetheory he has developed previously andcreates mass storage data files for easy
.28
Business/Economics 17
access. Much effortis_then taken toderive a simple model whose explanatoryvariables have the expected sign and arestatistically significant. (See the"estimated coefficient" and "t-statistic"for the explanatory variables in Table 1of Appendix B.)
In addition, the student must work toobtain the best over-all fit of the dataand minimize the degree of autocorrela-tion. (See the "adjusted R squared" andthe "Dublin - Watson statistic" of Table 1in Appendix B.) This process usuallyinvolves strong trade-offs between thelatter two statistics and provides thestudent with a real world sense of fore-casting problems rarely encountered byeconomic majors.
After numerous structural changes ofthe model, the student selects that modelwhich most closely achieves two objec-tives. First, the model should have ahigh degree of explanatory power over thehistorical data period; second, its fore-cast errors should be minimized. Thelatter is diagnosed by dividing the his-torical data into a sample period and aforecast test period. The model is thenused to generate predicted values of thedependent variable that may be comparedwith the actual values. This process isespecially important during establishedturning points of specific cycles forvariables classified as "cyclical."(See the actual, predicted, and forecastvalues of a selected variable in Table 2of Appendix B.)The Forecast
lincestructure of the model ischosen, the student uses his best judgmentof the values of the exogenous variablesduring the forecast period. This periodis defined in the short-run, for example1979:4 through 1900 :4, to minimize fore-cast errors. These values are selectedbased upon expectations of professionaleconomists and business leaders concernedwith macro-economic forecasting asreported in recent issuccof numerousjournals and magazines.(")
The student then plugs these expectedvalues, or a range of values, into hismodel to produce "conditional ex anteforecasts." These forecasts may Si--altered if student expectations castdoubt upon the likelihood of the model'sresults, i.e., the "judgmental forecasts."It ie in this last-pilaffsthat the-Studentcomes to appreciate the limits ofcomputer-based modeling and the impor-tance of pereonal, informed judgment,common to all large-scale forecastingmodels.
18 NECC 1900
Once each student's structural model---anorforecast-have been completedi-the
forecast of GNP and inflation is done byusing a reduced fora model with the exog-enous variables and predetermined endog-enous variables used in each student'sown structural model. This step permitseach student to discuss his research pro-ject with ether class members, while pro-viding each with an awareness of theinter-relatedness of their work.
CONCLUSIONBy the end of the semester each stu-
dent has become fully immersed in ahighly technical research topic. Eachhas related previously learned economictheories to empirical model buildingusing regression analysis on time seriesdata. With the aid of the TSP softwareand the college's computer hardware, eachstudent has acquired a sense of accom-plishment rarely experienced at theundergraduate level. The blend of'eco-nomic theory and computer-based techni-cal analysis often opens the eyes ofundergraduate majors to a whole newworld of excitement in learning.
FOOTNOTES(1) Cochrane, D. and G. Orcutt,
"Application of Least-Squares Regressionsto Relationships Containing Auto-Correlated Error Terms,* Journal ofAmerican Statistical AssoiiiT1757VO1.44(1949), pp. 32-61. (See Table 3 ofAppendix B for an example of theCochrane-Orcutt adjustment factor.)
(2) Almon, S., "The Distributed LagBetween Capital Appropriations and Expen-ditures,* Econometrica, Vol. 30 (1965),pp. 178-96. (See Table 4 of Appendix Bfor an example of the Almon DistributedLag coefficient )
(3) The text used in the course is:Chisholm, R.,- and -G. Whitaker, Fore-casting Methods, Homewood: R. r-rrain,1971.
(4) Three sources of information aremost often used: (1) Federal ReserveBank publications (e.g. St. Louis'sReview, New. York's Quarterly Review, andWIW%'s New England Review) and largecommercial bank publications (e.g.Citibank's Mopthly Economic Letter, ChaseManhattan Bank's Business rti-alrf andInternational Finance, and Morgan
-----------Guarantyks-MOrin570iiranty-8urvey)1 (2)national business magazines (e.g.Business Week, Forbes, and Fortune):and (3) newspapers e.g. Nei-YEFEtinesand the Nall Street Jourarit
29
Tools and Techniques for Instruction
HYPERTEXT - A GENERAL PURPOSE EDUCATIONAL COSUVTZH TOOL
Darrell L. WardNorth Texas State University
Steve BushThe University of Texas Health Science Center at Dallas
The Hypertext computer system is described in this paper. Although its applications are varied, themajor emphasis here is the use of this system in the educational environment. The major featuresas visible to the student and instructor are developed. For the instructor, these features includethe use of Hypertext as an organization and lecture presentation tool. For the student, Hypertextprovides a model of information that mirrors a library yet permits a computer - assisted instructionapproach as information is perused.
ItiTRODUCITON
Hypertext, as an information organizing fa-cility, wee first described in 19/0 (1). Themajor emphasis of this paper will be the use ofHypertext in an educational setting. Hyper-text has been implemented at The Universityof Texas Health Science Center at Dallasand is currently being used within the MedicalComputer Science Department at that facility.The implementation details of this systemwill not be described, though it is appro-priate to describe the computing environ-ment within which Hypertext now functions.
Hypertext operates on a highly reliable,dual processor Tandem -16 minicomputer system.The Tandem-16 was deeigned to satisfy criti-cal computing functions, thus the architec-ture reflects the design criteria with ahigh degree of redundancy. The total sys-tem philosophy is to run non-stop and, infact,ersingle hardware failure does notcrash thio-Itiretem or contaminate the data in
any way (1.1.: The implementation language(Tandem mis TAL Application Language),_ as
block-structured, IIGOL -like language. TheTandem-16 is capable of supporting page-modeterminals, which permit user interactions tooccur on pages of data, much like pages of books.
With the above as background, the remain-der of the paper will alarm the way Hypertextpresents information to both the instructor
and the student. Section 2 will describe themodel of information that Hypertext presents toits users.Section 3 will introduce the instruct-
or's use of Hypertext in creating lecture mater-
ials, proaenting information in the classroom in-teracting with students, and maintaining currentmateriels in the subject areas. Section 4 willdemonstrate the student.s use of Hypertext. Thissection will describe the environment with respectto one course, although general use by the studentcould benefit the total educational p:ocess..Finally, section 5 wila summarize the ideas pre-sented and describe areas of further researchusing the Hypertext tool.
MCOEL or THE HYPERTEXT ENVIRONMENT
The basic unit of information in Hypertext isa pages the user is provided an environment ofpages and relationships among those pages. Eachpage in Hypertext is owned by an individual fromthe community of users and may be declared private,.preventiktother users from reviewing its con-tents. However, the philosophy of Hypertext is toprovide an environment of information (haring,thus typically most pages are public, accessibleto the community of users.
Each Hypertext user enters the systemthrough a page designated as the top page orentry page. When the user is identified to
19 3(j
20 NECC 1980
Hypertext and verified as a valid user, the toppage- of that user-is -immediately.presented to him.The user is then in command mode and mays
1) edit information on the current page.2) view another page via a menu selection for-
mat available on the current page.3) view another page via an 4mplicit relation-
ship between the current page and the "next" page.4) view another page by explicitly naming the
page of concern.5) establish a new page and formalise a rela-
tionship between the new page and the current page.6) establish a relationship between the cur-
rent page and some other already existing page.The above operations are by no means exhaustive,but identify some of the major functions that re-late to the educational use of Hypertext.
BALATTOVSHIPS AMONG PAGES In HYPERTEXT
Each page within the Hypertext system is giv-en a unique page number that is available for dis-play along with the informational contents of thepage. This system-assigned number is an explicitmethod of estabiishing inter-page relationships.Any page can be created or altered to point to
Urals to do
o prepare Mot vane* 3:00 p.a. mottoso trade rung:3 ocouo reid OM Articleo wart on SIN ptoposaio moth on DINS lecture
noteso call J. Smith tat*
utak yagy SS
4e0 sots
(I) Wags to do
(2) Orr.. neaten
(3) reurso kotttlelo
page 23
another page by inserting a "HYPERRIMP" to thatpage number. Any number of pages may be-ietatedto a particular page by linking them in the abovedescribed manner. The Hypertext system imple-ments the linking by providing the user a numericselection menu of pages (numbered sequentiallyfrom 1) which are accessible from the currentpage, permitting the user to surely select thepage of interest by depressing the appropriatenumeric key.
The above environment of pages is illustratedin the following example representing one user'stop page and its set of possible relationshipswith other pages. (See figure 1 fora graphicmodel of the example.) In this example the toppage provides for the selection of one of threeother pages. The other pages includes
1) a page containing "things to do" with noadditional relationships.
2) a page containing the current classestaught by the instructor (e.g. FORTRAN, COBOL,etc.) and additional access to the rosters ofeach of those classes.
Covers GGGGG tS
(1) FORUM
(2) COSOL
(3) Structured otos.
peso 27
Class ootettels
(1) FORUM
(2) COWL
(3) Structured Frog.
(4) 06.S
%($) Op. Sys. psis 66,
See Urns to totPASOAN clan: negotiate0
tOtalthS class rooter
Ed Illttfane
Mitten JonesShirley 1011
Jeff herd
Pent 96
COSOL cleat' Teeter
Soo JestsJill SmithShirley lilt
Jerry Knowles
Nab 73 4
Structuted frog.close soscss
40MOTOR
Yet
Tess 34
Figure la. An example Hypertext organisation for an instructor
31
Dope. eel.cher UNIVACnor MC were Involved/* the devevelopom ofFORTNAlktin lecke, ledffietdevplepimac effect.am le an linter,
TORUN eouctin
(1) Introdoctlat(2) txpresslomo
Asslinevnt state.(4) I/o
($) Logical Ii
(10) Nettles sublime.
pate 123
Introduction toFOAMY
text
Peie 124
1.
.11LMWhich compeer *pee:-handed the developmentof MIRAN?
(1) UNIVAC(2) ISM .....(3) DEC
Ma 00,
... continuation ofTOMO text ...
Page 03,
Figure lb. Hypertext example continued
Tools and Techniques for instruction 21
CTuelt, Mender theIrallershlp of MopDachas led the Vayfar the developmentof TAMAN
pato 111
3.C)
22 NECC 1980
page_containing_courses.already-develop--ed in Hypertext and the capability to access theindex of each course via the menu selection pro-cess. Each course (only the FORTRAN cotirse isillustrated) will consist of pages linked togetherwith the possibility of many routes through thecourse materials depending on the user's responses.
TRAVERSING PAGES IN HYPERTEXT
Hypertext offers its users a variety of op-tions for traversing its pages. This sectionwill not attempt to describe all the options butwill try to convey the atmosphere offered by Hy-pertext. There are several aids available to theusers as they peruse Hypertext, including:
1) the ability to retrace pages alreadyvisited.
2) the ability to place landmarks (either ina temporary mode or permanent mode) on pages sothat the user can directly return to a visitedpage of interest.
3) the ability to proceed forward throughpages.
The traversal of pages can be illustrated via theprevious example (the instructor's Hypertext).Consider a typical set of operations that aninstructor might wish to accomplish:
1) add "grade COBOL exam" to the list ofthings to do.
2) drop "Ed Jones" from the roster of theFORTRAN class.
3) review the FORTRAN lecture material thatwill be presented in class the next day.We will assume the user has successfully loggedonto the Hypertext system and is currentlypositioned at the top page.
To accomplish-select the "things to do page" by depressing1
-indicate the desire to edit the page (EDITcommand)
-alter the page to reflect the added item(grade COBOL exam)
-back up 1 page (now positioned at the top
page)
To accomplish 2:-select the "Classes" page by depressing 2-select the "FORTRAN class" Peg* by depress-ing I-indicate the desire to edit the page-alter the page by deleting the line contain-ing Ed Jones
-back up 1 landmark (the top page is an im?plicit landmark, thus we are now back to thetop page
To accomplish 3:-select the "Courses" page by depressing 3-select the "FORTRAN course" page by depress-
ing 1
-review the index, select- the - appropriate - --
topic of the lecture and depress that key-review the pages of the lecture one at a .
time, altering the contents of any pages asdesired and selecting the next page by de-pressing the RETURN key or function key 16(the last page of the topic normally pointsto the index page)-return to the entry page by going back 1landmark
As the instructor peruses the lecture contents,there is ample opportunity to thoroughly test outall branching situations (via the back page func-tion) and to alter any text that has changed or isincorrect. Also, the instructor can quite easilyconstruct additional materials and link those atthe appropriate point while traversing the coursematerials.
THE INSTRUCTOR ENVIRONMENT
The previous section hinted at some of thefacilities available to individual instructors.This section will explore those aspects relateddirectly to the teaching function and how Hyper-text can assist that function in most instances.Previous work has shown the delete utility advan-tages of incorporating the computer into the pre-paration and delivery of course materials (3).The current Hypertext system significantly extendsthe previous work by providing the followingadditional functions:
1) the ability to traverse course materialsin a very flexible manner(landmarks, backing up,skipping, etc.).
2) the facility for embedding, within a page,a call to an external program thus providing foran open ended system with respect to simulations,demonstrations, etc.
3) a dynamic, easy to use organization toolfor creating and maintaining course materials.
4) a framework for combining materials to bepresented in class with CAT materials to be takenby the student outside of the classroom.
The flexible approach toward visiting pageswithin the system allows the instructor to createcourse materials that suffice both for in-classpresentation and for self-paced instruction. It
permits the instructor to include additionalmaterial based on the response of the class. It
also allows the instructor to leave more detailedexplanations, examples, and problems for thestudents to discover on their own outside of class.
As as example of the above consider the pagesof information shown in figure 2, a short segmentof a FORTRAN course with branching Possibilities.In an'isi -Clips fxesentaiion; the*irstruCiOrOiiirequest, from the class, responses to the questioncontained in page 200. It may be apparent thatthe majority of the class understands the concept;thus the instructor can select to disregard fel,.Itup of the question and go directiy to the nextblock of materials to be presented, beginning at
33
Tools and Techniques 'for Instruction 23
-4311th-of-thr-folltivInn
con a valid FORteanviesesslont
(1) 23
(2) X4(4)
(3) 2-1462
...4c 200
2.5 IS A ve114 0.1messLou - It 2s a commentset page 14$ (1) forthe complete set ofrut.,*
Mole 201
aftwt this is elpv.
lovely an Assieneeotstatement
pap 20)
Logical Expressions
Vela
"Pg. 101-51 is avalid even:440e leFORTRAN - see pose 145(1) fox the templetsset of expressionroles
vase 204
Figure 2. Typical text in an instructional segment
page 204. However, there may very well be somestudents that do not understand the question andinteractions that transpired and do not stop theinstructor to request additional information.This segment of the population could turn to theCAt mode of Hypertext and review the question ontheir own.
Finally, from the instructor perspective, ahighly reliable computing system is essential.So far, the experience with Hypertext has beenquite good: there have been no system crashesduring the approximately 60 hours of in-classpresentations given to date.
STUDENT ENVIRONMENT
Again, earlier work describing the classroom use of computers and student benefitsapply in full to the Hypertext system (3).Briefly, these include:
______1)_no.requirement for extensive note taking.2) a clear, oonsistentprensentation medium.3) an outside of class CAt facility for re-
view of the in-class materials.With Hypertext, as pointed out above, the ?unt-tions available to the instructor are significant-
ly extended. The student profits as well, as
pate 202
Hypertext allows him to access information theinstructor has created in a friendly, reliable,and flexible manner. Although the student doesnot have the capability to alter this information,he may get a hard copy of any pages that arereviewed. Of course, the instructor may permitthe student the ability to create pages. Forexample, consider figure 3 as a possible studentview of Hypertext. The top page permits thestudent to organize each individual course ofinterest within Hypertext as well as his total
educational environment. The student use of such
a facility, with easy access to a reliable com-puting resource, would indeed promote a creativeand exciting environment.
SUMWARY
Hypertext, as a tool to assist the presen-tation of information both in and out of class,has been presented. An-overview Of.the Hyper-text environment, as well as its uses from boththe instructor and student vantage point, has
been described.Use of the computer in the classroom has
been slow to develop due to the lack of support-ing software and hardware. A reliable computingsystem with a creative environment such as
34
24 NECC 1980
Top_page
Things to do
Current courses
Homework ensign.
Acquaintances
Upcoming exam
Wimp to du page
Current courses page
(00e pointer co eachcourse)
Homework saalgneenes
Pate
Upcoming cues p4$4(this may alni2tilpotterer to studyoec114om tor eachexams)
Figure 3* Example Nypertoxt for a student
Hypertext should facilitate the use of computers
to assist the educational processes* Some areas*notably the reliability and interface* have beenextended from earlier work (3)1 however, thereremain sone existing areas for future research.The ability to provide a user-oriented graphicfacility embedded within a 'yet.; such as Hyper-text is an outstanding problem. Also, the abilityto provide quality video display for a largeaudience requires additional work.
=MEN=1. T. S. Nelson. "No More Teachers's Dirty Looks."
_Cteputer.Decisions, September-_1970. -
2. Tandem Corporation. Programming Manual.Cupertino, Ca. Noveeber 1977.
3. D. L. Ward. "A Computerised Lecture Prepara-tion and Delivery System." Journal of Educa-tional Technology Systems, Vol. 6(1),
1977-1978* pp.21 -32.
4. R. Chong. "On-Line Large Screen DisplaySystems for Computer Instruction." Proc.of ACM SIGCSE-SIGCSE Joint Symposium,February 1976* pp. 189-191.
5. W. Trace. "The Use of ATOPSS for PresentingElementary Operating System Concepts."SIGCSE Eulleting, 711, February 1975, pp.168-171.
6. J. Rogers. "A Computerised Classroom forInstructor's Experimentation and Training."Commuters in,Educationi-0.-Lecarme and R.Lewis (eds.), IFIP -North Holland PublishingCo.* 1975.
35
Tools and Techniques for Instruction 25
A DYNAMIC PROCESS IN TEACHING TECHNIQUESby
Mall E. Moth, Ph.D.INFORMATICS INC.311M Yuma Sheet
Conga Path, Caffein* 31304(213) 1117-9131
ABSTRACTPrograms to teach people how to use computer
products (hardware and software) often do not meet theneeds of the diverse audience of users. The needs of onegroup may vary considerably from the needs of another. Alag exists between-the hydra-like growth of the computerbeldame and software industry, and the technology forteaching customers. This paper describes a practicalInaba technique which has proved successful atInformatics Inc.
'Teaching designs are based on identified procedures.This approach Is represented by a network of relationshipsand interactions. The process starts with recognizing theaudience segments for whom the courses should bedeveloped, specifying their learning objectives, andplanning a design customized to user needs. Brainstormingfor ideas and selecting the most relevant comes next. Afterthe form of presentation is drafted, at least two dry runsare conducted before representative audiences. FeedbackIs collected, analyzed, and evaluated; results are integratedinto the evolving course. A field test at a customer site isthe next step; more feedback is collected, analyzed, andevaluated. Finally, the validated responses are integratedinto the courseware to produce a tested and usefuleducations! program
INTRODUCTIONHelping customers learn how to use software
products is an important element in the design of thoseproducts. Some customers need technical training toinstall, implement, and support a product at their facility;others need to learn to use the product for reporting orprocessing data. Meeting those need!, Is the job of
specialists In the ProductCommunications Group of Informatics Inc.
The Product Communications Group Is a part ofTechnical Product Support; it Includes technical writers,editors, graphics designers, and education developers.These communications specialists work together todocument software products and to design courses for
training customers. Their work begins early in the designof the product; software documentation and their teachingtechniques are considered integral parts of the
oduct package.
During the creation of learning activities atInformatics, the education development specialist isinvolved in a set of relationships and interactions that canbest be illustrated as a network. This network designincludes a systematic process for seeking relevantinformation to improve both course designs and teachingtechniques. Figure 1 represents this process of continuousinteractions and decision making.
IDENTIFYING THE USERSIn the computer industry, hardware or software
products are manufactured to meet the needs of a widespectrum of users. The first step in teaching techniques isto identify the audience segments of this spectrum thepeople for whom the courses should be developed. Formost practical purposes, two major audience segments,the end users and the technical support personnel, canbe defined:
1. The End User Audience. This segment of usersusually includes managers, clerical personnel,and casual programmers. The end users are, ingeneral, not a data processing - oriented groupbecause their primary interest is in theirparticular job; data processing is only a means tosome other end.
2. The Technkwl Support Audience. This segmentincludes the support personnel, mainly data baseadministrators (DBAs), system designers, systemprogrammers, -sad application programmers.
Depending on the level of complexity of the product, aplan should be developed for subdividing each audiencesegment into smaller segments with users of similarbackgrounds and job requirements.
36
26 NECC 1980
THE DYNAMIC CYCLE OFEDUCATION DEVELOPMENT
Figure I
37
--IDENTIFYING USERS' LEARNING °enemaThe second step is to identify the learning objectives
which reflect the needs of each audience segment. For thetwo major groups of users, two sets of objectivesare identified.
The end users need to know how the product can beused to their benefit and how to obtain the informationwhich answers job.related questions and solves theirproblems. The following set of learning objectives mightbe developed to reflect end users' needs:
To obtain job-related information.To acquire a basic level of skills essential to theuse of the product.To explore and identify the capabilities of theproduct for solving problems.
The technical audience needs to know how to solvemore complex problems bow to keep the productrunning efficiently, maximizing the benefits to theirorganization. Tice learning objectives developed for thetechnical support personnel might be represented bythe following:
To understand the difficult concepts andprocedures which dcn1 with the activitiesassociated with the product.To explain and, where necessary, customize theproduct to meet the end users' needs.To identify the specific security requirements andthe constraints that can be imposed upon the userfor efficiency or security purposes.
PREPARING THE DESIGN PLANThe educational courses are developed to satisfy the
learning objectives which reflect the needs of the identifiedusers. The procedures start with gathering information.Tice education development specialist interviews a diversecross-section of specialized personnel; they contributeideas and techniques relevant to course development,which might shape the product learning process. People'scontributions are based on personal experiences in thefield during interaction with similar users or on ideas frompublished sources, or they are gathered from colleagues.
The neat procedure is analysis to find out how to fittogether the collected ideas, how to look for the simplestand most relevant approaches to be used in presenting thecapabilities of the product to the users, and how toproduce a design covering all of the important andpractical aspects of the product. Once the overall designpima is established, skills needed to accomplish eachobjective are specified and written down to form the basisfor group discussion. In industry, they refer to this groupas the document specification review committee. Thecommittee members are representatives of the samegroups who contributed ideas. They are called to reviewand comment on the design plan and the proposedspecifications. As a result of the document specification
I ;
Tools and Techniques for Instruction 27
review committee meeting, the education specialist reviser.-the form and content of the evolving course and rebuilds itinto a coherent design.
COURSE MEDIAThe media used in teaching are generally: handouts,
overhead projector tansparencies, the board, and theinstructor's guide. This guide should be prepared carefullyto provide consistent instruction. Wall charts are alsoimportant course media; they are developed forcontinuous reference during class session. Examples ofwall charts used for IMS/VS-envirorament data structureare shown in Figures 2 and 3. Figure 2 is intended for thetechnical group who normally looks at a hierarchicalstructure of logical statements from top to bottom.Fixate 3 is intended for the end users, presenting the sameconcept in an approach that deals with building arelationship within a framework of reference, a readableleft-to-right flow.
TIME AND PLACE CONSIDERATIONSTime and place must be considered in course
development techniques. The duration of the course, whenin the sales/installation cycle it will be taught, and where itwill be taught all affect the design of the course. Theeducation specialist has to ask some questions abouttiming: Is the class needed before the installation of theproduct? If so, this may give the technical group anopportunity to become aware of what is expected duringthe installation procedures. Should it be taught duringinstallation? When does the end user need theinformation? The education developer also has to considerlocation: Where will the course be taught? Will studentshave access to terminals? Will actual data bases be usedfor examples? The choice of customer site or a regionaloffice has impact in these areas. These decisions are madeby product management, marketing, and productcommunications after considering both technical andmarketing needs.
IMPROVING AND EVALUATING THE COURSEThe process of course development starts with laying
out a preliminary desks resulting from selected ideas andcommittee discussion. To improve the course, dry runs areconducted before an audience representative of the groupfor whom the courseware is being developed. These dryruns help determine the effectiveness of the instruction,which is measured by the feedback collected after each dryrun. Changes to the course are introduced, and the courseis retested in another dry run before a new audience; morefeedback is collected, evaluated, and compared to theresults collected from earlier tests. This comparisonmeasures the degree of improvement.
3 s
SAMPLE USER DATA BASES
PLANT SKILL
LPLANTFLANT.ID.FLANT.NAMEFLANT.PHONEPLANT.REGION
...whl.PROO.CDOE. EMP.NO.
MODPROD.DESCPROO.AMT EMF
EMP.NAMEEMP.SEX
PRODAITY
I--.SALYEAR.SALYTDSALDED
ED YEAR'EDDEGREEED.SCHOOLSAL ED
'DENOTES SEGMENT KEY FIELDS
SUB
SUO.NAME.SUS.GRADE
SKILL.CODE.SKILLNAME
pLANT.io'
EIN.NO.
Figure 2
39
Tools and Techniques for instruction 29
0
.
qb
BUILDING A FRAMEWORK(defining relationships)
PLANT. IDKarr mow. NAAE
PLANT. PHONEPLANT. REGION
PROM CODE
PRODUCT PRCS DESCPROD. NATPROD. CRY
DAR NOOWLOYEE Di? NOE
SALYEARSALARY ISALYTD
SAL.DED
YRECUCARON ED.DEGFEE
ED.SCHOCI.
SWECT 'MENSUB. NAME
PLANTSUB.ORAD
SKILL SIPSKILL. CODE FLAME, (mewSU.. NAPE
Figure 3
40
30 NECC 1980
VALIDATING AND INTEGRATING THEINSTRUCTIONAL DESIGNIn industry, product communications group should notrelease the course to the field before a field test at acustomer's location. After teaching the course to theintended audience at a customer site, the developer andcourse instructor collect written and oral comments thataid the developer in determining the degree of success inachieving the identified objectives for the course. The fieldtest itself indicates that the minimum evaluation effort inproducing a product learning activity is accomplished.Responses produced from the field test reflect the attitudesand feelings of the users to the course learning materials.In the process of validation, some of these responses areconsidered very useful to field instructors and areintegrated in the instructor's guide to assure that fieldinstructors are aware of users' expectations. Otherresponses may lead to the elimination of misconceptions inthe course presentation. These changes are made beforethe alai release of the hinting document; however, noattempt is made to customize the instructional activities toevery individual customer's needs. Examples arc set to atypical and general application familiar to users. Theseexamples can be modified, based on users' feedback, tobecome more realistic in the final touches for refining thecourseware product.
SUMMARYIn summary, the instructional design technique that
produces a successful learning activity should followthese criteria:
1. The learning activity or class material is designedto achieve specific learning objectives. Mainly,these objectives are to impart the knowledge ofthe product capabilities to the users and to enablethem to use the product efficiently.
2. The learning activity is designed to offerreplicable instruction, that is, instruction thatcan be taught in the field setting by any of thefield instructors. The instructor's manual servesas a guideline to assure consistent instruction.
3. The learning activity is designed so itseffectiveness can be tested and demonstrated.Typical problems, based on user needs, shouldbe solved in class. Hands-on practicalapplication of what has been learned should alsobe provided.
1. .USIONare products and their instructional activities
should be developed to work for the users' benefit. Usersare encouraged to be a part of this dynamic process ofeducation development. Their feedback is needed in orderto learn how the product and its learning process can bemade to work better.
Product educ. ion development techniques and theevaluation cycle have no time limit. They start withbrainstorming but never end as long as the users are activein monitoring the product and its educational programs toservice their job-related needs.
4r
Tools and Techniques for instruction 31
CONSIDERATIONS AND GUIDELINES FOR DEVELOPINGBASIC SKILLS CURRICULUM FOR USE WITH MICRO-
COMPUTER TECHNOLOGY
Robert M.. CaldwellDivision of Educatior.:2 e.t4diesSouthern Methodist University
Dallas, Texas-76276(214) 692-2347
The availahil. of low cost micro-computer technology is creating a revolu-tion in education. Institutions of alltypes can now take advantage of the manybenefits offered by computer-based educa-tion at a cost that is easily affordablefor most. In addition, advanced micro-processor technologies interfaced with awide range of audio-visual devioes in justa few months have increased the oapabili-ties of micros to include color, graphics,animation, and music and speech reproduc-tion. In short, microcomputers have madeavailable an extremely fldxible and power-ful teaohing medium at a prioe that isfinally cost-effective for most users.
In response to this growing interestin the use of microcomputers in instruo-tion, several major publishers currentlyoffer limited curricula in basic skillsfor delivery on microcomputers. A num-ber of school districts such as the DallasIndependent School Distriot and those dis-tricts affiliated with the MinnesotaEducational Computing Consortium also havedeveloped materials whioh cover a varietyof skill areas. Because of the signifi-cant expenditures associated with develop-ing these programs, however, most of themare limited in scope and employ a rathernarrow range of teaching strategies andmachine capahilities. In addition, fewof the individuals ourrently engaged ininstructional deveW "nct have had muchexperience in using Highly interactivemedium like the Mi4 umputer. As a re-sult, much of what passes for coursewaretoday neither helps to develop higherlevel cognitive skills nor challengeslearners to use higher order learningstrategies. Most lessons utilize a drilland practice format in which thecolputerasks a question and requires the studentto respond. Many so called tutorial pro-grams are no improvement. They merelypresent segments of expository text and
then display questions to test the student'scomprehension of that text. This sort ofof instruotion has its use but implies tomost educators that the computer's powerlay only in its ability to present textand ask questions. This form of instruc-tion serves to reoreate the very worst ofwhat presently ocours in a traditionalclassroom OUR ignoring all other teach-ing strategies coat oan help develop learn-ing styles and learner independenos(Garson,1980).
To complicate this problem further,some oompanies are now developing author-ing systems which will allow classroomteaohers, students and curriculum devel-opers to create their courseware througha process which utilizes templating and/ormenu selection. To be sure, many edu-oators will use these processes to createexciting programs which utilize thesystem's capabilities to its fullest.Many others, on the other hand, willduplicate the type of programs mentionedabove; they will need help, guidanoe andeducation about what microcomputers cando and how instruction can be presentedto take full advantav of the uniquefeatures of the new technology. Theywill need explicit guidelines that willhelp them devise ways to make contactsbetween the learner and the learningexperience more meaningful, more effi-cient and more productive. The purposeof this paper, therefore, is to presentspecific guidelines for designing instruc-tional programs that will be deliveredon microcomputers so that those programswill use the capaoity of the microcomputersystem in a way that will develop inlearners a range of cognitive skills andhelp learners develop useful learningstrategies.
32 NECC 1980
GENERAL FEATURES OP PROGRAM DESIGNOne of the most important factors
inherent in programs delivered on com-puter-based systems is their ability toadapt instruction truly to the individ-ual needs of each learner. With this inmind,then, programe of computer -basedinstruction should include the followinggeneral features (Caldwell and Rizzo,1979):
1. Learner Control over theinstructions iiWaia-re a feature ofprogram design often ignored by instruc-tional designers. Certainly there aresituations in which a presentation ofinstruction must be linear. Certain sub-jects and concepts require it. In mostother cases, however, considerationshould he given to allowing the learneras much control over the learningsequence as possible. Options should beincorporated into the instructionalsequence which allow for review of pre-vious frames; for decisions about thetype, difficulty level, and number ofproblems or exercises received; and foralternative branching routes that mightlead-to the accomplishment of lessonobjectives in less time. In short,students should be given the opportunityto advance, review, and'exit lessonsexcept where such control defeats thepurpose of the lesson. This ability oflearners to pace themselves provides adegree of individualization not presentin purely linear programs.
2. A system should be totallyindividualized and offer highly adaptiveand res onsive learning environments.By a ow ng self- pacing adk. individ-ualized branching, learners are helpedto select the pathway through theeateries that is most appropriate fortheir needs.
3. Programs should be modularizedand structured in coherent, hierarchiapatterns. This type of organizationalpattern allows for great flexibilityin program implementation becausecurriculum materials specifically ad-dress each necessary skill in a definedcontent area in a manner that providesfor development of skills not masteredor allows bypassing of skills whichhave already been mastered. This pro-cess can reduce student frustrationand increase curriculum effectiveness.
4. All skills to be masteredshould be carefully stated in or-formance objectives. The accuracEl-i7W.Biram is based on the spec fiedefinition of the objective in termsof performance competencies. Activi-
ties allow for precise diagnosis ofskills already mastered, remediation inskill deficiencies, and exact evaluationof learner progress.
S. Progress should be measured interms of mastery of performance objec-tives.
6. Strategies for dia nosie and4 prescription should be used. e effic-
iency of a program is due primarily to thediagnostic inventory made of the ekillsof each learner. This information canthen be used to place them appropriatelywithin the curriculum and to direct learn-ers to the instructional material mostAppropriate to their needs.
7. Programs should be, when possible,multi - sensory in format,
SPECIP:T GUIDELINES POR INSTRUCTIONALDEVELOPMENT
Programs of computer-based educationhave been developed in a variety of de-signs and formats. Some use drills arrang-ed in stionds while others are built arounda series of tutorial leesons. Many evenincorporate all or most of the featuresmentioned above. Within these programs,however, are characteristics of instruc-tional design that heavily affect thesuccess of the instruction. The followingie a description of some of the more com-mon characteristics of lesson design thatcan contribute to instructional effec-tiveness and some that can seriously de-tract from it.Creating Text and Graphic Displays
Many alternatives can be used to breakthe monotony of lines of text filling anentire screen:
1. Use graphics to box important sen-tences or paragraphs. Boxes made of linesor keyboard characters do nicely to alterthe visual display of text on a monitor.Reverse highlighting ae. color can also beused .effectively to accentuate visual dis-plays. In addition, microcomputers can beconnected to graphics tablets which makeextremely interesting displays in almostany size or shape.
Whatever is used, it is crucial tointerrupt continuous lines of text onscreen so that the screen does not lookcrowded and cause verbal overload in stu-dents. Too much text can have the effectof discouraging the learner, especially ifhe/she is a poor reader.
2. Allow the student control over thesequence of presenting text by breakinglarge portions of the text into discreteunits or segments which the student cancall up in sequence by merely pressing thespace bar or some other key activated forthis purpose. This procedure serves two
purposes: it allows learners to read at arote that is appropriste for them, and itbreaks the reading task into small segmeatseo that the reader ie not overwhelmed bypage after page of text on the screen.
3. Animations, graphics, cartoon char-acters and other creative devices serve tocreate variety end interest in the displayWhich appears on the monitor. Used cre-atively, these capabilities of microcompu-ter systems can contribute greatly toeffective instructional programs.
4. Double spacetext material wheneverpoeeible to enhance the visual effect.
5. Use color to enhance the display orto highlight' whenever possible. One pro-gram uses color-coded feedback to distin-guish it from other text and to emphasizekey concepts. Color is also useful in pro-viding prompts and to direct attention tovarious portlons of the screen.Creating the Instructional Sequence
1. As a general rule, try to showlearners rather than tell then. TU.over-use of exposition is the single biggestmistake instructional designers make.They feel as if they suet write a lectureinto each frame or prOline complex direc-tions, instructions or explanations whenthey can very simply use graphics or theability of the computer to erase, rewrite,flash and even animate to make conceptsclear. (The presentation which accom-panies this paper illustrates this pointwith k number of examples.)
2. Make lessons as interactive aspoeeible. Force learners to Makechoices; help them make decisions byproviding them with options and altern-atives. Simulation and dialog programsare the beet instructional strategy forpromoting interaction, but it can be pro-vided through other means as well:
a. MenusLearners may choose from a variety
of options within a leeeon or within aprogram by making choices from menus.These menus allow the learner flexibil-ity to pursue his/her interests or to con-trol the sequence in which topics are pre-sented. For example, s language leeeonWhich deals with predicates might allowstudents to access concepts which have noparticular instructional sequence. Figure1 illustrates Just such * menu.
b. stftticerfarmatie
Learners SIMMUAIMI Oven a variety ofactivities and choice Of content. Tutor'-isle should be accompanied by creativedrills or instructional games that rein-force skills and intonation and enhancethe interactive mature of the instruc-tion. These games and drilla can stimulatesotiVation hy capitalizing on a novelty
Tools and Techniques for instruction 33
Figure 1
PREDICATES
Which mould you like to study
1. Predicate Nominative
2. Direct Object
3. indirect Object
4. Predicate Adjective
5. Review
a. Linking verbs
b. Action verbs
c. Adverbs
Choose a number or letter
effect. Competition in the games can beagainst other students, the computer it-self, time limits, or performance criteriaset by other students (e.g. "the best scoreon this game to date is " or "the beettime for this game ie 613".
A word of cautiorli-appropriate here,however. In en attempt to bring novelty totheir programs, some designers have defeat-ed their own goals. In one such leeeonstudents are encouraged to save a tiny manwho is standing in a deep pit from a largeheavy stone which ie slowly rolling down ahill on the left side of the pit. Thisteat is accomplished hy solving $ Seriesof math problems successfully. With eachcorrect response the man moves up the rightside of the pit toward freedom. If thelearner fails to get the problem correct,however, the stone rolls closer and loseruntil it eventually falls into the pit andsquashes the poor fellow. In an observa-tion of this sequence with several child-ren, a number of them chose to see the manget squashed with the full knowledge thatto do eo they must fail every single mathproblem presented to them. This exercise,therefore, did little to develop math skillsbut did much to distract the learners fromthe true intent of the leeeon.
c. PromptsThe power of computer-based instruc
tion resides in its ability to ehape learn-er behavior toward learning outcomes in away not possible with most other media. Animportant factor in shaping hehavior ie theUse of prompts. Designers who ignore theuse of prompts often turn the instruction
44
34 NECC 1980
into a guessing game. For example, a numberof math programs currently available presentproblems then simply respond to the learn-er's incorrect response with a "No, tryagain." Thie type of feedback gives thelearner no information about how he/she iedoing; it only promotes guessing until thecorrect answer is discovered. Prompts suchas, "Too high, try again" or "Remember,carry the ones," on the other hand, pro-vide learners with guidance toward thecorrect answer. Without them, interactionbecomes meaningless. More about promptswill be said later in this discussion.
d. Task DescriptionLearners are very cation misled about
what it ie they are supposed to do or howthey are supposed to respold in a givenexercise because the instructional designerhas confused the task for them. The follow-ing i4 a perfect example: In this lessonthe learner is presented with the task ofdistinguishing complete and incompletesentences:
The store closed early.
Type comp (for complete)or inc(forincomplete)
The task here ie clearly discriminationbetween complete and incomplete sentences,but it ie confused by asking the learner totype "comp" or "inc." The instructionalsequence is marred not only by a poorlydesigned drill and practice but also by apoorly conceived response format. An im-proved version would simply ask the studentto respond by typing a 1 for complete (or ac) or a 2 for incomplete (or an i), thusfocusing attention on the discriminationtask rather than the typing task.
Task confusion can also be lessened oreliminated by providing learners with asample exercise before they are engaged ina drill or a quiz':
Input and Response1. The method an instructional de-
signer chooses to facilitate learner im-put and reeponee ie largely a function ofthe type of learning outcome desired. Res-ponse modes most commonly include typed res-ponse, touch reeponee or light pen response.The first criterion for choosing one of thesemodes is whether the response will helpreach the objective of the lesson. Oneillustration of this rule is in the ex-ample cited earlier in which etudente areasked to discriminate between complete andincomplete sentences. In this case a typ-ed reeponee was inappropriate to the objec-tive of the lesson. In fact, typed respons-es in general seem prone to high error ratesbecause of the nature of typing tasks (Cald-well, 1974). Typed responses also become
time consuming especially if learners areunfamiliar with the keyboard. This problemcan have the effect of distracting learnersfrom learning the concept being presented.A related problem ie that many errors intyping are judged by the computer to be in-correct responses, and therefore etudenteare routed to remedial sequences or to seg-ments of the lessons they have just beenthrough.
Similarly, if typed responses are notcarefully cued, learners will often be for-ced into closed loops; that is, they try tofind the correct answer by typing randomresponses. When this problem occurs learn-ers have little idea about what is expectedof them and become frustrated trying to findthe response that will advance them to thenext frame. Thie common error in instruc-tional design has an extremely negativeeffect on learners.
Some advantages can be found in thetyped response. however, which are not pres-ent in forced choice reeponee modes (Cald-well, 1974):
a. Learners seem to express more favor-able attitudes toward writing and the useof verbal language.
b. Spelling skills and the ability togenerate language improve.
c. Personalized feedback ie made poss-ible through direct entry of student names.
2. As mentioned earlier, response bydirect entry should not be required of astudent without an example or a practiceexercise. Thie can prevent learners frommaking errors in the practice exercises.These errors can become important, particu-larly if one ie using response data tobranch learners or to do item analysis.
3. A number of other helpful suggestion*can be found in the guide to developing ins-tructional software developed by the Minne-sota Educational Computing Consortium (1980).Some examples from that guide follow:
a. Be consistent in the way questionsare asked and the required format for res-ponses. Do not code answers unless it ienecessary. Request the responses of YES orNO rather than coding responsee such as 1YES and 00NO. When processing these answerscheck the first character to determine if itie "Y" or "N". If it ie neither, erase thestudent's answer and re-ask the question.
b. Ask questions while the informationneeded to answer them ie still on the screen.Don't ask a question then change the screento ask another one leaving the choices forresponse on the screen jut erased. Ifthe student must choose from a fixed eet ofoptions, then those options should remain onthe screen throughout the quiz sequence. Insome cases it ie helpful to allow etudenteto go back to formerly displayed frames to
refresh their memories. This review ieaccessed by means of the WNW key.
c. Put options for a question in aCOIMID. Do not imbed them in a sentence.Options having multiple words should becoded with numbers or lettere.
d. Answer checking after accepting'in-put deserves specific consideration. Don'task a question and expect a very specificanswer. For example, if the question ie"What is the capital of Minnesota?" somelearners might answer, "The capital ie St.Paul." If the response judging is too rig-id and is looking for just "St. Paul" thisstudent's answer would be wrong. It ietherefore very important to do keywordmatches rather than a match on the wholeanswer entered.
Also, when looking for correct answers,consider how many times to allow a studentto miss a question. A typical method usedby may designers ie to ask a question amaximum of three times. This way, thelearner has the opportunity to be prompt-ed twice. Any more than three tries seemsto indicate that the student has not mas-tered the concept and it ie beet to movehim/her on. However, it is a good ideato ask the same question again later.
4. The number of questions presentedduring a formative evaluation ie the sub-ject of considerable debate. Essentially,the number of questions presented in anyone exercise ie dependent upon the objec-tive being assessed. As a general rule,however, five (5) ie a number that seemsworkable for most learners. It ie asufficient number to assess progress to-ward mastery of the objective and requiresjust enough time so that the exercise it-self does not become tiresome for thelearner. Five items ie a good number forgeneration of item pools also if one usesthe guideline that three times the num-ber of items should be written as arepresented in the exercise.
5. One must also be wary of the type ofinformation learners are permitted to in-put. For example, a common practice inmany computer- ascieted lessons is to sek,"What ie your name?" or "What do yourfriends call you?" The purpose of such aquestion, of course, is to use this infor-mation later to provide personalizedfeedback or to use the learner's name inthe content of the presentation (e.g."MIMO to: Johnny"). This practice can some-times have disastrous effects if thelearner decides to get cute. One student,for example, was observed to enter "idiot"to just such a question as those citedabove. All successive feedback after thatresembled statements such ae, "No, Idiot,try again" or "Good work, Idiot." While
1
Tools and Techniques for Instruction 35
this type of response can delight 'earners,its overall effect ie one of distractingthem from the purpose of the lesson.
RWinforceftent-Considerable space has been devoted in
this paper to suggestions regarding respons-es to various types of imput. The sectiondealing with prompts serves as a good ex-ample. Generally, reinforcement should bespecific and directed at providing informa-tion that will help shape the learner's be-havior toward the desired learning outcome.As an illustration, consider the followingsample sequence:
ing to the word- -
Come)(Here students are required to transformthe word by retyping it with an -ing end-ing.) A correct response on the firsttry would warrant positive reinforcementsuch ae "great, super, well done, excel-lent." These and other reinforcing state-ment(' can be generated randomly from alist of 20 or 30 possible statements. In-correct responses on the other hand,should follow a pattern of prompting whichshould lead the student to the correctanswer. For example:
First incorrect response: Come)giggling (drop the e). The student ietiedlfrj reminding btm/her of the ruleunder examination (drum the e).
Second incorrect response: Come,comiing (come + ing). Here the learner
different response. In thesecond prompt the computer system animatesthe a in come and it drops from the equa-tion then-f6W-ing moves into place.
Third incorrect response: Come,gmipx. If on the third try the studentstill not type the correct response,the correct answer should be given, andthe student should then move on to thenext problem. This type of reinforcementpattern contributes to meaningfulness ofthe material to be learned because it:
a. Provides specific informationthat helps guide correct learner behaviortoward achieving the desired outcome.
b. Reduces the frustration oftenexperienced by learners in CAI programsthat simply provide a "no" response to anincorrect answer. A simple "no" ie un-satisfactory because it provides nospecific feedback that helps the learnerto discover correct responses. Insteadlearners are forced to guess until thecorrect answer ie found (Caldwell, 1070.
Reinforcement strategies dependlargely on the nature and type of learningbeing attempted. Reinforcement, however,can be made more meaningful if it is per-sonalized and specific to student
4V
36 NECC 1960
response sequences. (e.g. "You did well,Allen, but would you like to review pre-fixes before going on?")
In summary, the design of instruc-tional programs should incorporatevarious teaching strategies to addvariety to an overall curriculum and toaddress cognitive processes at all levelsof that domain of learning. Also, pro-grams should use to their best advantagethe features of color, graphics, and.animation offered by microcomputer tech-nology. Exciting new developments inspeech and sound periphekals also existand should be incorporated where relevantand appropriate. Also, learner controlover the instructional sequence allowsfor individual pacing and better oppor-tunities to achieve mastery throughbranching, diagnosis and remediation.This flexibility can add significantly tocomputer based programs if they aredesigned and implemented carefully.
The scope of this presentationlimits discussion of the many subtlevariables which contribute to effectiveinstructional design, but attempts tobegin an organization of some of thefeatures which seem to contribute tosuccessful programs. It is hoped thatit might stimulate other authors anddesigners to share their rationales andexperience in designing high qualityprograms of computer-based education.
REFERENCE NOTESCaldwell, R.M. "The Effects of Selected
Strategies for Teaching Reading toNon-literate Adult Learners UsingComputer-Based Education." Paperpresented at the annual meeting ofthe American Educational ResearchAssociation, San Francisco, April,1979.
. "Evaluation of a Program ofComputer-Assisted Reading Instruc-tion for Semi-literate Adolescents."Paper presented at the annual meetingof the American Educational ResearchAssociation, Chicago, March, 1974.
REFERENCESCaldwell, R. M. and Peter J. Rizzo.. "A
Computer-Based System of ReadingInstruction for Adult Non-readers,"AED8 Journal, Summer, 1979, 12,4.
Gerson, f7-177TWe Case Against MultipleChoice," The Computing Teacher,February-March, 1980, 7,4.
Minnesota Educational Computing Con-sortium User Services, A Guide toDeveloping Instructionanarefor the Apple II Microcomputer, St.Maul, Minnesota: The Minnesota
Chaired by David KniefelNew Jersey Educational Computing Network
THE AFFECTIVE AND COGNITIVE EFFECTS OFMICROONNPuTER-BASED SCIENCE EDUCATION
Ronald E. Anderson, Daniel L. Rlassen,Thomas P. Hansen, Minnesota EducationalComputing Consortium
ABSTRACTMicrocomputers are used increasingly
to deliver and enhance science instruc-tion, but the impact of these newtechnologies has not been extensivelystudied. An experiment was designed toinvestigate the effects of a brief CAIactivity on student attitudes, beliefs,and knowledge. Three hundred fiftyhigh school students were tested and re-tested before and after taking a 20-30minute lesson on water pollution. Theinstructional material was all deliveredby an APPLE II microcomputer using highresolution graphics and some color vari-ations. The findings support the claimsthat: (1) even a very brief CAI exposurecan be instructionally effective; (2)
understanding about computers, i.e.,computer literacy, can be improved as aconsequence of such CAI: (3) studentsbecome attitudinally more positive aboutcomputers from microcomputer CAI: (4)graphic enhancements may not improvestudent responses; (5) system malfunctionsmay revise student conceptions of them-selves as well as computers.
The micro offers a vast potential. withsome possibly questionable effects.Since it allows new kinds of person -computer relationships, we can not presumeto know its ultimate impact on individuals.
THE EFFECTS OF MICROCOMPUTING ON LARGECENTRALIZED TIMESHARING
Kent T. Rehrberg, Minnesota EducationalComputing Consortium
ABSTRACTimiesota schools have already acquirednearly 1,000 APPLE II and other micro-computers, even though they have hadaccess to a large central computer viaa statewide educational network. Thissituation offers opportunity to researchsome important questions such as: (1) Howdo people justify their acquisition ofmicrocomputers? (2) How are micros usedin the schools? (3) Does microcomputeracquisition substitute for or foster time-sharing use of a centralized facility?The MECC research serves as the basis forprojection of the role of microcomputersand instructional computing in the short-
' term future.
DISCUSSANTS
Harold Peters, Associate Director, CONDUITJohn Castellano Jr., Indiana University
37 48
Computers in Humanistic Studies
CREATIV/TY THROUGH THE MICROCOMPUTER
George M. Bass, Jr., Ph.D.Assistant ProfessorSchool of Education
College of William and MaryWilliamsburg, Virginia 23185
804-253-4289
"Creativity is a battleagainst fixed attitudes."
(Raudsepp, 1977, p.55)
In many ways teaching a college courseon creativity is very similar to teachingstudents hdw to use a computer. Both setsof students come with many preconceivednotions about the subject. Usually theseattitudes have not been built upon directexperience with computers or creativityexperiences but on hearsay and vague ex-pectations. This lack of experience hasoften led students to be fearful of theirown ability .o be fyeative or to masterthe computer. Yet a willingness to breakfree from theba frozen views and to seeideas and problems from a new perspectiveis needed in both areas for the studentto succeed.
During the fall semester of'1979, Itaught an advanced education course oncreativity. This course was designed tobe a seminar for students who had success-fully completed an introductory education-al psychology course. My overall goalswere to present students the necessarybackground about creativity as it hasbeen studied by psychologists and educa-tors; to introduce them to instructionalstrategies which may help develop crea-
38
tivity; and finally, to increase theirown creative abilities. Eleven studentstook .the course.
COURSE ACTIVITIESIn order to distinguish this course
from just another elective, I wanted touse activities not typically found inother courses they had taken. Enteringnational contests, participating in crea-tive growth games, working on word puz-zles and logic problems, reading and solv-ing detective stories, producing poetryand written compositions, and using amicrocomputer were introduced throughoutthe semester. (More traditional activi-ties such as assigned readings, classpresentations and projects, and a finalposition paper were also used to increasethe learning process.)
Because the microcomputer is such anew educational technology, it is anideal way to break the chains of studentexpectation and release any hidden poten-tial. Yet the fear of the unknown or theexaggerated visions of computer power(such as HAL in 2001) were a real concern.In order to redarfhis possibility, Ibegan the first class with TLC -- TenderLoving Computer. Other exercises usingthe capabilities of our Apple II micro-
49
computer were later interspersed through-out the course.
MICROCOMPUTER ACTIVITIES --OBJECTIVES AND RESULTS1. Course Syllabus
Since this was the first time thecreativity course had ever been taught,I felt it necessary to give students mycourse objectives, assignments, and re-quirements during the first session.However, I wanted to communicate thisbasic information in a nontraditionalway as mentioned earlier. I also wantedthe students in the class to meet oneanother and to learn of any personal ob-jectives each individual might have forthe class. To accomplish these objec-tives, I wrote a program on the microcom-puter to interact with each student.After requesting the student's name andsome background information (interests,reasons for taking the course, etc.) andengaging him in chit chat, the programpresented relevant information about thecourse and then introduced the studentto the capabilities of the microcompu-ter. Using a variety of programs cur-rently available for the Apple II, thestudents were presented with computergraphics, text manipulation, and gameplaying experiences, e.g., Rock/Scissors/Paper, Dragon's Maze, Star Trek.
The results of this experience werequite positive. The students had neverseen a microcomputer before and weresurprised with the variety of things itcould do. As each student took a turn,the rest of the class crowded around thecolor monitor to watch the action. Theygot to know each other through the typedconversation with the Apple II. Theyalso got to see many of the programsavailable to them outside of class. (Fortwo students this introductory sessionled to a semester-long journey with StarTrek after almost every class!) Thiscomputerized syllabus generated interestand set the tone for an unusual class.2. "How I Spent My Summer Vacation"
AssignmentOne of the most common assignments
that teachers give when students returnto school in the fall is to write a themeon how the student spent his summer vaca-tion. In order to get the students inthe class to begin thinking of novel waysto solve problems or accomplish tasks. Igave the class this typical task, butasked them to solve it in a creative way.When they next came to class with theircreative solutions, I shared with themone solution using the microcomputer. Iprogrammed a story which contained an
Computers in Humanistic Studies 39
overall structure but required the addi-tion of certain words and information fromthe user to compose a complete tctle of howthe Apple II spent its summer. Besidesshowing how unique paragraph-length sto-ries can be created using a microcomputer,this exercise laid the groundwork for afuture class discussion on form versuscontent in creating something new.
While the students' creative solutionsto this assignment were indeed varied (apicture scrapbook, a summer job advertise-ment, "first grader's" paper, fantasystory), their response to the Apple IIstory was enthusiastic. They worked as agroup supplying the adjectives and otherwords requested by the program. However,the first time through the story, they hadno idea how these words were going to beused. The resulting narrative was a hu-morous, if not entirely accurate, accountof what a microcomputer might do over thesummer on a deserted college campus. Thesecond time through the program the stu-dents chose their words more carefully tofit into the structure of the story theyremembered. While this second experienceresulted in a new version of the tale, insome ways it was not as entertaining asthe first version. In resulting classdiscussions the notion of appearance ver-sus substance, "how said" versus whatsaidilith,are role of the creator and hisaudience in giving meaning to a creationwere all related to this computer/human-generated story.3. The Apple II Pops
The apple II microcomputer can makemusic-like sounds through the Apple'sbuilt-in speaker. A number of commercialprograms are available to create thesesongs with minimal programming. Using theFORTE Music Interpreter by Rainbow Comput-irs Inc., I showed the students how toprogram and save songs they composed usingthe Apple's speaker. I wanted the stu-dents to construct a nonverbal creation,to become more comfortable with using themicrocomputer as a tool, and to learn anintroductory programming approach.
At the class concert during which eachaspiring composer introduced his creativeeffort, it became readily apparent that myobjectives were achieved. Although thevariety of songs (from birdlike melodiesto atonal experimentations to K-Tel musicadvertisements) again showed the diversityin the class, all students completed theassignment and expressed their growingease with the Apple. They commented onthe computer as most forgiving, but alsomost demanding. Mistakes were easy tooorrect by retyping the note or line, butputting commands in the correct syntax was
50
40 NEOC 1980
required for the program to work. In ad-.dition, the students expressed an appre-ciation of each other's efforts regardlessof previous musical experience.4. Aphorism Construction
Adapting an Aphorism Generator pub-lisheAby J.D. Robertson (Creative Com-utin , August 1979), I programmed thepple to construct aphorisms such as"Alcohol is the liver of stress" throughrandom combinations of lists of the threekey nouns. The class was asked to generaate sayings using the format " is the
of ." as well. Pour stints werechosen to do this individually, Is oftwo and three students were chosen to doit collectively, and the remaining twostudents worked as a team to select thoseaphorisms generated by the computer whichthey thought most creative. After eachindividual or team had chosen their besttwo sayings, these aphorisms were read tothe rest of the class without identifyingthe origin. Each class member separatelyranked the three best aphorisms; theserankings were then tallied and the high-est rated sayings identified. My objec-tive was to stimulate a discussion ofindividual, group, and random efforts increating original products.
Although the results were certainlynot definitive (one pair had written bothof the top ranked sayings), they did leadto alively debate on the difference be-tween random replacement and meaningfulcombinations. The idea of "creativity isin the eye of the beholder" was also rein-forced by this exercise. The writer andthe reader put meaning in aphorisms &M-other creative products.
IMPLICATIONSThe use of these microcomputer activi-
ties added greatly to the overall achieve-ment of the course goals. One reason forthis success was the close fit between thecourse topic and the capabilities of thecomputer. Indeed, creativity with its"battle against fixed attitudes" stressesa concern with new images and experiences.Torrance (1979), a major figure in crea-tivity research, has even recommended athree-stage instructional model to en-hance incubation and creative thinking..Activities using a microcomputer can beimplemented at each stage in his model.Stage 1 is aimed at heightening anticipa-tion; it tries to motivate learners torelate each learning task to meaningfulexperiences. The course syllabus andsummer story certainly attracted the stu-dents' attention and stimulated theircuriosity and imagination. Other compu-ter activities such as role playing games
and simulations should also warm-up stu-dents and increase their anticipation.Stage 2 burns this initial interest intodeepening expectation. Song developmentis one example of the yellow; informationprocessing patterns which Torrance sug-gests. Stage 3 simply tries to take crea-tive thinking and keep it going. Seeingthe various capabilities of a microcompu-ter in this course should permit studentsa better understanding of future .techno-logical applications in their own lives.
Although psychologists have perhapsfocused more on the instructional demandsof creativity and problem solving coursesthan other subjects, the psychologicalprinciples underlying such instructionalplanning and design van be generaliz..d toother college subjects as well. Treffin-ger and Huber (1975) have emphasized thatthe content of any course can be analyzedaccording to an instructional systems ap-proach. Such an approach focuses on iden-tifying specific instructional objectives;diagnosing characteristics of enteringlearners; developing sequential, learninghierarchies and teaching strategies toaccomplish the objectives; and assessingperformance to determine learner achieve-ment. Following such a basic instruction-al model will allow any teacher to esti-mate the value of particular educationaltechnologies in a chosen course.
The specific issues for incorporatingthe computAr into this current educationaltechnology have been widely discussed dur-ing the past decade. Usually the chiefcriticisms against such a move have re-volved around cost and teacher resistance(Trow, 1977). Yet these concerns havemostly been associated with computer-assisted instruction using a large time-sharing system. Recently, there has beena shift from CAI programs aimed at indi-vidualized teaching of a single subjectmatter toward computer-managed instructionin which the system directs the entire in-structional process (Splittgerber, 1979).Even with this move toward CM/, therestill remains an implementation probleminto the public schools. Economic, tech-nological, and instructional design con-siderations have kept the potential bene-fits from being realized.
But a revolution is in the making:With the recent rise in accessibility ofmicrocomputers, the cost argument has beengreatly deflated. With the groWth inmicrocomputer design and capabilities,many of the technological concerns aboutaccess, memory storage, and program lan-guages have also been met. With the grow-ing commitment to instructional' designrelevant to CAI and CMI, better software
is becoming available. Nevertheless, itteacher resistance is not overcome, allthese changes will probably be for nought.
The experiences detailed in this paperreveal one possible strategy to introducethe applicability and benefits of micro-computers. By incorporating course con-tent with microcomputer capabilities,students can gradually come to appreciatethe vast educational potential in thisnew technology. With such a focus oncomputer-enriched instruction, the teach-er is less likely to fear displacementfrom the teaching/learning process. Theemphases of CEI experiences are on stimu-lating class discussions and group inter-actions, not taking over the teacher'srole. As Eisele (1979) has pointed out,the possible uses of microcomputers inthe classroom are many: as tools forcreative problem solving, games and simu-lations, drill and practice; testing andtest construction. By taking advantageof these uses to enrich classroom activi-ties, today*s educator can be accountableand productive without feeling deperson-alized or replaceable.
And isn*t that a reasonable way towin the battle against restrictive fixedattitudes?
SUMMARY AND CONCLUSIONSSince many students and teachers are
hesitant to use the computer because theyare unfamiliar with what it can -- andcannot do, it is imperative that waysbe developed to introduce gradually amore realistic appraisal of the compu-ter's capabilities. If these benefitsare provided through computer-enrichedactivities which do not reduce the mainrole of the teacher, more acceptance ofsuch activities should be forthcoming.With the advent of microcomputer systemsalmost any small college can afford thecomputing power that was available onlyto large institutions just a decade ago.An example of such an application in a_OPllege course on creativity was pro-vided to illustrate the feasibility ofsuch a CEI approach.
Computers in Humanistic Studies 41
REFERENCES
Eisele, J.E. "Classroom Use,of Microcom-puters." Educational Technology.October 1979, P. 13-15.
Raudsepp, E. Creative Growth Games. NewYork: Jove Publications, 1977.
Splittgerber, F.L. "Computer-based In-struction: A Revolution in theMaking?" Educational Technology.January 1979, p. 20-26.
James W. Gets=Department of PhilosophyUniversity of Notre DaneNotte Dame, Indiana 46556
(219) 283-6471
Ovet the last thtee years, my colleague PaulMOWN* and I have been at wotk on an interactivecomputet ptogtam called EMIL which we use to helpteach our courses in formal logic. Since June of19/9, we have been devoting our efforts (thanks tofunding from The National Science Foundition) toimplementing a computerised "copilot" for EMIL
which the students can call on to obtain advice onhow to tackle problems that they find too diffi-cult to handle on their own.
The methods we are using to provide out stu-
dents with advice are easy to Implement and quiteeffective. They represent an approach to comput -*tired education that doesn't fit neatly into thestandard categories (i.e. recotd keeping, multi-ple choice CAI, testing, games, simulations,etc.), but one which has the potential for wide-spread application. The goal of thin kind oftutoring ptogran is to prompt the student todevelop and use creative strategies solve problemswhich do not necessarily have any one
correc answer. The advice-giving program isSocratic; it asks the students leading questionsconcerning the problem before them, questionswhich help them analyze and resolve their diffi-culties.
A typical course in formal logic requiresstudents to find formal proofs. Students arepresented with a sot of formulas of logic andasked to apply certain logical rules to them toderive some othce formula. The derivation whichthey are asked to construct consists of a ssqueues of formulas each of which follows fromprevious members of the sequence by one of therules, and which ends with the formula.they areto prove. This kind of learning is importantnot only in that it provides the foundation forthe mastery of the basic concepts of. logic, but
also because it gives the students the opportun-ity to learn some of the practical problems andstrategies involved in creative thinking, parti-cularly in creative thinking characteristic ofmathematics and the sciences.
Giving students practice in the creative so-lution of formal problems particularly import-ant in science education. Too often scientificknowledge is presented as if it is descended from
heaven, or required same form of superhuman intel-ligence for its discovery. Very little attentionis payed in science education to allowing studentsto appreciate the thinking processes which go intothe analysis and mention of scientific problemsin a real setting. There is a tendency to obscurethe very human process of fumbling around, of try-ing out strategies, of assessing failures, and of
cresting better lines of attack, which are allpatt of the scientists' daily life. A course inlogic gives students the opportunity to refinetheir skills et problem solving in an environmentwhere the difficulty of the problems they conftontcan easily be adjusted to their growing abilities.
In the standard sort of course, students' abili-ties at finding proofs vary widely, so that thosewho do not have an initial knack are severely or.wilted. Even when strategies for proof - findinge re carefully discussed in class, some students in-variably complain that they can't do a new problemon their own in spite of "understanding" the lec-tures. Pert of the problem for these students isthat they cannot convert a verbal explanation oftechniques. into s flexible tool for dealing with anew situation.
With a bit of tutoring, most students with thee.difficulties improve rapidly. If students thinkout loud while attempting a proof, a gentle nudgelure and there often leads to success. If theydon't understand the rules, or simply haven't both-ered to learn them, guiding then through a proof ortwo tends to straighten things out fairly quicklyand to improve their confidence and activation.Just as in teaching most skills, the effective
strategies involve letting the student perform thetask under guidance; lecturing on the proper pro-cedure and telling students to go home and do like»wise is relatively ineffective.
Of course there are good teems why tutoring isnot widely used in an introductory coots* in logic.These classes are usually quite large, so tutoringsimply-takes too much of the tes6er's UM*. Notonly that, grading exercises in proof-finding istedious, so teachers tend to give students rale..tively few exercises that require then to create aproof. Sven if students can learn on their own,they simply don't get enough practice to develop
53
any skill, unless they catch on right off the bat.Very often, the teacher relies on exercises thatrequire a single answer, such as those that askthe student to fill in the justifications for thelines of a proof that is already completed. Thisprocess does familiarise students with the rules,but it gives then no practice in the art of find-ing a proof.
Computers make it posigible to simulate the tu-toring situation. Students can enter their proofsat the terminal, and the computer can be program-med to see whether each line of the student's so-lution follows from previous lines and to describethe difficulty if anything goes wrong. When thestudent gets lost the computer can make sugges-tions about how to proceed.
This use of computers in education is particu-larly interesting because it departs radicallyfrom the multiple- choice foneshich has become al-most paradigmatic of computerised teaching materi-als. A proof-checking program doss not requitethe student to cowl tv ady preselected answer, butto find a solution by any of a potentially infin-ite somber of lines of attack. In a sense, theprogram does not demand an answer, but *imply pro-vides an ongoing check of the student's progressin achieving a result. It does not require a setresponse so such as provides tool that studentican use in their own way to develops skill.
Compared to multiple-choice programs, proof-checking progtams make heavy demands on the com-puter, the teacher, and the student. The coeput-
-- or must interpret the student's Name at theterminal, determine whether they ere correct, andrespond in an intelligent way. The student andteacher must familiarise themselves with the pro-cedures for operating the computer erosion, andsuet put up with the inconveniences caused by hav-ing to use a computer which is generally overbur-dened already, sad which occasionally malfunctions.They gust also put up with the inevitable mistakesa programmer sakes in designing the logic teachingsystem. And yet, if we are ever to develop com-puter teaching syetems that provide students withtwig' for learning, rather than merely with ongo-ing multiple choice exeminatians, we must over-come these difficulties. Working out effectivestrategies for proof - checking programs can pavethew for developing less authoritarian stylesof cceputerised education in other areas.
Our logic tutoring program, called EMIL, hasseveral advantages over other programs of theeau* kind. First, there are a large number oflogic textbooks, each with its own version of therule of logic. EMIL is the only program thatlets teachers supply the program with the set ofrules to be used with their textbook, instead of
forcing them to use the book which goes with aset of rules written into the program. Second,E MIL is extremely gentle with the student's input,and generally repairs typing mistakes rather thancomplaining about them. This le important becauseour students ere, for the most part, unfamiliarboth with the typewriter keyboard and the notation
Computers in Humanistic Studies 43
of logic. Third, the program lets students tamlines st the bottom of the proof Which they hopeto derive later so that they can work the proof
backwards if they like. The reason we allow, andin fact encourage working backwards, is that ef-fective proof-finding strategies require an ana-lysis not only of the formulas already derived,but of the formula to be proven as yell. Often theproof - finding problem can be considerably simpli-
fied by using the goal formula as a guide-post fordetermining what the steps previous to it oust
look like in the completed proof. Our programallows students to record the results of usingsuch strategies right it the terminal, instead ofsubmitting a polished product to the computer. Thefourth advantage of our program is the main topicof this paper. Since September of 1979 EMIL has
been giving students good advice about how tosolve problems that they ere unable to do on their
own. In this way it is providing a good portionof whet can be offered by s human logic tutor.
There are three main strategies for designingcomputer program that can offer advice on proof
finding. The first is simply to store a completedversion of the proof that the student is workingon and to store a list of comments that ere in-tended to help a student who asks for aid on com-puting a particular line. If the calmest the stndent :eceives proves isahelpful, the student canoak to see the next line of the stored proof, orindeed any number of lines, up to and includingthe entire proof. -
This hint strategy requires that a completedproof must be stored in the computer, along withappropriate comments, for every problem studentswill work on. It also presupposes that there islikely to be only one reasonable sequence of stepsthat leads to the conclusion. If studentsp-proach a problem in an unusual way, there may notbe enough similarity between their proof and thestored proof for the computer to be of any help.finally, it presupposes a top -to- bottom pattern of
proof construction. But frequently the very nextJeeps in a proof will not reveal a strategy thatLeeds to success; suck strategies must rather beexplained with reference to what happens muchlater in the proof. This sort of hint routinedoes not help students appreciate the global str-ategies which require knowledge not just of whetsthe proof has been, but also of *ere it Lavine,and these are generally the most useful strategies.
Mother technique is to write a program thatallows the computer to generate a solution to thestudent's problem and to recognise standard *Stil-
ettoes during the course of that solution. Thisstrategy eliminates the need for storing a proof,with commentary, for each problem to be attempted,since the computer generates its own solutions.But this strategy runs the risk of generatingstrange proofs, which students are unlikely to to-capitulate. Also, each formulation of the rulesof logic will require its own custom-tailored pro-gram for generating proofs. Furthermore, the pro-gram to generate comments must be very carefully
5.4
Os.-r
44 NECC 1960
written to avoid misleading advice. Worst of all,
this strategy still does not help students to seeglobal strategies; like the stored-proof strategy,this strategy uses a top-to-bottoet approach toproof construction and confines itself to givingadvice about the very nett line of the proof.
Another difficulty with both of these approach-es to the design for an advice-giver is that theprograms doss not attempt to construct advice onthe basis of whatezer progress the student mayhave already made on the ;nobles. The failure to
build advice *round the students partial successestends to discourage invention of novel, yet pron.loins, partial solutions, to devalue the students'own creat...te abilities, and to lower their self-confidence. It dampens the students' engagementIn the problem-solving process while reinforcingthem for stereotyped solutions.
The third approach to the design for an advice-giver, the one we have adopted, overcomes theseproblems by paying ..are attention to the techni-ques actually used by human logic tutors. One ofthe main things a human tutor should do is prov-ide students with effective problem solving toolsfor analyzing the situation they are in and forbreaking the problem into simpler subproblems towhich the same tools can be applied all overagain. An effective tutor does not give thesolution, or evenrpiedet-of-ft; bhi-tiotead-pro-vides an apprenticeship in 'AO art of asking therelevant questions, the r avers of which willlead the student to sae ow the problem can bebroken down into more enageable parts. Questionslike "Can you apply "-Us rule to lines you havealready derived?" lard "What rule could be used toderive s formul.. 4.4 this shape?" when presented
in a coherent sequence are very effective in help-ing students .sevelop strategies which they canlearn rt. vs. effectively in a wide variety ofproof - fading problems.
"e actual program that we use to give advice.us writs... by me in about four hours. The lm-placentation was ao easy because the advice pro-gram does little more than ask students a lead-ing question and thee branch to s new questionon the basis of their answer. Eventually, theprogram tuns out of questions to ask, and sped.fie advice is given the basis of the informa-tion prov44 ,- in the st4ant's answers to thequestions. se questicao can be thought of asbeing structured in a Use, with the path alongthe branches being determined by the students'armors, and the advice for each situation be-lug located st the tip of each branch.
Since programming our advice-giver was so
simple, the main focus of our attention hasbeen on the creation of a file of questionswhich have real pedagogical merit. Since thequestions axe not written into the structure ofour program, modifying the question tree in re-sponse to 4at we learn about effective advicehas been a painless process which dose not re-quire any programming expertise.
Our question file bee a very simple format.
(See Figure I.) Each record contains the text ofa question followed by a lint of accepted answers,each followed by a number which indicates whichrecord to Sump to in case the student responds withthat answer. The last item in each record beginsbegins with a "*" (which indicates that there.areno more accepted answers) and contains text whichis printed in case the student does not respondwith one of the accepted answers. Most of thequestions we ask are answered with "yes" or "no",
but we found the use of other sorts of answersmore conVe2:ent for certain questions. The textof the advice to he given is simply stored in thequestion file foil. ad by "*." This "*" indicatesto the program that this "question" has no accept-ed answers, and ro the program should stop the ad-vice giving process after printing it.
We have built a number of improvements intothis simple program. First, the sequence of thequestions should vary depending on how such thestudent has learned and how difficult the problemis. Our first solution to this problem has beento assign each problem that the student is us. It
on a level number and to use this number to roavthe question-asking program to separate quests...trees for each level we have defined
The second enhancement is motivated by the factthat we want to mention items in our questions,that change during the execution of the program,for example, the last line number the student hasfinished In the proof, or the name of the rulethat he intends to work with. Obviously the textof the questions in the file cannot mention spe-cific line numbers or rule names. Our solutionis to introduce variables, or fillins, to ourquestion text,
FIGURE I: SAMPLE RECORDS FROM A QUESTION FILE1. 'CAN YOU APPLY MP TO ANY PROVEN LINES' '7' 2
'N' 3' *ANSWER YES OR2. 'APPLY MP TO THESE LINEC"*.3. 'WHAT IS THE NAM CONNECTIVE or YOUR GOAL
FORMULA?' '4' 4 'V' S t->1 6 **PLEASE MOWER4, V OR 41
that are then replaced with the cottespondingspecific information Suet before the question isprinted at the terminal. We have adopted a con-vention that words beginning with "4" are vari-ables, and so, for example, a line of advice onour question file might read like this: 'OUSHOULD APPLY 4kULE TO LINE 6GNUM'. This directs
the program to fill iz the specific informationabout the rule name and line number so that thestudent sees, for example, YOU SHOULD APPLY GOUTTO LINE S.
It may surprise you to learn that although ouradvice-giving program was running with these twoenhancements in September 1979, we wereworking on a central portion of the advt. ing
program in January 1980. That was because ea stillhad to program the most important improvements thedevelopment of subroutines which can answer allthe questions which are posed to students by theadvice-giver, and comment on any errors in thestudents' responses. Though students are gamer-
ally quite accurate in their responses to ques-'tions posed by our advice-giver, they occasionallymake eine:Nuts that can result in their receiving
bad advice. But informing students of their er-rors is not the only reason for giving the comput-er the ability to monitor the correctness of thestudents' responses. Once students run the advice-giver a number of times, they become bored at hav-ing to answer a number of pointless questions. Thequestions become pointless not because they aren'tneeded in analysing proof - finding priblems in gen-eral, but because the Student is r- l aware thata particular portion of the me' is not neededfor the problem being dealt with aen the com-puter is capable of answering qua time itself, wecan decide which questions, at mi. :h levels ofdifficulty, should be printed at the terminal, andwhich we should let the computer answer for itselfby examining the proof the student is working on.Experienced students may resent being asked anyquestions at all and nay prefer an advice-giverthat merely prints a specific piece of advice.However, we believe that for met students whoneed the advice-giver in the first place, posingthe relevant sequence of questions is such morevaluable to their learning problem-solving skills
than is their obtaining advice.At this writing we have a version of EMIL that
answers for itself all the'questioes we pose saveone, and we have a method of indicating in ourquestion file which questions are to be askedunder which circumstances. He *till need to do alot more research on how obtrusive the advice-giver ought to be as a function of the students'progress and cognitive style. However, the mainadvantage of our program is that we have completeflexibility over the circumstances under which theprogram types out the questions.
There is a final reason for programming thecomputer so that it can answer all the questions.When this is done the program can traverse thequestion tree on its own and come up with therelevant advice. Once advice is available, theprograms can follow it to construct proofs on itsown. Judging from extensive paper and penciltests, our advice tree turns out to be highlyeffective (though not totally effective) in solv-ing logic proof-finding problems. As a result,it is capable of solving for itself the vastmajority of problems we give our studs s. Thisprovides us with an 'sporran tool for .mprovingour program. By running a large number of prob-lems through our advice-giver, we can determinethe circumstanced under which it is unable to doa proof, and than use that information to crestse more sophist -.1ted version of our questiondata files.
Our approach to giving computerised advicehas a wide range of applications. It can beused, for example, to help college studentswith their physics homework, or with trackingdown the identity of unknowns in qualitativechemistry, to help medical students learn diag-nosis, or even to help people to determine what is
Computers in Humanistic Studies 45
wrong with their car, or whether to itemize theirdeductions. All it takes is a simple program torun the questions and a question file that iscarefully constructed to reflect the best strate-gies that people actually use to solve the kindsof problems which are at issue. Depending on thecontext of its use, sone or all of the enhance-ments to the basic program that we have developedcould be used.
It is worth pointing out exactly how our ad-vice-giving progra differs from the standardapproach to CAI using the multiple-choice format.The differences are not particularly striking fromthe programmer's point of view. In both cases,the program is designed to ask questions and toselect new questions on the basis of the students'mowers. The advice-giving programs generally re-quire a more elaborate branching structure, andthey may differ in being unable to evaluate thestudents' response. But the important differencesare the ones that are obvious to the educator, forthey have to do with the educational purposes ofthe program.
Multiple choice CAI attempts to get the studentto memorise the correct answers to a certain kindof question. The stress is almost entirely on en-suring that the student lutes certain facts. Inthe case of advice-giving programs, the answersare not pert of whet is being taught. If any-thing, it is the questions we would like the stu-dent to master. By exposing students, over andover again, to a sequence of questions that have
been proven effective in problem analysis, thestudent learns to develop efficient strategiesthat can be used over a wide range of problems ofa similar kind. Furthermore, the whole processof learning to adopt principles of problem analy-sis and decomposition is a valuable exercise ofproblem solving skills that can be applied tovirtually any domain where creative thinking isrequired.
Although advice-giving programs may not lookvery different from multiple choice courseware tott% programmer, they have radically different edu-cational goals, the most important of which is thedevelopment of problem solving ability. Given thesimplicity of the programming effort as comparedto games and simulations, advice-giving programsare particularly attractive for any educator in-terested in developing the student's creativity.
46 NECC 1980
Ir
FROM A THEORY OF READING TO PRACTICEVIA THE COMPUTER
Dale M. Johnson*Center for Educational Research & Evaluation
The University of TulsaTulsa, Oklahoma 74104
R. Scott BaldwinUniversity of MiamiCoral Gables, Florida
INTRODUCTIONThe purpose of the present paper is to
describe a project which uses the uniqueand high-speed capabilities of the com-puter to match adolescents with bookswhich the students are capable of read-ing and which they enjoy reading. Theultimate mission of the project is toincrease the amount students read, whichin turn would enable them to become__better readers. -Enhanced reading abilityand attitude would then motivate studentsto read mores thus, the probability of acycle of life-long reading habits for theindividual would be increased (Mathewson,1976).
Traditionally, reading has been viewedas a hierarchical skills arrangement.This conceptualization has been reflectedin reading instructional programs whichincorporate highly structured approachesto word recognition and comprehensionskills. The underlying assumption hasbeen that students who are thoroughlygrounded in word attack skills, phonics,sight vocabulary, word analysis, spelling,and other reading subskills would usethese skills in reading, thereby becomingable readers.
Computer assistance in building read-ing subskills is not unusual. Illustra-tive projects in CMI include the Wiscon-sin System of Instructional Management(WIS-SIM), the Instructional ManagementSystem (IMS), the Interactive TrainingSystem (ITS), and the Stanford Project(Splittgerber, 1979). Numerous CAI pro-jects in reading have been implementedsince the mid-1960s that capitalize onboth drill and practice and tutorialmodes (Atkinson, 1968; Atkinson, Fletcher,Chetin, & Stauffer, 1970: Atkinson & Paul-son, 1970; Madachy & Miller, 1976: andMorrison, 1979). Computers have alsobeen used to establish the "readability*levels of print material (Barry, 1979;
Barry & Stevenson, 1975: Fang, 1968; Mar-ten, 1978, and Walker & Boillot, 1979).
Recently, reading specialists haveposited that fluent reading develops as anintegrated process rather than as a loosecollection of reading subskills. Fluentreading entails much more than the sum ofspecific skills. Such skills provide anecessary but insufficient foundation forthe-development of nature reading-(Goodaman, 1976: Smith, 1971). Profound in itsimplications, the basic idea reduces to asimple formula--young people need to prac-tice reading in order to become good read-ers (Allington, 1977: Daniels, 1971: FaderMcNeil, 1968; and Squire, 1973). How-
ever, students who are in most need ofreading practice tend to read the least.The results are that the reading deficitaccrues with age, ultimately resulting insecondary students who can neither read ona level congruent with their abilities norenjoy reading.
RATIONALEThe basis for the project described in
this paper is that students do not readpartly because they fail to find readingmaterials which ar- comprehensible andinteresting to them. However, a varietyof books do exist that span the abilitiesand interests of virtually all students.Therefore, a sound argument can be ad-vanced that the problem is not one ofavailability but one of accessibility.
On the basis of current trends in read-ing, it could be reasonably hypothesizedthat if various student characteristicsand preferences could be matched withcorresponaing book characteristics, theprobability of the student increasing hisor her interaction with print would beenhanced. Further, the increased readingwould result in a better likelihood of thestudent becoming an improved reader anddeveloping more favorable reading habits
57
and attitudes. Therefore, a major taskbecomes one of maximizing the match be-tween student characteristics and perti-nent print characteristics 3v. order toselect a book that the student can readand will enjoy. Gilliland (1972) pre-sents this task as follows:
Readability is primarily concernedwith a basic problem familiar toall people who choose books fortheir own use, or who choose booksfor others to use. This is theproblem of matching. On the onehand, there is a collection ofindividuals with given interestsand reading skills. On the otherhand, there is a range of booksand other reading materials,differing widely in content,style, and complexity. The ex-tent to which the books can beread with profit will be deter-mined largely by the way inwhich the two sides are matched...(p. 12).The time-consuming task of matching
students and books has typically reliedon published bibliographies; -human recall,and trial and error. Such techniques areobviously limited. Finding books whichcontain all or most of a desired set ofcharacteristics is a complicated task, infact, it is virtually impossible if theset of characteristics is large and theresources for finding them are limitedto human intervention techniques. Simpsonand Soares (1965), Jongsma (1972), andStanek (1975) have found that librariansand teachers, regardless of their in-tentions, fail to accurately considerthe reading interests of young peoplewhen purchasing or recommending books.
Research over the past decade or soprovides clues as to what variables arcimportant for matching considerations.Robinson and Weintraub (1973) and Squire(1973) have confirmed the importance ofreading interests of students whenselecting reading material. Further, thereader's age, grade level, socioeconomicbackground, sex, and cognitive readingability will tend to influence readingpreferences (Ashly, 1970; Carlsen, 1967:Jungeblut & Coleman, 1965; Hansen, 1973;Scharf, 1973: and Soares & Simpson, 1967).Book characteristics which tend to beinfluential include amounts of dialogue,concreteness of language, type of narra-tion, and the degree of action and con-flict provided (Carlsen, 1967; Jungeblut& Coleman, 1965; and Simpson & Soares,1965). The reader's age, grade, sex,ethnic background, personal history,hobbies, and the book's length, theme,
Computers in Humanistic Studies 47
linguistic readability, etc., will com-bine in a predictable manner to determinejust how interesting a particular bookwill be to a particular reader (Imam &Patyk, 1967; Smith & Johnson, 1972).Thus, it becomes clear that the qualityof the match between a reader and a bookis a function of numerous personal traitsand textual characteristics: however, theliterature does provide a comprehensivelist of the most important variables.
THE BOOKMATCH SYSTEMIn keeping with the goal of getting the
most suitable books to students, the mostpertinent variables were incorporated in-to algorithms which were subsequentlycoded for computer processing. First,individual student preferences regardingprint characteristics are considered.Figure 1 shows the variables -on whichstudents can express their personalsentiments.
Three of the characteristics relate tothe characters portrayed in the book, onerelates to the setting or location of thestory, and then 08_tvical interest areasare considered. With respect to the "in-terest areas," the student simplycated how well he or she likes each par-ticular area from he list of 88 topicson a three choice response format codedas follows: (Y) Yes, very much, (S)Sometimes. It's OK., (N) No, I don't. Ofthe OS interest areas, 20 are human dramathemes (ecology, drugs, personal beliefs,loneliness, etc.).
Corresponding to each connecting lineshow in Figure 1, an algorithm was devel-oped and coded into a FORTRAN subroutinefor a main program. However, Figure 2displays additional algorithms for refin-ing the match between students and books.Figure 2 shows three student traits thatrepresent somewhat different constructsfrom the student variables (preferences)shown in Figure 1.
The student traits which are linked tovarious book variables are reading abil-ity, attitude toward reading, and gradelevel. For example, three algorithms usea measure of student reading ability to-gether with: (1) linguistic difficultyof the book, (2) when the story appealbegins, and (3) the length of the book.Conceptually, better readers can morereadily comprehend more difficult text,they can comfortably tolerate more intro-ductory background prior to the beginningof the story appeal, and they will tend tocope better with longer books than poorerreaders.
Students' attitudes toward reading arealso used (Figure 2) to help discriminate
5V
48 NECC 1980
among books. The algorithms use a meas-ured attitude score based on the follow-ing heuristics. Students having betterattitudes toward reading, as opposed tothose with poor attitudes, will be ableto cope with longer books, will not re-quire early story appeal, and will relyless on physical action as a prerequisitefor enjoying a book. Finally, the gradelevel, since it is associated with chron-ological age and maturity, is also com-pared to the linguistic difficulty (whichis measured in grade equivalents) of thebooks and to the length of the books.Generally, students in lower grades willnot enjoy longer and/or acre difficultbooks.
A FORTRAN main program was written forthe Xerox Sigma 6 system which includessubroutines incorporating the schemesshown in Figure 1 and Figure 2. The func-tion of the program is to receive individ-ual student variables as input and matchthem with each of the approximately 5,000books in the data base. Each variable ofeach book is compared with the respectivestudent variables and differentialweights are assigned according to the de-gree of match and the relative importanceof the variable. The weights are con-verted to points and summed, resulting ineach book being scored for each individualstudent. The books with the highest num-ber of total points are listed as a per-sonalized ,bibliography for each student.
The functional operation of the match-ing process (referred to as Bookmatch) isshown in Figure 3. As can be seen, thefirst function is to extract a subset ofbooks which are accessible to the studentfrom the book data base. In most cases,this subset is the school's library hold-ings which have been furnished by thelibrarian. Second, each variable of eachbook is compared with either the student'spreferences and/or traits. The algorithmsaccumulate a point total (score) for eachbook. Finally, the books receiving thehighest score (most accumulated points)are printed. This output includes theauthor(s), title, type of book (fiction,nonfiction, biography, etc.), and a briefannotation of the book.
Each student receives a printed listof his or aer own personalized bibliog-raphy. The bibliography includes the top20 books and must contain some nonfictionand biographies along with fiction.Summary reports for teachers, librarians,and administrators are also provided.
DATA BASEOf significance to the success of the
system is the data base which contains
information on books written for earlyadolescent readers (grades 5 - 9). Newvolumes are periodically being added tothe data base which presently containsabout 5,000 volumes established over athree-year period. Initially, pertinentbook variables were identified through areview of literature in the field of read-ing. The original pool of variables wasverified by a committee of public schoolreading teachers and university readingspecialists. Two of the original varia-bles were subsequently omitted from thelist due to variations in printing stylesacross separate editions of many of thebooks. These two variables were "numberof pictures" (or non-print figures) and"size of type."
Second, a master list containing thetitles of books to be included was com-piled from published recommendations fromwidely recognized sources including theAmerican Library Association, Library ofCongress, School Library Journal, AmericanGuidance Services, International ReadingAssociation, and the National Council ofTeachers of English.
Reviewers were tutined to evaluatecharacteristics of books and to prepareannotations. These reviewers consisted ofreading professionals witn masters degreesand with either adolescent reading orlibrary experience. The book evaluationswere checked for reliability and werefurther validated by a reading specialist.The annotations were refined by a profes-sional editor. Book data were then codedand stored for computer match of bookcharacteristics with student traits andpreferences.
SUMMARYBased.on the notion that practice in
reading will be facilitated when readersand appropriate literature are broughttogether, a project called Bookmatch ispresently being implemented. In order tomatch reader characteristics an' pre-ferences with textual characteristics ofbooks, a medium with capabilities for vastdata storage, retrieval, and rapid arith-metic comparison operations was required.The computer was the logical choice tohandle the task. During the Bookmatchprocessing, books are scored according tothe degree to which they favorably compareto student preferences and characteristicsOn the basic of these scores, the computerprovides a printout with a personalizedannotated bibliography containing the mostcomprehensible and interesting books thatare accessible to the individual student.
59
Computers in Humenistic Studies 49
BOOK VARIABLES STUDENT VARIABLES(PREFERENCES)
(SEX OP MAIN )1(
CHARACTER(S)
ETHNICITY OP
IN CHARA
)1(
AGE OF MAIN
CHARACTERS)
SETTING OF
THE STORY
INTEREST
AREAS OP2VE BOOK
4SEX
PETERENC)
ETHNIC
REFEREN
)(AGE
REPERENC)
(SETTING4)
PREFER=
Figure 1. Schematic fpr matching algorithms of studentpreferences and book variables.
6Q
50 NECC 1000
.
----lit
SEQUENCE or (INDIRECT)
STORY
(GRADE )4LEVEL
CHARACTERISTICS
.( BOOK
CHARACTERISTI S
PHYSICAL
LENGTH OF
BOOK
ATTITUDE AMOUNT ofPHYSICAL
READING ACTION
Figure 2. Schematic network for matching algorithms ofstudent traits (characteristics) and bookvariables.
61
.--IVARIABLES
INPUTSTUNS?
4
newLunn
sown=
STORELIBRARY'SMOLDINGS
MATCHARIFIRENOSS(sm. 1)
INPUTXa BOOR
4
CUMULUSBOOK TOTALSCORE
i(
SELY.CT
LIBRARYHOLDINGS
Computers in Humanistic Studies 51
BOOKDATABASH
SORT BOOKSBY POINTS(SCORES)
OUTPUT'
INGS ANDBEST BOOMS
(ANOTHER sTunetakftroP OR tisX'?STUDENT ./
Figure 3. Functional Schematic for BOOM= operation.
52 NECC 1980
REFERENCESAllington, R.L. "If They Don't Read Much,
How They Ever Gonna Get Good?" Journalof Reading, 1977, 21, 57-61.
tanfor Project. TechnicaTW47513Ft131.3117ard, Calif.: 'natl.. 'orMathematical Sciences in the ScSciences, 1970.
Atkinson, R.C. & Paulson, J.A. AnA roach to the Psychology of Instruc-t on. TcahlEil Report 157. 3TIEE7A,Calif.: Institute for MathematicalStudies in the Social Sciences, 1970.
Barry; a:c.- "Computerized ReadabilityLevels--Their Need and Use." Journalof Educational Data Processing770757
Barry, J.C. & Stevenson, T. "Using aComputer to Calculate the Dale -ChallFormula." Journal of Reading, 1975,19, 218-22.
Carlsen, G.R. Books and the Teen-a eReader. New YENT-Iirper & Row, 967.
Daniels, S. How 2 Gerbils 20 Goldfish200 Games IINTIFhemHew E12-1.7a. .iihTIORATaT The War--ail-stet Press, 1971.
Emans, R. & Patyk, G. "Who Do High SchoolStudents Read?" Journal of Reading,1967, 10, 300-4.
Fader, D.N. & McNeil. Hooked on Books:Pr ram and Proof. Nis7-173Fks G.P.Putnam 11-13nr-f568.
Pang, I.E. "By Computer: ?leach's Read-ing Ease Score and a Syllable Counter."Behavioral Science, 1968, 13., 249-51.
Gilliland, J. Readability. London:University of London Press, 1972.
Goodman, X.S. "Reading: A Psycholin-guistic Guessing Game." In Singer andR.B. Ruddell (eds.). TheoreticalModels and Processes of Reading.
Newark, Delaware: International Read-ing Association, 1976.
Hansen, H.S. "The Home Literary Environ-ment--A Follow -up Report." ElementaryEnglish, 1973, 50, 978.
Jongsma, B.A. "The Difficulty of Child-ren's Books: Librarians JudgmentsVersus Formula Estimates." ElementaryEnglish, 1972, 49, 20-6.
Jungeblut, A. & Coleman, J.H. "ReadingContent That Interests Seventh, Eighth,and Ninth Grade Students." The Journalof Educational Research, 196758, 392-T51.
!Carton, H.A. " G M Program Rates Level ofText Difficulty." Computerworld, 1978,(May), 9.
Madachy, J. & Miller D.J. The Use of CAIin the Lan ua e Pr ram at GaTigunt.Santa Bar ara: Assoc atran73EaiDevelopment of Computer -Based Instruc-tional Systems, January, 1976.
Mathewson, G.C. "The Function of Attitudein the Reading Process." In H. Singerand R.B. Ruddell (eds.), TheoreticalModels and Processes of Reading (2ndM=Newarrirei,Dwares InternationalReading Association, 1976, 665-76.
Morrison, F. TICCIT. Dynamic Phoenix,1976 (July), 45-6
Robinson, H.M. & Weintzaub, S. "Research'Related to Children's Interests and toDevelopmental Values of Reading."Library Trends, 1973, 22, 81-108.
Scharf, A.G. "Who Likes What in HighSchool." Journal of Reading, 1973, 16,604-7.
Simpson, R.H. & Soares, A. "Best-and-Least-Liked Short Stories in Junior HighSchool." English Journal, 1965, 54,108-11.
Smith, F. Understanding Readin . NewYork: Holt, Rinehart an W nston, Inc.,1971.
Smith, J.R. & Johnson F.D. "The Popular-ity of Children's Fiction as a Functionof Reading Ease and Related Factors."The Journal of Educational Research,1727437-7977467-----7--
Soares, A.T. & Simpson, R.H. "Interest inRecreational Reading in Junior High
School Students." Journal of Readig,1967, 11, 14-21.
Splittgerber, P.L. *Computer-Based In-struction: A Resolution in the Making?*Educational Technology, 1979, 19, 20-5.
Squire, J.R. "What Does Research Rev;a1About Attitudes Toward Reading?" InR.A. Meade and R.C. Smith (eds.),Literature for Adolescents: Selectionan -thiC -Columbus, Ohio: Charles E.Brirra, 1973.
Stanek, L.W. *Real People, Real Books:About YA Readers." ToE of the News,1975, 31, 417-27.
Walker, N. & Boillot, M. °A ComputerizedReading Level Analysis." EducationalTechnology, 1979, 19, 47-9.
*NoteDale Johnson will present paper.
64
Computers in Humanistic Studies 53
54 NEW 1960
NON - HARMONY:A VITAL ELEMENT OF EAR-TRAINING IN MUSIC CAI
in ear-training has become an importantassistant to classroom teaching in thefirst two years of college-level musictheory. The student's aural skills (whichare needed to deal with the three basicelements of music -- harmony, melody, andrhythm) can be greatly reinforced with
-such individual instruction. Part of thesuccess of this reinforcement is thatwhile these elements are integral partsof a complex multi-dimensional subject,they can be separated from each other toa great extent, so that classroom teaching(and CAI lessons) can concentrate onvirtually one element at a time. Inparticular, this system allows the lessadvanced student to concentrate hisefforts upon developing his understandingof one broad subject at a time. As hisskills and confidence grow, he is bettercapable of identifying these elements invarious combinations, and eventually ofunderstanding them in their musicalcontext.
As soon as the student has developedsome skill in each of these three areasand has begun to understand the integratedwhole, it is necessary to introduce theconcept of "non-harmony." As the nameimplies, this element lies outside therealm of the established harmonic strucTtare but relies on melodic and rhythmiccontext to differentiate among the variousforms that it takes. Despite its rathernegative title, non-harmony is a verypositive part of music an aspect ofthe concept of variety which helps todelineate style. In practice, non-harmony takes the form of short melodicand rhythmic patterns which augment theharmonic context. These patterns fallinto approximately ten categories forwhich terminology is fairly standard.Each term refers to a type of non-harmonictone which defines its own special pattern
and context. Although each pattern israther specific, melodic and rhythmiccontext can vary somewhat and harmoniccontext can vary gz:atly. For thisreason, the necessity of drill and prac-tice of such variance presents a logis-tics problem in the classroom but isgreatly helped by random access and tran-position available with CAI.__
METHODThe ear-training CAI system at North
Texas State University employs the Auto-matic Music System (ANUS) designed byProfessor Dan W. Scott of the ComputerSciences Department. The ANUS systemconsists of a Motorola 6800 microprocessorwith a CRT terminal and specialized musichardware (MUSOR). The microprocessortranslates the musical score and accessesthe MUSOR, which is the actual sound-producing apparatus engineered byProfessor Scott. The ANUS system isconnected to an BP 2000-F Timeshared BASICcomputer, which provides additional filestorage capabilities. This system uses aflexible score language for notation inputand offers a variety of timbre and articu-latzon possibilities. Of primary impor-tance is the facility of rapid playbackwhich is a substantial advantage overother computer-based music systems in itscost range.
The non-harmonic tone lessons wereconstructed using a simple melodic andrhythmic four-voice chorale style. Thissimplicity was not due to any systemlimitations, as the system is capable ofrapid and complex articulations of rhythm,a wide range in melody, and up to sevensimultaneous voices for harmony. Rather,the simplicity,was'maintained to limitthe number of variables of elements otherthan those being tested and to give afairly normal musical context. For eachexample heard, the basic context was a
65
five - to seven-chord progression of fourvoices in quarter-notes. To this strictlyharmonic setting, one or more non harmonictones were added in eighth-note motion.Since there were ten categories to becovered, the following factors determinedthe order in which the terms were pre-sented. Group I consisted of the fivenon-harmonics that normally occur off thebeat or in a rhythmically weak positionas compared to the note of resolutionwhich comes directly after. Group IIconsisted of the four non-harmonics thatoccur on the beat or in a rhythmicallystrong position, and usually displace ordelay a note of the harmony. The lastcategory normally has no such rhythmicdefinition. Due to the rhythmic contexts,Group I is considered easier to hear thanGroup II. Also, in actual practice inmusic literature, the patterns in Group Igenerally occur more often than those inGroup II.
Within Group I, the five unaccentednon-harmonics (with their standard abbre-viations) were presented in this order:1. _unaccented passing tone (OPT)2. neighboring tone, upper and lower
LN) also sometimes called"auxiliaries"
3. anticipation (ANT)4. escape tone or echapee (ET)S. changing tones (CT)
Within Group I/ the four accented non-harmonics were given this order:
The ir'arnal order of the two groups waschosen to present the most-often heardcategories first, in ordtr tarstmrtthestudent with the most familiar patterns.Theonly exception to this practice occurredwith categories 8 and 9. The less-oftenheard retardation (18) was placed beforethe appoggiatura (19) in order to presentit directly after the suspension, on whichit is patterned. Also this order wasselected because Jummary lessons are usedin various places between individual pre-sentations, and a t meaty to show simi-laritieu and differences between the sus-pension and retardation would logicallycome before proceeding to the next (andless - related) non-harmonic.
The general need for both immediateand cumulative summaries of such a largenumber of elements resulted in seven suchlessons being inserted into the sequencealready discussed. While each of the tensingle-element lessons seemed well served
Computers in Humanistic Studies 55
by the examples (which were carefullywritten to provide maximum exposure tothe variations encountered in music), thesummary lessons gradually grew in numberof examples since they demonstrate avariety of combinations of these elements.Of course, this variety is furtherexpanded by random selection and trans-position.
A standard format for the student'sanswer was kept for all of the lessons.Because of the complex situation presentedby non-harmony, the student's responsecould not be reduced below three elements:the rhythmic, melodic, and non-harmoniccontexts. (This sequence best repre-sented the student's understanding of thefactors involved, as determined by theauthor's experience in teaching.)
For the first element, the studenttypes a number from one to six, locatingthe beat with which the non-harmonic isassociated. (Although many examples areseven chords long, a basic precept of non-harmony is that it must finally resolveinto harmony. Therefore, non-harmonycould not,be.associated with the seventhbeat.) The second element of the stu-dent's answer is a capital "S" or "8*which indicates that the non-harmonicoccurred in the soprano or bass melodicvoice. (The choice was limited to theupper and lower voices to afford a measureof variety while avoiding the unnecessarycomplication of determining locations ininside voices when these voices can beextremely close.) The third and lastelement of the student's answer is theabbreviation that identifies the non-harmonic itself from among the ten pos-sible labels. Even though this part ofthe answer would be obvious in any of thesingle-element lessons, the act of typingthe abbreviation reinforces the student'saural impression and, establishes a patternof response which does not need to bealtered for the summary lessons. Allelements of the answer were separated bycommas, even when the example involvedmore than one non - harmonic response.
One other factor, a chord tone, wasadded to a few random examples withineach lesson. A chord tone may resemblea non-harmonic tone in rhythmic and melo-dic placement, but it is part of the har-mony (as opposed to non-harmony), and itspattern of usage does not necessarilyparallel that of a non - harmonic tone.This factor was added to insure that thestudent would truly listen for non-harmonyand not just attempt to label anythingthat moved! These chord tones do notrequire any response from the studentother than to understand that they were
CC
56 NECC 1980
notes which were not part of the non-harmonic answer.
RESULTSIn final sequence, there are seven-
teen lessons, delineated as follows:Nl. UPT - Unaccented passing tonesoccur most often of any of the non-harmonics and are located equallywell in soprano or bass.N2. UN, LN - Upper and lower neigh-bors occur mainly in soprano, asreflected Immthe examples.N3. ANT - Anticipations occur almostexclusively in the soprano andusually very close to the end ofthe progression.N4. Immediate summary of Lessons 1,2, and 3 (the most-often heardunaccented non-harmonics).N5. ET - Escape tones occur mainlyin the soprano voice.N6. CT - Changing tones involve adefinitive four-note pattern andoccur only occasionally in the bass.N7. Immediate summary of Lessons5 and 6 (somewhat similar patterns).N8. Cumulative summary of all unac-cented non-harmonics (Lessons 1through 7).N9. APT - Accented passing tones,like unaccented ones, occur in bothsoprano and bass.N10. Summary of APT and Urf (Lessons1 and 9). These two patterns occurquite often and are alike exceptfor rhythmic placement.
S - Suspensions occur in bothvoices in many cases.N12. RET - Retardations parallelsuspensions in every way exceptthat they resolve up instead of .
down.N13. Immediate summary of the ac-cented non-harmonics covered thusfar (Lessons 9, 11, and 12).N14. APP Appoggiaturas may haysslightly varied melodic patternsand most often occur in the soprano.M15. Cumulative summary of all ac-cented non-harmonics (Lessons 9,11, 12, 13, and 14).N16. Tap - Pedal tones are the leastoften used, but are the most obviousnon-harmonics to identify, occurringmainly in the bass.N17. Final cumulative summary of allof the non-harmonics (This lessoncontains the most examples).
SAMPLEThe following example shows Lesson Nl
(UP Lth comments on specific features.Cs L, 1/13/80, JCGT & ATC
C: CHORD PROGRESSIONS WITH UNACCENTEDC: PASSINGATONESH: FIRST ANSWER IS THE NUMBER OF THEHI CHORD AFTER WHICH THE NON-HARMONICH: TONE:SOUNDS.H: SECOND ANSWER IS "S" OR "B" FORH: SOPRANO OR BASS VOICE WHERE THE NON -H: HARMONIC TONE OCCURS.H s THIRD ANSWER IS THE ABBREVIATION (UPT)H: FOR THE NON-HARMONIC.Hs TYPE LETTERS IN CAPITALS (WITH SHIFTH s KEY).H: (SOME MOVING EIGHTH-NOTES MAY BE CHORDHs TONES.)The *H" statements are "hints" which thestudent may branch back to at any time toobtain reminders about valid answers andformat.T: $N, HERE'S A LESSON TO HELP YOU LEARNT: NON-HARMONIC TONES.Ts THESE WILL BE UNACCENTED PASSING-TONES.P: 2The "T" statements are texts which thestudent sees, and the "$N" addresses thestudent by his own name. The HP: 2" isa two-second pause (with the text still onthe screen) to allow the student to absorbthe information.CS:Ss 2Ts I'LL PLAY 5 TO 7 CHORDS.T: YOU TYPE: WHICH CHORD (1,2,3,4,5 OR 0T: 6) THE NON-HARMONIC OCCURS AFTER,Ts WHETHER THE NON-HARMONIC IS IN SOPRANO
OR BASS (S OR B) IN CAPITALS.T: THE ABBREVIATION FOR THE NON-HARMONICTs (UPT) IN CAPITALS.Ts TYPE A COMMA (,) AFTER EACH PART OFTs YOUR ANSWER.Ts FOR INSTANCE: 3,B,UPTTs MEANS THE NON-HARMONIC OCCURRED AFTERTs THE THIRD CHORD, IN THE BASS VOICE,T: AND IT WAS AN UNACCENTED PASSING-Ts TONE.Ps 10The"CS" then clears the screen so that thefollowing text can appear as a stationarygroup without rotating off the screen."S: 2" double-spaces the text of addedvisual impact. The text contains specificinstruction on the anwer format and givesa hypothetical student response withexplanation. Again, there is a pause often seconds for comprehension.Ts HERE'S AN EXAMPLE (WITH ANSWER), $N --T: $R 4CD 3C 2E GETs 4EET: 4G 3BD 2D GQT: 4E 3CU 2G CUQTs 4F 3C 2F ADOT: 4G 3GD 20 BETs CUETs 4GD 38 2F DOTs 4CU 3G 2E CH
67
Ti ?P: 20Ti eS 813
T: YOUR ANSWER: 5,S,UPT -- RIGHT, $N ?pi 10The next section plays a sample examplefor the student (lines 145 - 205 processthe sound) and then gives the correctanswer with a pause for comprehension.Ti REMEMBER:Ti TYPE A "," BETWEEN THE PARTS OF YOURTi ANSWER.Ti TYPE A "7" TO RE-HEAR THE MUSIC.Ts TYPE " / /HINT" FOR A HINT ABOUTTi ANSWERS.Ti (YOU MAY WANTTO USE PAPER AND PENCILTs TO JOT DOWN ANSWERS BEFORE YOU TYPET: THEM IN.)P: 10The student then sees the final reminderson basic procedures for dealing with themusical examples.Ti LET'S GET STARTED, SR,Vi 123456SBUPT,0: 10CRNIAP4,453 ).The actual examples are then announcedto the student."V"' gives-the validcharacters for the student's answer. Forevery lesson the first eight charactersare the same; the numbers 1 through 6are the possible answers to the firstquestion of beat (rhythmic placement);and "s" and "B" are the possible answersfor voice (melodic) placement. The lastcharacters label the non-harmonic andrange from a single one- to three- lettergrouping (for the single-element lessons)to the complete list of all non-harmonicpossibilities for N17. In this samplelesson of Ni, only the single abbrevia-tion of VET is valid. Finally, "Q"formats the questions. In this casethere are 10 possible examples fromwhich the program makes a random selec-tion and a direct comparison. The stu-dent completes the lesson when heanswers four out of five times correctly,And he is allowed three tries,befobe thecorrect answer is given by the computer.
Ten examples then follow which aresimilar to the sample example of lines145 through 190.
DISCUSSIONThe seventeen lessons of this non-
harmony project are all presented inparallel format to establish a consistentpattern of thought and approach and tominimize confusion as the lessons grad-ually become more complex. Each elementis introduced in a lesson devoted solelyto its problems and contexts, and summarylessons gradually combine the elements ina logical order. Data on students'
Computers in Humanistic Studies 57
responses are kept and made available tothe classroom teacher for evaluation andadvising.
SUMMARY AND CONCLUSIONSMusic poses interesting instructional
problems because it shares with otherfields (for example, chemistry and math)a well-defined vocabulary for musicalelements. However, musical teaching mustinclude showing the relationships amongthese elements. Non-harmony is an essen-tial component of these relationships,and thus forms a model for the solutionto the problem of teaching.relationshipsin many fields.
SELECTED REFERENCESApel, Willi. The Harvard Dictionary of
Music. HarvainfaVerexty Press,Ca ridge, 1969.
Bales, W. Kenton and Joan C. Groom."AMUSs A Score Language for Comput-er-Assisted Applications in Music."Presented at the 1979 Associationfor Educational Data Systems Conven-tion3*Detroit, Michigii.
Hamilton, Richard L., and Dan W. Scott."A New Approach to Computer-AssistedInstruction in Music Theory."Presented at the 1977 Conference onComputers in the Undergraduate Cur-riculum, Denver, Colorado.
Hofstetter, Fred T. "Computer-BasedRecognition of Perceptual Patternsin Harmonic Dictation Exercises."Presented at the 1978 Associationfor the Development of ComputerInstruction Systems Conference,Dallas, Texas.
Killers, Rosemary N., W. Kenton Bales,Richard L. Hamilton and Dan W. Scott."ANUS: The Computer in MusicInstruction." Presented at the 1979Texas Music Educator's AssociationConference, Fortlorth, Texas.
Ray, Douglas, and Rosemary N. Killam."Melodic Perception Development andMeasurement Through CAI." Publishedin the proceedings of the 1979National Educational ComputingConference, Iowa City, Iowa.
68
Computer Literacy
A CASE FOR INFORMATION LITERACYDr. S. D. Schiaming
ISM Corporation10401 Pernwood RoadSatbeeda, MD 20034
301/897-2090
INTRODUCTIONAlthough digital computers have
been present in educational institu-tions for over 25 years, only a smallpercentage of students have been ex-posed to them. As the cost of com-puting continues to decline, it is be-coming financially feasible to intro-duce the "majority of students to com-puting and data processing.
Dr. Andrew R. Molnar (1) is an.ert1-cultte spokesman for the expanded inte-gration of the computer luto Americaneducation. Citing our shift from anindustrial society to a knowledge -basedsociety together with the internationalchallenge to our technological leader-ship, he makes a compelling case forcomputer literacy. Dr. 4elner is notalone to his views. The te:: "computerliteracy" is being increasingly usedand discussed at educational confer-ences and in the educational litera-ture.
The objective of Ott paper is toexplore the future content of computerliteracy education. Bose' upon anassessment of current computer /dataprocessing education, trends in tech-nology, ma a projection of the futureinforastioe environment, a set of threefuture skill categories is defined.
Computer /Data Processing Pro-,fessionals (2) - the computerexperts
. Inforaatioa System Profession-als (3,4) - the integrators ofthe computer and business pro-CerfeSThe End Users - the informationsystem customer
The first two categories are, ofcourse, key areas requiring educationalplanning and investment. However, from
a statistical point of view, they willrepresent s small minority of the popu-lation interacting with future informa-tion systems. it is the end user major-ity for which computer literacy isprobed in this paper. The result ofthis exercise is a recommendation forthe creation of a course entitled "AnIntroductory Course in informationLiteracy" for the majority of under-graduate students who are pursuing amajor other than computing.
TECHNOLOGY DIRECTIONSThe 1980s will conceivably see an
order of magnitude growth in computingpower. Responsible for tnis growthwill be the continued decline in castof electronic components together withthe elastic demand for data processingservices. The concept of the computerrill rhease as processing capability isdist.- uted into many devices. Today's'hobb t with a microprocessor at homecepa. .. of communicating with a distantlarge computer is a beginning exampleof distributed processing. One of thereasons for this trend is that thecost of computing has dropped below thecost of anuaicstion. Thus it becomesdesirabl ;o satisfy local data pro-cessing kegnirements at the sourcewhile communicating only that datarequired by other sites. Communicationcosts will also continue te, declineas newer forms of broadhead communi-cation, such as satellite transmission,become commonplace. Again, as in egecomputer industry, change will he drama-tic in the communications industry.There can be little doubt that informa-tion technology growth will continue tohe significant through the latter partof this century.
NOTE: The views expressed in this paper are those of the authorand do not necessarily represent the views of the internationalSnail:ass Machines Corporation.
58
69
The key implication with regard toeducational curricula is that not onlywill computers change, but the informa-tion envlrossent in which they operatewill change; the way data elements aremoved, stored, and used will be alteredas wall as the way they are processed.
An analogy to transportation aye-teas can be drawn *ere. Vehicles, road-ways, and parking facilities are compo-nents of a transportation system justas computers, communications, data, andapplications are components of an infer-nation system. Vehicles are notstudied in isolation by urban planners,and correspondingly computers shouldnot be studied in isolation from othercomponents of the information system.
CURRENT COMPUTER/DATA PROCESSINGEDUCATION
For the purpose of this discussion,it is convenient to think of threeclssses of students that correspond tothe skill categories presented in theintroduction.
. Computer Majors - tho t prepar-ing for careers in da,a process-ing.
. Information Systems Majors -those preparing for careers inmanagemenc with a major in infor-mation systems.
. Other - those with career objec-tives other than computing ordata processing . . . the futureend users.
Dr. J. Baubles's most recent inven-tory of computing in American education(5) provides insight into the relativedistribution of these classes: in ex-cess of 901 of the students fell intothe "other" category while 301 of theinstitutions offered courses in thefirst two cstegoriss. From these sta-tistics, it would sees the computermajor is receiving sore than his fairshare of reseurca$.
The status of information systems(IS) education is addressed in a recentarticle (6) reporting preliminary 'find-Legs of the ACM Curriculum Committeefor Information Systems. Prof. Jay F.Nuns-alter, chairman of the committee,
.id: "The IS field integrates eye-.efts analysts, statistics, managementscience, accounting, economics, fin-ance, marketing, production and compu-ter and communication teahnolo-gime . "The U.S. has nearly fivecomputer science departments for everyIS department according to a recentstudy not connected with the ACM Laves-
Computer Literacy 59
tigation. Nevertheless, the nation hasa much higher demand for personnel,such as IS graduates, who have a combin-ation of technical and organizationalskills than for computer scienc t. grad-uate with 'solely' technical skills."
With the previous discussion as abasis, certain conclusions can be maderegarding future educational require-ments of the three categories of stu-dents previously defined.
FUTURE EDUCATIONAL REQUIREMENTSVocational and Computer ScienceProfessionals - The demand forthis type of professional to-gether with the curriculumneeded to prepare him or herappears to be relatively wellunderstood and served by theeducational community.
. Information System Profes-sionals - As reflected by theNunamaker study, the informationsystem concept has been slowerto emerge. However, due to theefforts of the ACM (3,4) and thenation's business schools (42 ofthe 52 satisfactory IS under-graduate programs were compo-nents of business or managementcolleges) (6), the problem isunderstood and undoubtedly stepswill be taken to improve thequality and quantity of informa-tion system graduates.
. All Others (the future endusers) - It is this group forwhich current computer literacyappears to be inadequate. Theywill take their places in market-ing, production and administra-tion with a cursory knowledge ofthe computer but not the informa-tion system in which they func-tion. A hard-earned lesson inthe computer indubtry is that tomajor cause of information sys-tems fatless has been that theydidn't do what the users ex-pected The end users must rep-resent their needs as input tothe information system designprocess.
The following course descriptionaddresses the minimum elements of aninformation literacy offering.
AN INTRODUCTORY COURSE IN INFORMATIONLITERACY
There are four key components inthis information literacy proposal:computers, application programming,
7u
60 NECC 1980
data, and communications. The computerportion could be satisfied with thetraditional "Introduction to Computing/Data Processing" course currently avail-able at many institutions. There is noneed to elaborate here on the contentsof this type of course.
Application Programming.Coding-in a high-level language is
not sufficient to understand the roleof application programming. In fact,for the generalist, it may be of second-ary importance to an awareness of theapplication requirements/preliminarydesign process. The following topicsshould be addressed:
. Requirements Study- Documenting information flow
in an organization. Prz:tminary Application System
Dep4Ln- Transaction definitions- Screen formats
. Generation of a SoftwareSpecification
. Design of an ApplicationSoftware Module- Flowcharts
RIPO diagrams. Coding of an Application
'Software Module- Coding- Testing- Debugging
DataMuch as the treasurer is respon-
sible for the money resource in -anorganization, the business communityhas come to realize the value of theirdata records andthe need to assignmanagement responsibility for theirsafekeeping and use. An infornsticliteracy course should include thefollowing subjects.
. The Concept of Data as a Re-source
. Data Characteristics- Physical relationships
Logical relationships. Storage Alternatives. The Role of Data Base Adminis-
tration- Data dictionary concepts
CommunicationsThe fourth and final component of
the information literacy offering, com-munications, should impart an awarenessof the regulatory, cost, and technolo-gical aspects of the links in a contem-porary information system,
. Industry Structure- Common carrier services
. Switched
. Leased
. Value added - packetswitched
. Satellite. Other
- Tariffs. Information Networks
- Combining count:Ideation°and data processing
- Network architectures. Centralized. Distributed
. Design Tradeoffs- Nigh volume - batch
Interactive- Public vs private
It should be noted that each ofthese four topics is a complex special-ty in its own right. No pretense ismade that skill in any of the topicswould be developed through the course,but rather a conceptual knowledge ofthe elements in an information systemwould be gained. Also of equal impor-tance, the role of the end user in aninformtion system design would be estab-lished.
A two - semester sequence would bedesirable; however, a one - semester sur-vey course could also be of value. Thebusiness school, because of its prece-dence in information systems curricula,would be a logical choice to ofer thisintroductory course.
In conclusion, it is hoped thatthis presentation stimulates thinkingin the educational community withregard to its roil in preparing atmdents to coexist with information tech-nology in the 21st century.
NOTES1) Andrew R. Molnar, "The Next Great
Crisis in American Education: Com-puter Literacy," ARDS Journal, Asso-cief'-.1 for EducationalMISys-tems, Volume 12, No. 1, Fall 1978,p. 21.
(2) " Curriculum '781 Recommendationsfor the Undergraduate Program inComputer Science," Communicationsof the ACM, Vol. 22, No. 3, March1149.
(4) R. L. Ashenhurst, ed., "CurriculumRecommendations for Oteduate Pro-
fessional Programs in InformationSystems," 4 retort of the ACM Curri-culum Committee on Camputer Educa-tion for Managemen, 2972.
(5) J. Hamblen, "Computer Education inNigher Education -- _Status, Alter-natives and Needs," AFIPS, :978.
(6) "ACM Cites Dearth of DP Programs,"Computerworld, November 5, 19)9.
7o
Computer Literacy 61
82 NECC 1980
A BYTE OF BASIC
Judith A. Hopper
Arapahoe CountyDistrict 6
Grant Jdnior HighLittleton, Colorado 80122
(303) 795-2560
INTRODUCTIONThe average U. S. citizen has not the
foggiest'idea of how computers work andhow pervasive their influence actually is.Consequently, he has no idea of what todo when a computer makes a mistake; he hasno idea of how to vote on local, state, urnational issues involving computers (e.g.,the establishment of a national databank)s he is, in short, culturally dis-_advantaged.
It is therefore essential that oureducational system be modified in such away that every student (i.e., every pro-spective citizen) become acquainted withthe nature of computers and the currentand potential roles which they play inour society. It is probably too late todo much about adults, but it would bedisastrous to neglect the next genera-tion.
- Committee on Computer Education,Conference Board of the Mathe-
computer fundamentals is one veryimportant factor in building an informedcitizenry for the future. Students whoare computer literate will have .settercareer opportunities and less careershock. They will be better able to copein a world with rapidly moving and ever-changing technology. Thus, in this eracalled the Age of Information, schoolscan no longer delay in bringing thecomputer into their curricula.
A BYTE OP BASICSeeing the need to get some kind of
program with computers started in herschool district, the author of this paperwrote a proposal to pilot a computerliteracy and BASIC programming class atthe junior high level. The district pur-chased an Apple II microcomputer andthings were off and running.
A class of fifteen eighth and ninthgraders was chosen for the project. The
class, to be one semester long (eighteenweeks), met five days a week for fifty-five minutes. The Apple II is a 32Ksystem with a Centronics printer. Xt wasfelt a hard copy would be a benefit inthat the students would have somethingtangible to take home to show to theirparents.
The class was begun by acquainting thestudents with computers in generals whatis a computer system, what are its parts,and what does each part do. Using theirknowledge and expanding what they knew,the class discussed various types ofinput/output devices as well as particularusage, i.e,, in business, industry,science, and research.
In order to get the students on thecomputer as quickly as possible, flow-charting and algorithms were tae nexttopics presented. As it was important forthe students to be able to break down aproblem into its component parts andanalyze each stzp of the solution, flow-charting was introduced as a pictorialrepresentation of this pzocess.
Simple programming was introduced nextto acquaint students with the computer.Feeling comfortable with the computer andlearning the on-off procedures in usingthe computer were the main objectives.The introductory programming was simple toinsure each student initial success at theterminal.
Egg cartons were then used to buildpersonal computers for each student. Inbeginning to write more difficult programsthe students had to put their programthrough their "EGG 12" system first to besure it was doing what they had intended.Once it was demonstrated that it would runon their computer, they could run it onthe Apple. Using this technique forcedthem to evaluate their own %ark and makecorrections as needed.
As the class progressed through thesemester, more and complex programming
73
was introduced. Most all of the commonBASIC statements were used (LET, IF-THEN,TAB, GO TO, REM, FOR-NEXT, PRINT, etc.)and also some library functions (SOB, ABS,INT, RND). In addition, as the machinebeing used has color graphic and stringvariable capabilities, these were soonadded to the repertoire of the studentprogrammers.
Interspersed among the programminglessons, other topics were presented. Inanswer to the question of what does thecomputer do with the program once it hasbeen typed in at the terminal, machinelanguage and the binary system were dis-cussed. Changing numbers back and forthfrom base two to base ten, and adding,subtracting, and multiplying in base twowere some typical class activities. Acomparison was also made of a programwritten first in BASIC, then in.assembler,and finally in machine language.
*he concept of language interactionresulted in the introduction of theorigins of BASIC and the history of com-puting. Beginning with Stonehenge, theabacus, and the slide rule, the classproceeded through Pascal, Babbage,Hollerith, Aiken, Hopper, von Neumann,and others. This discussion of thedevelopment of the computer and itschanging faces served to illustratewhere we are in terms of both currentcomputing capabilities and conjecturesabout the future.
The class enjoyed learning to usevarious computing devices, the abacusbeing of particular interest to them.The slide rule and calculator were alsointroduced as well as a ten-year oldHewlett-Packard.2000 series computer,which required marking Hollerith cardswith a pencil. As the students wereexposed to these devices, they learnedboth the limitations and the capabilitiesof each.
An important part of any literacyclass is not only learning what a computercan and cannot do but also keeping up withsome of the literature available on com-puters. Throughout the course thestudents had to read a number of articlesfrom various periodicals acquainting themwith what was going on in the world ofcomputing. They had to bring currentarticles in from magazines and newspapersand write reports. The vocabulary ofcomputer -ese was carefully ()is:111890d sothat as complete an understanding aspossible was available. Any neY developmeats of computer usage were also pointedout and discussed.
Field trips were made to observe theuse of computers outside the olassro,m,and several people working in t:o com-puting field were invited 'do speak to the
Computer Literacy 63
class. They discussed various jobsdirectly related to using the computer,the training required, the responsibili-ties that go with the jobs, and the avail-ability of openings in the field.
It was hoped that with the completionof this course the students would have aworking background knowledge of program»ming, that all fears of a computer wouldhave been extinguished, and that theywould be aware that computers have greatpotential and capabilities but also limit-ations. If the development of thecomputer continues at the rapid pace atwhich it is now moving, the students'world will truly be a computerized world.Any knowledge of the field they can gainnow will serse to make their lives thatmuch easier upon completion of theireducation.
FINDINGSThis first class was an enthusiastic
group of fifteen students. They wereexcited about working with the computerand could hardly wait for some hands-onexperience. This enthusiasm continued.hroughout the semester. Each time thestudents wrote a program, they immediatelywanted to test their programming abilities,gaining a great deal of satisfaction inwatching the computer do what they hadintended.
But while there was enthusiasm, therewas also some apprehension among thestudents. One of the primary goals of theclass was to overcome this fear. Thestudents need to be comfortable and feelat home with the terminal. Providing themsome initial success with computers was away of achieving this goal.
The biggest problem for the class wasthe availability of only a single terminal.With fifteen students writing two andthree programs a week it was difficult toschedule time so they each could use thecomputer at least twice a week. Some camein before v-hool, others during lunch andafter school, as well as during the fullhour being used during class. It is hopedthis problem will be alleviated nextsemester with the acquisition of fouradditional microcomputers.
Another problem encountered was the lackof typing skills in the students. Havingto use the one-finger-hunt-and-peck methodconsumed a lot of time when they:werereedy-to type in their programs. As ther.4.ass progressed through the semester theybecame better and faster at getting theirprograms in, but it remained a time-consuming problem throughout.
Now that the first class is over, thestudents are asking for a second class ata higher level. They are very interestedin the many facets of the machine and
74
84 NECC 1 980
would like to pursue its capabilities.Unfortunately a second course in computingis not currently available at the juniorhigh level (it is at the high school), butit is hoped that these students can beused as resource personnel. With theirability to program and run demonstrations,they may be called on by teachers wantinginformation or simulations. By using thecomputer in other classrooms, it canberme an integral part of the entireschool curriculum.
FUTURE OUTLOOKThe computer programming and literacy
class at Grant Junior High was a pilotclass for the district. Before jumpingin with both feet the district felt itbest to try an initial program, evaluatethat one, and go on from there.
Concurrently with piloting this class,visitations to other school districtswith a "computer curriculum already in full -swing were made. As at Grant, thereappeared to be good student interest, butalso considerable frustration over theinadequate number of terminals available.
Then, with recommendations from visit-ations, attendance at several conferences,and research in the area of computereducation, a proposal was formulated forthe implementation of microcomputers intothe school district. The proposal callsfor one junior high and one senior high inthe district to be fully equipped to beginthe implementation. It is hoped thatthese two schools working together wouldbe able to iron out any problems thatwould arise so the full district-wideimplementation would progress smoothly.
The proposal for the computer imple-mentation specifies definite curriculumrecommendations. At the elementary levelthe gifted and talented program alreadyis doing some programming. The proposalcalls for some subtle exposure of theother elementary students -- at least toacquaint them with computers so thattheir fears are eliminated.
At the junior high level it is plannedthat all students at some time duringtheir thtee-year career take a computerliteracy course (possibly nine weeks).Structured within the class would be abasic introduction to computing -- whatthe computer can and cannot do. Inaddition to the literacy course a classin computer programming would be offeredfor those students who would like topursue the programming aspect. Also itis hoped other disciplines (math, science,social studies, and English) would want touse the computer as a demonstration toolfor simulati Is, data processing forexperimental r classroom recruits,
tutoring individuals, or any other func-tions -- limited only by the imagination.
Promtaming in higher level languagesas well as practice in developing softwarewould be offered at the senior high level.The proposal also recommends the computerbe used as a tool in higher level mathcourses, in science and social scienceclasses, and in the business departmentfor data processing.
As staff members at each of the schools,in addition to the students, would needexposure to the machines, the proposalalso provides for inservices and actualclasses for faculty members not familiarwith computers. %t is hoped that teachersin various disciplines would then beinterested in the computers and want touse them as a demonstration tool in theirclassrooms.
The school board has just recentlyaccepted this proposal for a two-yearimplementation of microcomputers into theschools in the district. The hardwarerequirements are not merely cost-effectivebut also offer the opportunity to improvethe quality of education. Where once thistype of technology was used by a fewprivileged persons, today it is becomingubiquitous. A degree of flexibility hasalso been built into the proposal sochanges or enhancements to improve theprogram can be made.
Education is missing an opportunity ifnot a responsibility if it fails to pro-vide its students with the necessarybackground to move into the "informationage." We need to start now to provide thepublic with the understanding of thisinformation science and technology. AsJ.C.R. Licklider put it "People mustmaster the technology or be mastered byit."
EQUIPMENT LISTHigh Schools:
Apple II 48KVideoDouble diskRP OscillatorFirmware CardRadio Shack 4C!32K additionDouble disk15 Radio Shack 4K Level IIMaster slaveInterfacing connectorsPrinters where needed
Junior High Schools:Apple II 48KDouble diskPrinter8 Radio Shack 4K Level IIMaster slave
A COMPUTER WORKSHOPFOR ELEMENTARY ANDSECONDARY TEACHERS
Herbert L. DershemJohn T. WhittleHope College
Holland, Michigan 49423(616) 392-5111
INTRODUCTIONAlthough computers have been used by
secondary teachers for a long time, onlyrecently has the microcomputer made iteconomically feasible for the elementaryteacher to use the computer in the class-room. IA addition, new technological de-velopments have made it possible forsecondary teachers outside the fields ofmathematics and science to use the compu-ter as a classroom tool. Recent reportsand recommendations (1,2) have emphasizedthese facts and indicated the need tomake teachers aware of the potential thecomputer possesses as a tool in theclassroom.
This paper describes a two-week work-shop which was offerod by the authors inthe summer of 1979 to provide teacherswith knowledge of computers as they areapplicable to the teacheks' classes.
ORIGIN AND OBJECTIVES or THE WORKSHOPThrough working with students and
teachers at both the elementary andsecondary level, the authors becameconvinced that the computer is a valuableeducatipnal tool. But it was apparentthat the computer could be used effec-tively only if the classroom teacher wasaware of its potential. We believed thatonce the teacher learned to use the com-puter, he or she would be able to use itin creative and innovative ways in theclassroom. The biggest step was over-coming the teachers' fear and awe so thatthey could implement their own classroomcomputer applications or adapt those ofothers to their own needs.
With this in mind, we designed a two-week workshop with the following objec-tives:
1) The teaoner will be sufficientlyfamiliar with the operation of a computerto instruct others in its use.
2) The teacher will know where to find
Computer Literacy es
resources for ideas, activities, andprograms related to the classroom use of
'computers.3) The teacher will know the BASIC
language well enough'to write simple pro-grams, to introduce the language to stu-dents, and to read any BASIC program andmake minor modifications to it.
4) The teacher will have sufficientunderstanding of tI.e way a computer worksto explain it to students.
5) The teacher will know the techniquesand approaches most frequently used ininstructional computer applications andhave experience in their use.
6) The teacher will have Cosigned andimplemented at least one computer activityfor his or her classroom and be capableof developing others.
7) The teacher will know the types ofcomputer equipment available for class-room use and be aware of advantages anddisadvantages of each.
The workshop was intended for teacherswho had no previous experience with compu-ters and who wished to explore ways inwhich they could improve their teachingby use of the computer.
WORKSHOP FORMATThe workshop was divided into two
parts, a laboratory and a lecture, each ofwhich met for one hour and fifteen min-utes every day for two weeks. Upon com-pletion of the course, the teachers re-ceived two hours of college credit.
The BASIC language was taught in thelaboratory portion of the course. Actual-ly,"two laboratory sessions were held eachday, one before and one after the lecture,half of the participants attending eachsession. Ten Radio Shack Level II 16KTICS -80 computer systems, loaned to thecollege by Radio Shack,were set up in thelaboratory and each participant was seatedin front of a unit. The textbook, "Using
76
66 NECC 1980
BASIC "by Julien Hennefeld (3), was chosenWcause it contains many sample BASIC pro-grams to illustrate concepts. A set ofthese programs was placed on cassettesfor each laboratory; then during the lab-oratory each participant would load a pro-gram from cassette. Next, the instructorwould talk about the concept illustrated,have the participants run the program,and usually ask them to make modificationsto see the effect. The topics covered inthe laboratory during the first week wereas follows:
Day 1 - How to use the computerLET statementsPRINT statements
Day 2 - GO TO statementsiF-THEN statements
Day 3 - INPUT statementsFOR-NEXT statements
Day 4 - Subscripted variablesGrapiics
Day 5 - String variablesPRINT@ statements
During the second week the laboratorytime was used for each participant todesign and implement a useful classroomcomputer activity. The instructors as-sisted the participants in the designand programming of the activities. Addi-tional features of the BASIC language andadvanced programming techniques were cov-ered as they were needed. On the finalday of the workshop each of the partici-pants presented the results of his or herproject to the entire group.
In the lecture portion of the workshop,the participants were divided into ele-mentary and secondary educators. The top-ics of the lectures for each group follow:
Day 1 - How Computers WorkDay 2 - Techniques and Approaches in
the ClassroomDay 3 - Elementary - More techniques
and approachesSecondary - Survey of Resources
Day 4 - Elementary - Survey of Re-sourcesSecondary - Problem SolvingPrincipals
Day S - Computer LiteracyDay 6 - Design Considerations for
instructional HardwareDay 7 - Experience with instructional
SoftwareDay 8 - Survey of Computer EquipmentDay 9 - Elementary - More experience
with instructional softwareSecondary - How to TeachComputing
Day 10 - Presentation of Projects
Both laboratories and the lecture wereheld in the morning. The room containingthe computers was also left open in theafternoons so the participants could re-turn and work on their projects. Althoughthis afternoon work was not required, manyteachers did take advantage of this oppor-tunity. in addition, several had TRS-80systems at home or at their schools whichthey used in the afternoons or evenings.
RESULTSTwenty educators from three local
school districts attended the workshop.Of these, twelve were secondary teachers,four were elementary teachers, and fourwere administrators. All four administra-tors attended the first week only. Also,some of the teachers who had previouscomputer experience took the second weekonly.
A brief description of the projectsdeveloped by the teachers is found in theAppendix. Both the instructors and theparticipants were amazed at what has beenaccomplished in a two-week period. Anenthusiasm for computer use was generatedby the workshop so that participants wentback to their schools and pushed for thepurchase of a computer for their classroom.
A further indication of the success ofthe workshop is the demand for it to berepeated this summer. This demand is com-i%g from teachers who saw the effect theworkshop had on last year's participants.As a result, the same workshop will beoffered twice this summer as well as onceduring the academic year. Hope College hasnow established its own microcomputer lab-oratory, so the computers will not be bor-rowed systems.
The only modification to the workshop.which we plan is not to allow participantsto register for just one week. Those whotook only the first week missed the im-portant experience of putting togethertheir own software. it was also difficultto adjust the workshop to those who camein during the second week. As a result,the workshop will not serve the needs ofthose teachers with some previous computerexperience.
interest has been expressed is a work-shop specifically for teachers who arealready using computers. We are consider-ing offering such a workshop, which wouldcover advanced programming and softwaredesign techniques. in this workshop wewould invite each teacher to bring along astudent so that they could participate asa teacher-student team. From our observa-tions, much of the software developed isactually done by such teacher-studentteams with the teacher doing the design
and the student the programming.
REFERENCES1) Milner, S. "An Analysis of Computer
Education Needs for K-12 Teachers.' Na-tional Educational Computing _ConferencePromeedNg1773WCity, 1979.
2) Taylor, R., Poirot,'.7., Powell, J.,andtHamblen, J. "Computing Competenciesfor School Teachers: A Preliminary Pro-jection for All but the Teacher of Com-puting." National Educational ComputingConference PromeedailtriBirCity, 1979.
J. EitasBASIC: An Intro-duction to Computer Prograiragi. WinnWiW7Ntlaidt, Inc., 1978.
APPENDIX. PROJECTS COMPLETED BY PARTICI--. PANTS IN THE COMPUTER WORKSHOP
Physics experiment simulation - A fallingbody experiment is simulated by the cam-pater. The student is asked to providethe appropriate formulas to calculateexpected results.
Geometry drill - A drill and practiceexercise is conducted using geometricterminology.
Spanish drill - A drill and practice onSpanish grammar and vocabulary is con-ducted entirely in the Spanish language.
Ordering - This program is intended foruse at the elementary level. The studenttakes a list of items and places them ina described order. Three implementedorderings are alphabetical, fractions, anddecimals.
Parts of Speech - The student is given aword list and after picking a word fromthis list, leads the computer to identi-fying the word by answering the computer'squestions about its part of speech.
Word house. - This program is intended forstudents learning English as a secondlanguage. The student is required to placeeach of a list of words into its propercategory and word house.
Presidential drill - This program drillsstudents on presidents and their terms ofoffice.
Golf statistics - This data collection andanalysis program simplifies the paper workof the golf coach.
Career counseling aid - A student inter -sated in a career in aocounting can sitdown with this program and find out what
Computer Literacy 67
options exist in this field and the edu-cation necessary for pursuing each option.,
Test generator - This test generationprogram simplifies the high school mathteacher's job of creating examinations byrandomly choosing problems which haveparameters that may also be randomly se-lected.
Carrying drill - This program drills thestudent on carrying skills in multiplica-tion by presenting randomly generated pro-blems. If the student responds incorrect-ly the program carries out the multipli-cation, carefully specifying each carryvalue along the way.
Mortgage payoff - The student can usethis program to examine the effects ofvarying the parameters on mortgage pay-offs.
68 NECC 1980
MICROCOMPUTERS
AND COMPUTER LITERACY;
A CASE STUDY
Robert J. EllisonHamilton CollegeC).inton, NY 13323
315-859-4138
I win discuss the role of the micro.computer in computer education at HamiltonCollege. The focus will be on the kirdsof applications I have found useful and onthe methodology that was used to developthe systems on the microcomputer. My ini-tial interest in the microcomputer hasbeen its use as an aid to teach computerliteracy. I will discuss the developmentof an interac ive statistical package thatis the core df a three week introductionto computing and programming. The packagepermits elementary programming and filemanagement to be introduced eithout an exttensive introduction to programming.
While the microcomputer can be an ex-cellent instructional aid, it quite oftenis inconvenient for program development.Most of our applications could not be de-velope4 on a larger machine, however,since we are very dependent on programmingthe screen and the graphics displays. Iwill discuss our experience with the UCSDPASCAL(TM) system which is now availableon a number of different computers. Ithas significantly aided the design andmanagement of the project. The system'sextensive use of prompt lines makes it easyfor the novice student to use.
It was my goal to develop systemswhich could be easily modified and expanded.The microcomputer is not large enough tosupport a universal package; furthermore,the more complete systems usually alsobring with them a more complicated controlstructure. Since our goal is to introducecomputing to an audience with little if anycomputer experience, we wanted a simple andoften specialized user interface. But ifwe are not going to offer a fairly com-plete package for the micro, then weshould design a system which can beeasily modiried. The statistical packageI will discuss was written using a libraryof procedures which can be integrated intoa custom system. The package has been sup-
plemented by both the mathematics andbiol-ogy departments for their own special ap-plications. Some of the materials willalso appear in the development of a filemaintenance system in the advanced comptterscience course on data structures. Thelimited professional support for academicprogramming in a small college practicallyforces us to write generalpurposesoftwear.
THE ENVIRONMENTHamilton College is a liberal arts in-
stitution with an enrollment of 1600 stu-dents and a faculty of about 135. Over 600students a year will take courses whichuse the computer. The computer science of-ferings consist of a two-course program-ming sequence (6) and two advanced courseswhich emphasize file structures and topicsin data base organization. The first pro-gramming course is taken by about one thirdof the students before they graduate. Theacademic computing is done using a remotebatch entry to the IBM 370/168 at CornellUniversity. We will shortly be able to usethe interactive system at Cornell also.The advanced computer offerings are taughtusing PASCAL and microcomputers.
r am exploring the possibility of ap-plications which do not require the fullcapability of the IBM/370 or which cantake advantage of the simpler operatingenvironment of the microcomputer. Whilethe Cornell system supports a quite adequatenumber of statistical packages and progimmrming languages, for the novice tLe systemis still too complex to be easily used. Anumber of applications involving elemen-tary program instruction or simple statis-tical analysis might be better served on asmaller machine or one specially configuredfor the use.
Learning the control language for theuse of a computer system is a special prob-lem for those non-computing courses thatwant to explore the use of the computer as
79
a tool. The focus is too often on the de-tails of the system rather than the prob-lem to be solved. A well designed micro-computer system might be of service herealso.
We have selected the TERAK microcom-puter as our primary unit. The CPU is aDigital Equipment LSI-11. The TERAK waschosen for its graphics capability, a 320by 240 raster image, and the wide range ofsoftware available. Operating systems forthe TERAK include DEC's RT-11, the UCSDPASCAL(TM) operating system, which we usemost often, and the Cornell Program Synthe-sizer (13), which I will discuss below.The TERAK can also be used as a terminalto Cornell, and with the use of a communi-cations package text files can be trans-ferred between the two computers. Programsexist to convert files between the RT-11and the UCSD disk formats. Ihavedevelopedsoftware to transfer the graphics andcharacter buffers to a Printronix whichhas a plot capability.
GOALS 4
Our most immediate need was foresys-tem to support a new offering for computerliteracy in our winter term. Hamilton fol-lows a 4-1-4 calendar. It was my intent tooffer a short course which couldhelpmeetthe demand for computer literacy. Sinceour first programming course is designedfor a general audience, I did not feel theneed to teach programming. On the otherhand it seemed important to introduce stu-dents to the use of the computer to storeinformation and hence to the concept of afile.
I felt it was quite important to in-volve the student more than as a passiveobserver of the computer. Ourgreatestprob-lem was finding programs and applicationswhich involved the student with more thanselecting an option from a menu. Iampar-ticularly interested in examples which in-volved the maintenance of files on the com-puter. An immediate application was the useof a text editor and text formatter. TheUCSD screen editor is very easy to use. Ihad also written the formatter (9)inPASCAL.I added user-defined macros with the out-put directed to the printer, disk file, orscreen. While the text editor could be usedto discuss files and updating of informa-tion, I wanted a more complex example.
The use of a statistical packageseemed a possibility. The choice was inpart motivated by use of the SAS system(12) which makes extensive use of both per-manent and temporary files. A SAS run con-sists of a number of procedures to list,sort, and modify the file. Most procedurescreate a file to pass data to the next
f
Computer Literacy 69
step of the process. SAS also supportslimited programming in PL/1. Inmost casesthe program code is applied to each recordof the file. It seemed that a package suchas SAS could serve to introduce the con-cept of a file, program variables, andelementary programming. Using files as theprimary way to pass data is ideally suitedfor a microcomputer with limited memory.
The use of SAS presented some prob-lems for our short course. Our computingon the Cornell system is primarily doneby remote batch entry. Although I preferredto use an interactive system for thiscourse, the IBM control language for theinteractive system was too complex forour audience. It also seemed difficult tointroduce in such a short time the syntaxand semantics of even an elementary partof PL /l.
My solution to the above problem wasmotivated by the availability of aprograra-ming system for the LSI-11 (13) called theCornell Program Synthesizer. The systemincludes an editor for the language PL/CS(S), which gives a template for each majorconstruction. For example, a single key-board command generates the template
If (condition) then(statement)
Statements are checked as entered. Thesystem is designed for teaching program-ming and includes elaborate editing andtracing commands. I wanted to include someof these features to assist in enteringelementary programs for the statisticalsystem.
THE STATISTICAL PACKAGEThe system supports real, integer,
and character variables. The core of thesystem is an elementary file managementsystem that includes modules to open andclose file, get or put the next record,and retrieve or store individual variablesin a record. Each statistical file consistsof a data file and a directory file thatlists the variable names and types. Thefile management system is stored in thesystem library and can be called from anyPASCAL program. The system is designed forinstructional use, not to support large-scale statistical analysis. It can serveas a tutorial for the use of a more com-plete, flexible, and faster system such asSAS. The system consists of the followinggeneral programs.
Create:he record and directory files are
created from the text files made by theeditor.
70 NECC 1980
List:The directory and the data are listed
on either the screen or the printer. Vari-able headings are automatically printed.A single record may take several lines.
Sort:The file may be ordered by any of the
variables in ascending or descending order.Multiple sort keys can be specified. It mayalso be used as a procedure in other pro-grams.
This module supports an elementaryprogramming language to permit modificationof the file. The statements are promptedand errors are noted on entry. The programis saved on the disk for documentation.Currently, an individual statement can beedited as it is entered but cannot bechanged after it is accepted. I will notinclude elaborate editing, since most ofthe programs are rather short and are easilyentered again.
The statements supported and the tem-plate provided include;
Assignment:
(variable name) = (expression):
New variables are declared simply.byentering the name. The variable type isautomatically set by the type of expression.
Conditionals:
If (condition) then do:(statements)end;
Select;when (condition) do;
(statements)end:
when (condition) d0;(statements)end;
otherwise do;(statements)end;
end;
The SELECT statement has been recentlyadded to the IBM version of PL/I. Only thestatements associated with the first truecondition will be executed. The OTHERWISEblock may be empty. The SELECT was usedinstead of an IF..ELSE construction as theformer seemed to be less confusing. Ichose not to implement the GO TO statement.
4.4
It not only complicated the design of theinterpreter for the language but alsocreated the possibility of infinite loops.
Input /output
Output (file name*
The current record is written to theselected file. When the file is openedthe user has the option of specifyingwhich variables will be written. Therecord has all the variables in the inputfile as well as any new variables intro-duced by the program. Currently two outputfiles are permitted. The program promptsfor an output file if none is specified.
If an expression or variable is notcorrectly entered,a list of the variablesand their types can be requested beforethe correction is made. The program isautomatically indented to reflect thenesting of conditionals.
The following is a sample program.The variables QU/ZI, QUIZ2, FINAL, andMIDTERM are on the input file. The vari-able TOTAL is created by this program.The lower case entries were supplied bythe program.
TOTAL = 0 ;if NOT(QUIZI MISSING) then do;
TOTAL = TOTAL + QUIZIend;
if NOT(QUIZ2 MISSING) .aen do;TOTAL = TOTAL + QUIZ2 ;end;
TOTAL = TOTAL + FINAL + MIDTERM ;output DEMO ;
The system was used for the firsttime this winter. Our program editing wasnot as elaborate as apparently was needed,but the overall student response was quitepositive. To the student, the basic pro-gram unit is the file which for a generalcourse on computer literacy was appropri-ate, The limited programming also helpsexplain how a computer workst We currentlyplan to use the same system with our first.programming course to introduce files andinteractive computer systems. A variant ofthe system will be used in the first yearbiology sequence.
The system is easily expanded. Newprograms are written in PASCAL with thePASCAL input/output functions replaced bythose in the new file management system.It is particularly easy to have a studentdevelop an independent PASCAL program fora specific file and then replace the input/output functions to generate a generalpurpose program. The system, especiallythe file management section, has been an
8j
excellent source of examples for the advancedcomputer science students.
DEVELOPMENTI would like to discuss the way we
have managed the development of this sys-tem and the use of UCSD implementation ofPASCAL. A small college often does not havethe p ogreaming support available fox thedevelopment of large systems. The prolif-eration of a number of specialized pack-ages would only compound the maintenanceof these programs. While the initial effortis greater, I have tried todevelopgeneral-purpose programs which can serve in otherapplications, for example, the file man-agement system which supports both thestatistical package and general file main-tenance programs for an advanced computerscience course on file structures. Thegraphics library, which I have not discussed,is used by mathematics, biology, and physics.
It is clearly desirable to reuse soft-ware, but keeping each of the various sub-systems supplied with the latest revisionscan btl difficult to manage on a small sys-tem. Programming in PASCAL and the UCSDimplementation of it have both assisted inthis task. A vital feature of PASCAL isthe ability to define new data types.PASCAL includes the simple variable typesinteger, real,and character. But for thekinds of systems programming involved inthis project, we needed more elaboratestructures. For example, we had to main-tain a directory for each file which main-tains the variables and their type. Thedeini+ion of such a directory had to beconsistent throughout the system. In PASCALwe could define a directory in severalsteps.
Name : namostr;Position integer;Typeofvars (realtype,integertype,
chartype) ;Length; integer { length in bytes)Max ; real;Min real;end
We can then declare the type directory as
directory m array(1..maxvars]of vardescript;
Computer Literacy 71
A procedure GETDIR whose purpose is toread the directory could be declared withthe following heading:
Procedure GETDIR( VAR D; DIRECTORY);
If it is necessary to add more informa-tion to our description, such as the num-ber ok missing values, then we only needadd a line such as
missing ; integer;
to the original definition. That changeis then reflected in all procedures whichuse that data type.
The UCSD system has included somenon-standard features. I have tried toavoid the use of them when possible, buttheir definition of a unit has been quitevaluable. A unit is a separately compiledlibrary of subprocedures. It can also in-clude type definitions. I keep the systemtype definitions in the library. Thosetypes are then copied into each programwhich uses that unit. Changes in the com-mon definitions are then reflected in allprograms which use that module. It hasbeen an excellent tool to maintain consis-tency in student written programs.
The procedures and definitions canbe included in the program by the simplecommand
uses {unit name };
The statistical system has a unit devotedto file management, another to parsing andevaluating expressions, and a third unitto provide run-time support. In additionwe have other units with graphics proce-dures or random number generators. The useof units does not affect portability cfthe system. The source code of the proce-dures could be inserted instead of theuses statement.
SUMMARYWe are continually surprised by the
ease with which major changes can be madeto the system whose length now exceeds5000 lines. The use of PASCAL and the UCSDunit has played a major role. It has beenvery easy to develop the system incremen-tally and to quickly train students toassist in the programming. PASCAL is wellsuited to the kind of systems programminginvolved in this project.
We have been pleased with the studentreaction to the statistical system. Thecombination of that system along with thePL/CS programming system has provided aquick introduction to programming and theuse of files for our course on computers
82
72 NECC 1980
and society. We are planning to developtutorial materials so the system can beused by other courses. Further evaluationof the system will have to wait until ithas been uzed by other disciplines. Wehave been successful in having other fac-ulty use the modules to develop their ownspecialized software. We are now evaluatingwhat additional modules should be placedin common units.
REFERENCES
1. Austing, R.H., Barnes, B.H., Bonnette,D.T., Engel, G.L., and Stokes, G.S."Curriculum '711; Recommendations forthe Undergraduate Program in ComputeiScience--A Report on the ACM CurriculumCommittee on Computer Science." Comm.ACM 22, 3(March 1979), i47-L66
2. Austing, R.H., and Engel, G.L. "A Com-puter Science Course for Small Colleges."Comm. ACM 16, 3(March 1973), 139-147
3. Committee on the Undergraduate Programin Mathematics of MAA. A compendium ofCUPM recommendations. /IAA 11 (1975),52S-570
4. Conway, R. Primer on Disci lined Pro-gramming UsIRFE/ag.-Iritt top 9:1-
5. Conway, R. and Constable, R. "PL /CS --A Disciplined Subset of PL/I." Techni-cal Report 76-293, Department of Compu-tar Potence, Cornell 1976
6. Ellison, Robert J. "A ProgrammingSequence for a Liberal At College."Proceedings of the 1980 SIGCSE Sympo-sium on Cdbeuter Science Education
7. Grogono, Peter. Pro rammin i. Pascal.Addison-Wesley,
S. Jensen, Kathleen and Wirth, Siklaus.Pascal User Manual and Report. 2ndarfan:SFr_rWViiEreg 1974
9. Kernighan, Brian and Plauger, P.J.Software Tools. Addison-Wesley 1976
10. Lopez, A.A., Raymond, R., and Tardiff,R. "A Survey of Computer Science Offer-ings in Small Liberal Arts Colleges."Comm. ACM 20, 12(Dec 1977), 902-906
1.1. Nevison, J.M. "Computing in the LiberalArts College.`' Science 194 (Oct. 1976),396-402
'12. SAS Users Guide 1979 Edition. SAS17WIENZIWIT
13. Teitelbaum, R.T. "The Cornell ProgramSynthesizer; A Microcomputer Implemen-tation of PL/CS." Technical Report 79-370, Department of Computer Science,Cornell University 1979
14. Van Loan, Charles. hComputer Sciencernd the Liberal Arts Student." Techni-cal Report 79-376, Department of Com-puter Science, Cornell University 1979
3
InVited Session
PERSONAL COMPUT/NG:AN ADVENTURE OP THE MIND.
Paul HazenApplied Physics LaboratoryJohns Hopkins University
ABSTRACTfirgrants from the IEEE Computer
Society, The Johns Hopkins University,Radio Shack, and other agencies, theInternational Instructional TV Cooperative,source of instructional TV materials toall educational TV networks nation-wideand internationally, has finished and ismarketiol the implementation of a six-course national educational TV seriesaimed at the pre-college level in thearea of personal computing and computerliteracy. The name of the project is"Personal Computing: An Adventure of theMind.*
The objectives of this new series areto illustrate the uses of perscAalcomputing, to demonstrate the interfaceof humans and machines, to identify thefundamentals of communication in personalcomputing, and to motivate students to beAnnovative in their own applications ofpersonal computing. Since the personalcomputer is viewed by many as a mindmultiplier, a furthcr objective of thiseducational TV ser.las is to greatly in-crease the number of minds that can bemultiplied by taking personal computingto millions of children in classroomsacross the country.Education and informational programs
are closeli allied in that both attemptto communicate facts, concepts, andideas. Both need to be designed withspecific objectives in mind. Some of theobjectives to be discussed .e bothattitudinal and informational in nature;that is, they deal with feelings as wellas facts. The underlyi:q thrust through-out is that . . . LEARNING CAN BE FUN!
8.173
invited Sessions
EDUCATIONAL COMPUTING: EMT, PRESENT, FUTURE
Ronald W. CollinsDept. of Chemistry
Jefferson Science Bldg.Eastern Michigan University
Ypsilanti, MI 48197(313) 487-0106
ABSTRACTDuring the past ton years considerable
effort has been devoted to optimizing therole of computers in education. A vari-ety of tutorial CAI programs have beenwritten, large data bases of questionsfor computer-generated exams have beenamassed, sophisticated software forgraphics has been developed, numerousdata reduction and simulation programshave been written, and the use of on-lineclassroom computing has been studied.In addition, the emergence of minicom-puters and microprocessors has loweredthe cost of computing. Nevertheless,the overall impact of'the computer oneducation has been minimal. One of themajor deterrents has been the poor trans-portability of programs from one computerto another. As a result, educators areoften required to develop their ownsoftware for use in courses; however,many instructors have neither the timenor the experti:,e to do so. Will thenew stand -alone home computer systemsimprove this situation? Possibly, but
74
for most educational uses, the need forprograMining time and exper 4.ca will stillbe high. Furthermore, the several gener-ations of changes in hardware have notyet led to pignificant improvements incomputer-basd pedagogy. Educatorssimply must give more consideration tothe question of how the computer canreshape the way and atyle it hich weteach. To date, most instructional com-puting has simply becn appended to thetraditional pedagogical framewc.k, thusencouraging the self-fulfilling prophecythat the final impact will be minimal.The future? By 1990 (or 2000 at thelatest), there should be a number of newand outstanding examples of true computer -based/- derived /- oriented projects ineducation, rather than a mere continua-tion of the current computer-augmented,yet traditional approach.
THE OPEN UNIVERSITY
Prank LovisThe Open University
Milton Keynes, England ME7 6AA(0908) 653371
William DornMathematics Dept.
Universik of DenverDelver, CO 80210(303) 753-3529
ABSTRACT--TH-TTE2 Britain's Open University willreplace the two second-level computingcourses which it has been running since1973. There will be wily one new coursesince we ha's come tc realize that runningtwo second-level courses is unnecessaryand extravagant. The new course is ex-pected to maintain the current enrollmento . 700 students per year.Production of the new course is well
underway, and in this paper its contentand presentation will be compares withthose of its predecessors. Tile newcourse was designed after careful consid-ation of the views of both students andfaculty members from which emerged theone urgent and clear demand that thecourse should relate closely to thecommercial world of data processing.The main themes of the course are
practical computing, data structures,files and file processtAg, systemsanalysis,and design, and the socialimpact of computing. Introductions tooperating systems, data bases, anddistributed computing are also included.The course components comprise writtenmaterials, computer programmingactivities, 16 television programs, threeteaching audio-cassettes, and sixbroadcast radio programs.
MAW Someone 75
The present paper will concentrate on:(1) Why the DP bias was considered
essential and how it shows-up in thecourse.
(2) The content and presentation ofthe first four main themes listed above.In this connection the systems analysisand design package bears specIsl mention,as it is being developed in collaborationwith the O K.'s National Computing Centreand will involve the student in a practicalproject.
(3) The specification and future imple-mentation of OUSBASIC, a computing languagedesigned to enable the student to writeand run structured programs in the unique- -and awkward -- distance teaching environmentof The Open University.
The paper will be illustrated with twoexcerpts, of ten to fifteon minutes duraticeach, from films of two of the televisionprograms- -The Introduction to Files and"ile Processing and Systems Analysis andDesign.
c*c
Science and Engineering
DEMOGRAPHIC TECHNIQUES IN ECOLOGY:COMPUTER-ENHANCED LEARNING
A. John Getz, Jr.Department or Zoology
Ohio Wesleyan UniversityDelaware, Ohio 43015
(614) 369-4431
INTRODUCTIONJudging from the availability of CON-DUIT programs on ecological topics (5 of8 in biology in the Summer 1979 issue ofPipeline), there may well be more com-puter use in courses related to ecologythan in courses in other areas of biolo-gy. For instance, valuable CONDUIT pro-grams exist that aid students in learningabout population growth using the logis-tic equation and related models, inter-specfic competition using the Lotka-Volterra equations, and energy flowthrough the trophic levels of variousecosystems. Additional areas in ecologyare appropriate for computer-enhancedlearning. In particular, the use of thecomputer to aid students in under-standing life tables is described here.Such work not only helps students under-stand these complex tables, but also per-mits them to quickly and easily get agrasp of the consequences, in terms ofbosh population growth rates and age dis-tributions, of alternative reproductivestrategies.
SIGNIFICANCE OP LIFE TABLESIn many ways, a survivorship and fertili-ty life table represents the ultimatesynthesis of data on the life historycharacteristics of a given population orspecies of organisms. Both age specific
76
death rates (1 ,).and age specific birthrates (mx) must be known and transcribedinto appropriate format. For survivorshipdata, this means knowing the proportionof individuals surviving at the start ofeach age interval; and for fertilitydata, this means recording the number offemale offspring per female in each ageinter.-al. From these data, numerousdescriptive functions can be readilycalculated by formulae of varying degreesof complexit (see Mertz, 1970 or Krebs,1978 for a fuller description of thesequantities). These fUnctions or .uanti-ties include the following.
(1) The gross reproductive rate, G.R.R.,is a hypothetical quantity that indi-cates the multiplication rate per genera-tion if females suffered no mortalityprior to completing reproduction at therates specified in W fertility tables.
G.R.R. = E mxx=0
(2) The net reproductive rate, Ro, is theactual multiplication rate per genera-tion if the population follows both themortality and fertility schedules in thelife table.
8 '7
no 01MxmOi! 1
(3) The mean length of a generation, 0,is given by:
0 alxmxx
Ro
(4) The innate capacity for increase inthe particular environmental conditionsfor which the table was written, rm, isa factor that gives instantaneous growthrates. The calculation of rm must bedone by trial and error in the formula:
-rmxL. e lx.s.x 1.
x 0
(5) The finite rate of increase,X,..isthe factor by which population sizechanges por age interval in the lifetable, e.g., per year if the life tableis in years.
/' a 0 1111
(6) The stable age distribution is theproportion of organisms that would be ineach age category if the population con-tinued to grow indefinitely according tothe schedules in the life table. The pro-portion of organisms in each age categoryx to x+1 is given by
I lxex
sE "4im0
where I, like x, is a subscript indicat-ing age.
While the data that are used to generatea life table in the first place are quiteintuitively understandable, e.g., thenumber of female offspring produced by anaverage six year old female, the derivedterms, especially rm, are much less easi-ly grasped. Because a thorough conceptionof r is desirable in conorete terms(suck as numbers of offspring and num-bers of seasons of reproduction) before astudent starts working with logisticgrowth models, a thorough familiaritywith all the various life table parametersis highly important.
RATIONALE :( 'ONPUlER USEThe reasons using the computer in lifetable analyst flan be divided into argu-ments against h,nd caloulation and argu-
Science and Engineering 7
ments for the computer. Co:18149ring firstthe arguments against hand caltulation,there is the unfortunate, but nonethelessreal truth that many undergraduate stu-dents are terrified by the sorts of sum-mation signs and exponents intrinsic toseveral of the demographic formulae. Atbest, such students may laboriously plugthrough their calculations and thenfinish with absolutely no confidence inthe accuracy of their results. At worst,some of these students may not even beable to perform the calculations. Clear-ly, neither of these situons is con-ducive to students' learning aboutecology. On the other hand, for under-graduates capable of manually performingthe calculations required, such an exer-cise rapidly becomes a numbtaglr mechani-cal process that, too, is not conduciveto their learning about ecology.
As for the arguments for use of She com-puter, three points.oan be made. First,for students ever more used to punchingnumbers into either calculators or com-puters, one additional application ofthe computer generates very littleanxiety. Math phobias can stay submerged.Second, the fast and accurate resultsprovided by the computer maintain boththe interest and confidence ox the stu-dent. Moreover, if the student wishes totest himself and his ability to manuallyarrive at the same answers, the computeroutput does provide such a check. Final-ly, because the tedious parts of thiexercise are done by a tireless computer,the student is free to concentrate oninterpreting the biological significanceof his findings. This is precisely theconcentration that most ecology profes-sors presumably desire, and I have found. ) useful in the present instance.
DESCRIPTION AND APPLICATIONS OF LIFETWith the above points in mind and twoyears' experience teaching about lifeicables without the aid of a computer, Iwrote a short program, LIFET, in BASICfor our PDP 11/TO at Ohio Wesleyan Uni-versity. LIPET would be suitable for useon other similar RSTS/E time-sharingsystems or on minicomputers, and a pro-gram listing is available on 7equest. Inbrief, the program asks the student toenter the age specific survivorship andfertility data for the population inquestion and then proceeds with calcula-tions of the first three of the sixquantities defined above. The innate'capacity for increase is calculated bytrial and error with active participationby the student in the trials. When the
tti8 (6,,
78 NECC 1980ts
student .is satisfied with the value ofrm that he has attained, the programuses that value to aalculate the Finiterate of increase. Finally, the studentis given the option of obtaining astable age distribution or not. A sampleprogram run is given in Figure 1.
There are at least three sorts of appli-cation of this proc:1Lm. First, and mostsimply, is rarely giving the studentssome real data from a natural population(e.g., from Vinegar, 1975 or Medica andTurner, 1976). The students can organizethe data into life table format for
RUN LIFETDO YOU WANT AN EXPLANATION OF THIS PROGRAM (YES OR NO)? YES
THIS PROGRAM CALCULATES LIFE TABLES AND DEMOGRAPHIC PARAMETERS. THE ONLYINFORMhTION NEEDED TO MAKE THESE CALCULATIONS IS THE AGE-SPECIFIC SCHEDULE OFDEATHS AND BIRTHS WHICH YOU SUPPLY AS 1(x) and m(x) VALUES WHERE:
1(x) = PROPORTION OF POPULATION ALIVE AT THE START OF AGE INTERVAL x; ANDm(x) = AVERAGE NUMBER OF BIRTHS TO EACH FEMALE ALIVE AT AGE x
TO RUN THIS PROGRAM, MERELY ENTER THE MAXIMUM AGE TO WHICH FEMALES SURVIVE(RECALL THAT LIFE TABLES ARE FOR FEMALES ONLY) AND THEN FOR EACH AGE CATEGORYENTER THE 1(x) and m(x) VALUES SEPARATED BY A COMMA. THESE DATA CONSTITUTEALL THE INPUT NECESSARY TO MAKE DEMOGRAPHIC CALCULATIONS USING THE FORMULAE INKREBS (1978, pp 160-170). SEE THIS REFERENCE FOR DEFINITIONS AND EQUATIONS.ONLY THE.EXACT r CANNOT BE CALCULATED WITHOUT ADDITIONAL INPUT ON YOUR PART.THIS IS BECAUSE THE EQUATION MUST BE SOLVED BY TRIAL AND ERROR, i.e., BYSUBSTITUTING ONE VALUE FOR r IN THE EQUATION
-rxSUM e 1(x) m(x) = 1
AND THEN SEE/NO HOW CLOSE THE SUM COMES TO ACTUALLY BEING EQUAL TO I.FOR /JUR FIRST TRIAL, USE THE VALUE FOR r APPROXIMATED BY Rc AND G. IF ALLREPRODUCTION TOOK PLACE IN THE SAME YEAR, THIS WILL ALSO BE THE EXACT r. IFTHE SUM YOU GET IS GREATER THAN 1, INCREASE YOUR EST/MATE OF r. IF THE SUM ISLESS THAN 1, DECREASE YOUR ESTIMATE OF r. INITIALLY MAKE FAIRLY LARGE CHANGESIN r AND ONLY LATER MAKE MORE SUBTLE CHANGES AS YOU ADJUST THE SUM TO1 +1-0.005. ONCE YOU ARE SATISFIED WITH YOUR VALUE FOR r, STOP ADJUSTING ITAND THE PROGRAM WILL CONTINUE TO MAKE OTHER CALCULATIONS USING THIS FINALVALUE OF r ON WHICH YOU HAVE DECIDED.
WHAT IS THE LAST AGE AT WHICH SOME FEMALES ARE ALIVE? 3FOR X = 0 WHAT ARE THE OBSERVED VALUES FOR1(x),m(x)? 1.0,0FOR X = 1 WHAT ARE THE OBSERVED VALUES FOR1(x),m(x)? .9,2FOR X = 2 WHAT ARE THE OBSERVED VALUES FOR1(x),m(x)? .7,1FOR X 22 3 WHAT ARE THE OBSERVED VALUES FOR1(x),m(x)? .5,1
THE GROSS REPRODUCTIVE RATE, G.R.R. = 4THE NET REPRODUCTIVE RATE, Ro= 3THE GENERATION TIME, 0= 1.56667AS APPROXIMATED BY Ro and 0, r=.701242
THE PROGRAMWHAT IS THESOLVING THEDO YOU WISHWHAT IS THESOLVING THEDO YOU WISHWHAT IS THESOLVING THE
WILL NOW CALCULATE AN EXACT r by TRIAL AND ERROR.VALUE FOR r THAT YOU WISH TO TRY? .7EQUATION WITH r= .7 GIVES AN ANSWER OF 1.1277TO TRY AGAIN WITH ANOTHER VALUE FOR r(YES OR NO)? YESVALUE FOR r THAT YOU WISH TO TRY? .8EQUATION WITH rom 8 GIVES AN ANSWER OF .995479TO TRY AGAIN WITH ANOTHER VALUE FOR r(YES OR NO)? YESVALUE FOR r THAT YOU WISH TO TRY? .79EQUATION WITH r= .79 GIVES AM ANSWER OF 1.00784
Figure 1. Sample output from LIFET
Science and Engineering 79
Figure 1 continuedDO YOU WISH TO TRY AGAIN WITH ANOTHER VALUE FOR r(YES OR NO)? YESWHAT IS THE VALUE FOR r THAT YOU WISH TO TRY? .795SOLVING THE EQUATION WITH r= 795 GIVES AN ANSWER OF 1.00164DO YOU WISH TO TRY ',GAIN WITh ANOTHER VALUE FOR r(YES OR NO)? NO
THE FIRM RATE OF INCREASE, LAMBDA= 2.21444THE EXACT INSTANTANEOUS RATE OP INCREASE FOR THE POPULATION, ra .795
DO YOU WISH TO CALCULATE THE STABLE AGE DISTRIBUTION FOR THIS POPULATION? YES
GIVEN A STABLE AGE DISTRIBUTION, TIE PROPORTION OP ORGANISMSIN EACH AGE CATEGORY, C(X), WOULD BE:Of 0 ) = .626875C( 1 ) = .254776c( 2 ) m .C89485c( 3 ) = .288641E-1
x 1(x) m(x) 1(x)m(x)012
121°
31.56667
1.9.7.5
0211
GREs= 4
01.8.7
Rots
.53
r= .795 lambda= 2.21444
DO YOU WANT TO CALCULATE ANOTHER LIFE TABLE (YES OR NO)? NO
themselves if desired, or the data canbe presented as a table from the start.Either way, this students have the ex-citement of finding out for themselveswhether various populations in natureare growing or shrinking and at whatrates. If population growth is occur-ring and data are also available on num-bers of individuals alive in each ageclass, the students can also investigatewhether or not such population growthhas been going on at the given rates fora number of years. If it has, the ob-served proportions of individuals ineach of the age classes should be thesame as the proportions given by theLIFET calculations of the stable agedistribution. Students seem to enjoychecking this assumption of 1:Se tablemethodology.
A second application is the use of datanot just from one population of aspecies, but from multiple populationsof a single species. The wcrk by Tinkleand Ballinger (1972) on intraspecificcomparative demography of a lizard isespecially appropriate. By analyzing theage specific death rates and fertilityrates using LIFET, students can enhancetheir understanding of the populationconsequences for real al.lmals of vari-ations !n life history strategies. Theseparticulan data, in fact, indicate thatsome of the natural populations are just
holding their own and others are decreas-ing in size. With analyses completed,students can either write about or dis-cuss hypothetical alterations in repro-ductive strategy that could improve thelot of the marginally surviving popula-tions.
A third and final example of an applica-tion for this program is to give studentsfree reign to devise an optimal repro-ductive strategy in each of severalhabitats tot a totally LypothOicalorganism. A few guidelines need to belaid down at the outset, and then stu-dents can be set free to uoe theircreativity. For example, one needs tospecify how many kilocalories total afemale has to produce eggs in her life-time, the range in sizes (caloric con-tents) of eggs to be permitted, and theharshness of the various habitat' interms of probabilities of the survivalof offspring to age 1 from each possibleegg size Zn each habitat. Survivorshipthereafter an either be specified or:eft to the students' own devices. Thegeneration of an optimal strategy, oreven possible strategies, for each en-vtronment then "bequires that studentsgenerate a number of life tables. Suc-cessful completion of such an exercise,in my experience, results in an excel-lent appreciation of the merit° anddrawbacks of iteroparity and semelparity
9 u
80 NECC 1980
in various situations.
SUMMARY AND CONCLUSIONSFor the past two years, I have used aninteractive computer program, LIFE?, asa means of enhancing my instruction ofdemographic techniques in an under-graduate course on animal ecology. Ihave found that I have been able tocover more examples and nuances of lifehistory studies since I initiated theprogram than,I,was able to before. Byobviating the time-consuming busy workof life table calculations, I canrealistically expect my students to ac-complish far more sophisticated and ex-tended problems than were possiblebefore I wrote LIVET. In my opinion,students who have had the benefit ofusing this program have gained a fargreater understanding of life tablesand demography than did my earlier stu-dents. Certainly, they have had moreopportunities to do so in a challengingway. What's more, they enjoy it. Thislast point, in and of itself, can be astrong recommendation.
REFERENCESKrebs, C. J. 1978. Ecology ; 'The Experi-mental Analysis of DietributionmndAbundance. 2nd ed. New York: Harper &Row. 670 p.
Medica, P. A., and F. B. Turner. 1976."Reproduction by Uta stansburiana(Reptilia, LacertITTa, Iguanidae) inSouthern Nevada." J. Hernetol. 10:123-128.
Mertz, D. B. 1970. Notes on methods usedin life-history studies. In: Readin ein Ecology and Ecological Gene. es,J. H. Connell, D. B. Mertz, and W. W.Murdoch (eds). New York: harper & Row.PP. 4-17,
Tinkle, D. W., and R. E. Ballinger. 1972."Seelo orus undulatus: A Study of theIntrespeci ic Comparative Demographyof a Lizard." Ecology 53:570-584-
linegar, M. B. 1975. "Demography of theStriped Plateau Lizard, Sceloperusvirgatut." Ecology 56:172-182.
6
Science and Engineering 81
MICROCOMPUTERS AS LABORATORY INSTRUMENTS:TWO APPLICATIONS IN NEUROBIOLOGY.
Richard F. OlivoDepartment of Biological SciencesSmith CollegeNorthampton, Massachusetts 01063
The low cost of microcomputers hasbrought them into the same price range asordinary laboratory instruments, and theirability to collect and store data makesthem an extremely welcome addition to a
laboratory. At Smith College, we have developed two applications for using microcomputers in neurobiology. Pne applicationis suited to advanced st,Jents and researchers and involves the use of a microcomputer and a digitizing tablet to moosOre photographic data; the other is suitable for routine use in a neurophysiologycourse and involves an analog/digital interface for collecting and displayingtransient data. In describing these twoapplications; 1 shall emphasize a numberof aspects that 1 believe are of generalinterest in introducing microcomputersinto undergraduate laboratories.
HARDWARE: THE AIM 651 chose Rockwell's AIM-65 microcom
puter for laboratory use. Like its smallercousins, the KIM and SYM, AIM-65 is a relatively inexpensive singleboard computerthat is based on the 6502 microprocessor.The AIM includes an input/output interfaceWith two 8bit parallel ports, which weuse to connect the computer to laboratoryinstruments, plus two timers and an inter-,.rupt register. The AIM also has a full
kaebOard; a 20character alphanumeric displays and a thermal printer. I consider itan advantage that the AIM does not use avideo monitor, since video would make thesystem less compact and more expensive;the ATM's oneline display is adequate forprompts and date, end the outpOt portsprovide a means (with a digitaltoanalogconverter) of plotting data at high reso!Utters on an oscilloscope or chart recorder. The AIM'. printer further provideseach group of students with inexpensivehard copy.
The AIM is also relatively convenientto program. It has an extensive (8K) moni
for that incorporates a versatile editor,a disassembler, and a pseudoassemblerthat permits writing programs in mnemonicsrather than op codes. A full symbolic assembler, which 1 used for writing our programs, is optional, as is an SK BASIC(both of these are supplied as readonlymemories that plug into designated socketson the AIM board). The assembler and BASICsockets will also accept 2716 erasable,programmable readonly memories (EPROM;).Once a program is debugged, it can be loaded into an EPROM to be installed in theassembler socket, permitting a student torun a program by typing "",° the monitorcall to the assembler. The student doesnot need to be an accomplished computeruser) the programs are highly interactiveand provide abundant prompts for entry ofdata and for menu choices.
In addition to the AIM, its enclosure,and a power supply, two other pieces ofhardware are necessary. The AIM'. 4K bytesof onboard memory are not sufficient forextensive digitization, so that a supplementary memory board is desirable. 1 chosethe Memory Plus, an SK board from The Computerist (P.O. Box 3, S. Chelmsford, MA01824), which has the additional featureof an EPROM programming circuit. Othermanufacturers also make A1M /KIM compatibleSR and 16K memory boards. The other majorpiece of hardware required is ananalog/digital interface. Commercial A/Dboards that can interface to the AIM'sinput/output ports are available, but theones that I am aware of are too slow fordigitization at the rates needed (about 10kHz) in a neurobiology lab. Consequently,we built our own analog interface, which Ishall describe later. At present we havetwo AlMs at work. One is a prototype system with 12K of memor And a single analoginput/output channel; its cost is about$1000, and we are seeking funds to buy andbuild six more such systems for routineclass use. The other AIM has 4K of memoryand is devoted exclusively to taking data
82 NECC 1980
from a HiPad digitising tablet (HoustonInstruments). That ssetoe's cost, including the digitising tablet, is about *1400.For ,comparison, these prices are of thesame order of magnitude as a laboratoryoscilloscope or a chart recorder.
APPLICATION ItIN/111FACINS TO A MIMING TABLET
1 shall first describe our system forcollecting data using Niro/ digitisingtablet. The program has been through several stages of design and is now in heavyregular use. It illustrates several aspacts of interactive laboratory computingthot I believe r of worst importance.
Nourephssiologists typically msisurathe electrical changes in nerve or musclecells in response to stimuli. The stimuliand responses are displayed on an oscilloscope screen and era usually photographedon 35-me film by an automatic camera. Thephotograph* serve as primary data for subempnt analysis of the traces and forillustratiogs for publications. Analysisraquiros measuring the amplitudes and timecourses of the events, which is usuallyden by projecting the film in an enlargeronto graph war and (boiler, microcomputers) to/lovely counting 'soar's. A digitising tblot under the graph paper ishuge improvement* it takes data utmeticells, *poets up the *nise!' by a factorof at least ten, and is intrinsically marsaccurate than hand analysis. For those whohave never seen a small digitizing, tablet.it consists of a flat surface (ours is 11inches suer*) in which wires are imm/des), plus a cross hair cursor that Isplaced on the surface and returns the X,Vceardintos of its position. In using adigitising tablet. two problems most besolved: interfacing the microcomputer tothe tablet, and creating an efficient proves for recordists and analysing data. I
shall describe the interfacing aspectsfirst.
The electronics in the HiPad tabletoutput data in Home' formats. any ofwhich meg be suppressed by jumpers at theoutput connector. Different strobe linesaro available for each of the formats. sothat a microcomputer can be made to attendto one format and to ignore the others. 1
chose to take data in binaryceded decimal(BCD) form rather than in straight binary(hexadecimal), the alternative parallelformat. The main reason for this choice isthat the final output had to be in decimalform since hexadecimal data are incemproheNtlible to most biologists. I believe*at an important principle casts here: astudent or researcher should never be
skid to read or generate hexadecimalnumbers. Since oh* 6502 microprocessor canbe set to do BCD arithmetic, it made senseto start, compute, and end in BCD ratherthan to convert to or from hex.
BCD DATA FORMAT
cooloascommate
OTOMMKthsus um mood
aloud Ward 1 1 1 1 1 1 1 x lxixlx totem:emu 0 I 0 I sot kw 2n0 OYUX kith 2o1 SO are SD 310 BYTEX Allil 4th SO LSO AA OrMY kth 0 1 0 1 Op 1430 sir srmY kith tad SD the SO NO OrMY Aith thlt SD LED Oh 10111
DATA TIMM
11476
OM SINK
Mt= 1.
In its BCD format, $0, HiPd tabletoutputs seven bytes of data for each digitised point. The contents of these sevenbytes are shown in Fig. 1, which alsoshows the timing of the data. Note thatOpt, 1 is a control word. bytes 2, 2, and4 contain the Xcoordinate. and bytes 5.
6. and 7 contain the Vcoordinate. Thetiming diagram shows that the BCD strobeline goes high as each byte is output.Thus. a subroutine to read data into themicrocomputer must test for the presenceof the strobe pulse, then read the currentbyte, store it, end repeat these operations until seven bytes have been read. I
connected the BCD strobe to control lineCAL and the light data lines to the A1M'sPort A. In the 6502 system, input/outputports and associated registers are ros/from and written to as memory locations,so that operations on the strobe pulse anddata resemble reading data from memory.
The full subroutine to take a datapoint is shown in Figura 2. When the sobrotine is called, the microprocessorenters a threeinstruction loop in whichit remains until the strobe pulse occurs.In this loop. it checks the interrupt flag
9 3
register 4SFR, part of the 6522input/output chip) to see if line CAI ishigh. It then reads Port A. checks for thetpected control word, and if it is found.proceeds to read and store the remainingsir bytes of data. Once the sir X,Y byteshave been stored, they ere repacked toplace the sign of each coordinate inseparate byte for more efficient handling.In the overall program. the OETPT subrou-tine is followed by another subroutinethat tests whether the point just takenlies in menu area of the digitizing ta-blet, if so+ the program jumps to the op-propriet* section, otherwise the point istreated as a legitimate data point.
GET POINT Pt NIPAO AXPT DATE: 11/2049
CV ORRENTT(VA) V410LF7(VA). y
SCALEAS/DIV. 44290
WAIOIV: 1000NVIDIV! 0021
01 OEIPT054CC9 LOA IFR0090 AND $02OFAS DEO OETPT130024 LDA PORTAA900 AND WO1053 CAP OFFO1C93
EI916 OEIPT
*6031CC9 LOX 0000503 STROO1110000 LOA IFR
O ARO 0,42MO946000 STA001
DE1045 LDS PORTA1020 ST*
NXXVOATA.X
00 I
DOER CPX 0$042e2915 ONE STR09100F4
itEPACK DATA1900 LDV 04iSNIFT 4 PITS0306 SNFOAT11,4 LSO XVINITA30C9 RON XVOATIS1WO ROO 3V5A10 4205A9 LSD XVDRIA43INC ROR XVOATIS4740 4 DOR
OEVXVOATx,0010.33
APOLDS. ONE SMFDAT
INS
P! A.
PRMIE 5100 lOS R1CLOCK 1A:30 11/28/79
IDASE.....
40020ASNA
49000VOASE 40034 NV
40004 AS
IPEAK 43040 NA00136 MS
VPERK 40094 NV
50220 NSVORIS 44047.20 NV
40444 MS
...... .....wm
FRAME 0031 IDO A1CLOCK 21.30 11/25,79
IOASE 10000 NA40014 AS
VOASE 44031.60 NV46004 MS
IFEAK 43140 NO416MS
VPEAK 400595 40 NV40220 NS
VAALF 4040 0 MY444 os
PIONS 3.
In addition to solving these aspectsof interfacing, it also is necessary toWive an efficient program for data collec-tion. Hy assembly-language program goner-otos almost 2K bytes of object code, a
substantial amount. On initial entry. var-ious constants and tables ore set equal tozero, after which the date of the etperi-ment and several scale factors are re-quested by prompts to the user. The main
Science and Engineering 83
part of the program then begins; in thissection, a series of points is taken for
each frame and stored in tables in memory.Figure 3 shows the print-out made duringthe early part of running the program+ andFigure 4 shows pert of the tables that areprinted after all frames have been meas-ured.
Several aspects of this program are ofgeneral interest. First, every measurementthat the user must make is preceded by anexplicit prompt on the All's display.Prompts make the program easy to use andare always good practice in interactivecomputing. Second, the program can bere-entered without going through the ini-tialization routine; existing data arethereby retained in memory unless they aredeliberately erased. Rosy re-entry makesthe program more forgiving of errors, suchas accidental escape to the monitor, andit also makes it possible to print tablesor re-enter scale factors at intermediatestages in the measurements. Re-entry isaccomplished by two means. The first wasmentioned above: the left edge of the di-gitizing tablet is reserved for a menu.and each point token is tested to see ifit falls in the menu area. The menu isavailable whenever the program is espect-ing a data point, which in practice ismost of the time. The other method ofre.-entry uses the three user-defined keyson the ASH's keyboard. During the initial-ization routine, these keys are programmedto initiate jumps to certain sections of
0.111.0
EXPT DATE. 11/21/79 EXPT OATE. 11/20/79
041FAO VANN TAMA 19115 FRO VPNV V/205 TINE
030 34.00 3440 013601 31.60 1140 0134ME 31.40 MO 004
the program Thus, the user can redirectthe execution of the program fairly conveniently and without losing data alreadyentered.
Another aspect of general interest isthat for each frame measured, the programstores data in a temporary buffer untilthe last measurement for that frame istaken. The data are then transferred automtically to the tables in memory. Thepoint of the temporary storage is to allowthe user to escape and remeasure a framewithout contaminating the final tables ifan error in measurement is suspected. Anyone who hes used program that does notpermit correction at the time of entryknows the frustration that accompaniesbeing forced to continue in-a process oneknows to be incorrect. ln this case, bypermitting a penaltyfree escape in themiddle of a frame. a compromise is achieved between ease of programming and nocorrection routine at all; relatively fewmeasurements are made in one freme, sothat repeating a frame is not onerous, andthe final tables remain accurate.
ln summary. this application of microcomputers to data collection illustratesseveral aspects that 1 believe are characteristic of good interactive programs. Theprogram is permanently in an EPROM and canthereby be called by single keystroke.without requiring that the user know howto run a tape or diskloading routine:also it cannot be accidentally writtenover. The program provides abundantprompts. Data input and output are alwaysin decimal format. The program can bereentered and redirected without erasing
existing data, making it more convenientand more forgiving of errors. Errors ofwhich the user is eware can be correctedat the time of entry. Each of these features is. 1 think, important in bringingmicrocomputers into the laboratory. Thephilosophy behind this approach is to makethe computer friendly to user who is notexpert in programming. As colleague expressed it. "You don't expect students tobuild an oscilloscope in order to use one,so why expect them to program the microcomputers that they use?"
APPLICATION 2:ANALOGTODIOITAL CONVERSION
Our second application will eventuallybe used by more students than the firstone, but the program is still in an earlystye*. so 1 will describe it only briefly.Students in neurobiology laboratories typically rbcord highlyamplified, transievitelectrical signals from the nerves of animals such as frogs or crayfish. The sig
nals are observed on oscilloscopes, and arecurrent difficulty in the past has beenthat the transient signals are difficultfor students to measure or even observe.Storage oscilloscopes are helpful, butthey are expensive and they do not providehard copg. Routine photography is slow.awkward, and expensive, particularly ifPolaroid film is used. Microcomputers, onthe other hand, offer the possibil.ty ofdigitizing signals and playing them backfor analysis. Playback can be repetitiveand fast for observation on an oscilloscope. or it can be oneshot and slow forwriting out on a chart recorder. ln addition, in many experiments signals are recorded from single nerve cells rather thanfrom groups of tens or hundreds of nervecells. ln such cases, the information ofinterest is the time interval between in-dividual electrical impulses (action potentials) in a nerve call. Histograms ofinterpulse intervals and post stimulusfiring rates are routinely computed in research laboratories: we can now make suchtechniques available to undergraduates aswell. Finally, some neurophysiological experiments (such as recording evoked potentials from the scalp) require signal averaging tecause the signals are much smallerthan the background electrical noise. Onceagain, microcomputers can makesignalaveraging tecnniques available toundergraduates.
Digitizing transient signals and signalaveraging both require analogto digital conversion, but presently availableAtoD boards that interface to a parallelport are too slow for use in a neurobiology lab. An action potential can be asbrief as 1 msac, and if its shape is to bpreserved in digitization, at least 10samples should be taken. This requires asampling interval of 100 usec. Analog Devices (P.O. Sox 2110, Norwood, MA 02062)manufactures an integratedcircuit, SbitAtoD device that can make one conversionevery 25 usec and that costs less then$25. I have used their device in the tnalog converter board whose block diagram isshown in Figure 5. The converter requirestwo control lines, one to start conversionand one to signal that the data are ready,plus eight date lines. These line: Oraconnected to Port A and to CA1 and CA2 ofthe AIM. Port S is connected to adigitaltoanalog converter, for playbackof digitized data. The Port S controllines, which are not otherwise needed, areused for input and output of triggetpulses.
My interactive program for A/D conversion, like the HiPad program, will reside
riga an EPROM end will make the microcomputer sees like smart instrument to theuser. The major input routine providescontinuous diiititton with scrollingdisplay, so that the oscilloscope screenresembles moving chart. Receipt of atrigger pulse causes digitization to ceaseafter' a preset interval, followed by continuous replay of the lest three pages ofdata tthree pages are about the limit fo.
flickerfree displeya. The user eon elsechoose to output the digitized data slowly, for playback to chart recorder.Chart recorders have frequency passbandsfrom DC to 100 Hz, at most, while the original signal has 'portent components upto about 5 kitzs thus an espnsion of thetimescale by a factor of i00 is necessaryto writs the data /accurately on a chartrecorder. The timescale can be almond/adeasily by inserting a timing loop in theplayback program.
Digitization at 10,000 samples persecond would appear to require huge amounts of immoral, but a comparison with theoscilloscope sweep speeds that aro used bynevrobiologisfs to display data shows thatthe memory requirements are not excessive.-A single sweep typically displays from 10Mee to 0.5 sec of data so that 194 bytesof memory would hold two to eighty swims,In *any cases, when longer sweeps of dataare of interest, high rates of figitization are not needed, and the effective capacity of the memory can be extended.
Thus, in summery, the relatively lowcost of icroceputers opens menu new possibilifiee for using them in Cie laboratory. The programming effort that is required can be extensive, end some of theinterface hardware may have to be constructed, but the results give studentsenormously greeter cap.bilities for nalyzing data. It is !.portent that undergraduates in science receive experiencewith such techniques, for they wiil workin world in which computerized canon.-tion and analysis of data will be routine.
1111=11migr4amr.
SAW IN
Science and Engineering 85
pm A
CAI
CA2
Cal
'tent 5.
96
86 NECC 1980
CLASSICAL MECHANICS WITH COMPUTER ASSISTANCE
A. Douglas Davis
Department of PhysicsEastern Illinois University
Charleston, IL 61g20
-This paper describes the use ofcomputer and elementary numerical
analysis in the teaching of a coursein classical mechanics. This is atraditional, rigorous, calculus-basedmechanics course. The use of-thecomputer allows students to solveproblems somewhat before the analyticalsolution is developed. Such priorsolution allows them to anticipatethe analytical solution and greatlyaids in their understanding. Computer-generated solutions also allow theinvestigation of interesting problemswhose analytical solution would other-wise be beyond the scope of thiscourse.
INTRODUCTIONThis paper describes the use of
a computer and elementary numericalanalysis in the teaching of a coursein classical mechanics. This is nota computer-based mechanics course.Rather it is a traditional, rigorouscourse in classical mechanics. Computer-generated solutions are used as simplyone more tool to teach the real physicsof the situation.
The first week (three SD-minutelecture.) is devoted to teaching"conversational BASIC" just enoughBASIC that even students with no priorexposure to computers can go to aterminal and write the simple programswe shall use. Thus, students canimmediately write, run and change theirown BASIC programs. The TAB functionin BASIC allows even neophytes toobtain graphic results very quickly.In addition, the results of someinstructor-written computer programsare used. Both student-written andinstructor-written programs give studentsanother tool-- another point of view,enother handle -- to use in developing
physical intuition and a solid under=standing of what's going on in a givensituation.
Computers are used in the followingareas:1. Introduction to variable forces.2. Introduction to integral calculus.3. Investigation of harmonic motion.4. Alternate approach.S. Central force orbits.
INTRODUCTION TO VARIABLE FORCES
In any good, solid introductoryphysics course students will have solvedessentially all of the basic probltisinvolving motion caused by a constantforce. Variable forces require the useof calculus for a solution and areusually not covered in detail in anintroductory physics course - even if itis nominally calculus-based.
Harmonic motion is the motion of abody under the influence of a linearrestoring force and can be written asF X where F is the force, X is theposition, K is the spring constant whichdetermines the strength of the force, andthe negative sign indicates that theforce always acts to move the body backto the origin. A classic example is abody of mass M attached to a spring withspring constant X. But harmonicoscillators have more wide spread usethan that. A thorough understanding_ ofharmonic oscillators is useful, evennecessary, for understanding such diversethings as automobile suspension systems,radio receivers, and ultra-violetabsorption by the atmosphere.
A major foundation of classicalmechanics states that a force acting ona body will cause it to accelerate. Thisacceleartion is directly proportional to
r
the force and inversely proportional tothe mass and can be written in the formof a = F/m or, as is more usually done.in the form of F ma. Beginning withthis in the second week of class studentswrite a simple iterative program to solvefor the motion of a harmonic oscillator.The essentials of the program are:100 LET F K * X110 LET A = F/M120 LETVV+A* D130 LETX=X+V* D140 LETT=T+ D160 PRINT T, X, V160 GO TO 100where F is the force; K, the springconstant: A, the acceleration; V, thevelocity; X, the position; T, thetimes and 0, the time increment T.
Pditional details of the program allowD to be small for greater accuracy yethave only a manageable amount of datato be printed out. A more sophisticatednumerical analysis routine, like the Runga-Kutta method, could be employed (I).But since the object of all this is' tounderstand the physics of the motion ratherthan extreme numerical accuracy, thesimpler method seems preferable. Whileclassroom discussions are usuallylimited to writing a program in BASIC,this procedure is readily adaptable to ahand-hald programmable calculator (2).
The TAB function in BASIC allows astudent to get graphic output readily bychanging the PRINT statement to:160 PRINT TAB (40 + 30 * X): "*"which centers the output on column 40when X 0 and has a scaling factor of 30to PRINT an asterisk in column 70 orcolumn 10 if X has a value of +1.0 or-1.0. If considerably larger or smalleramplitudes are expected the scaling factoris changed accordingly.
Such graphic output allows thestudents to see the motion -- andinvestigate its dependence upon Seriousparameters -- in some detail before webegin the rigorous analytical solution ofthe same peobiem. Knowing more aboutthe behavior of a system makes findinga mathematical solution all the easierand more meaningful once it is obtainedby more traditional means.
INTRODUCTION TO INTEGRAL CALCULUSThe fact that an integral represents
the area under a curve is of vital impor-tance in physics. Yet students sometimescomplete the average course in integralcalculus knowing an integral as simply an
Science and Engineering 87
operation, the antidifferentlation of afunction. To stress the idea of anintegral as the area under a curve or asan infinite sum, a homework problem Isassigned that asks for the sum of theareas of small rectangles bounded by thequadrant of a circle. The quadrant of acircle can be broken into five, ten,perhaps 20 or even 100 small rectangleswhose individual areas can be calculatedby hand. As more and more smaller andsmaller rectangles are used, their totalarea comes closer and closer to theactual area of a quadrant of a circle,wrZ /4 or 0.71164r2. An integral, since itis a sum of an infinite number ofInfinitesimally small rectangles givesexactl that result. To make this pointum* s a ably clear, the students areasked to continue by breaking thisquadrant of a circle into 100, then 1000,and finally 10,000 tiny rectangles. Thisproblem would be unrealistic and futileto attempt by hand. but is an easy andinteresting problem for the computer.
INVESTIGATION OF HARMONIC MOTIONF = -KX describes a simple harmonic
oscillator. Addition of a frictionaldamping force turns this into a much morerealistic damped harmonic oscillator.Such a damping force might representmass and spring wiggling under water, orin cold molasses, or the resistance In aradio or the shock absorbers on a car.Its inclusion drastically changes thetechnique and approach necessary for ananalytical solution. But long before thestudents are concerned with the detailsof an analytical solution, they haveamply investigated the behavior or motionof this damped harmonic oscillator. Theonly change to the earlier computerprogram is to redefine the force, includ-ing the damping force.100 LETF=-K *X-C* VWith this change, the students can nowinvestigate the motion for various initialconditions and various values of thedamping coefficient C. Underdamping,critical damping, and overdamping becometerms of real significance describingcertain particular and characteristicmotions of the oscillator.
Resonance Phenomenon .or the behaviorof a driven or forced harmonic oscillatorhas application throughout technology.Tuning of a radio is but one example.Long before the students hear the ominousphrase "inhomogeneous second-orderdifferential equation," they will havelearned much about the characteristics of
96
88 NECC 1980
solutions to Just that. Again, onesimple change to the initial computerprogram is all that is required. Theequation defining the force now becomes100 LET F = -X*X-C*V+E*SIN (10T)where E is the strength of an externaldriving force which varies sinusoidallywith angular frequency w. Resonancecan clearly be seen and investigated bychanging various parameters and observingthe effect those changes have on theoutput.
ALTERNATE APPROACHSome very interesting real-world
problems are difficult to solve analyt-ically. Others can be solved analyticallywithout too much difficulty, but under-standing the full meaning of the solutionmay not be entirely clear to students.For both situations a computer-generatedsolution is a very useful alternateapproach.
It is an easy matter to discuss thetrajectory of an object thrown and thenacted upon by Earth's uniform gravita-tional field -- as long as frictionthrough the air is neglected. That's areasonable approximation for many sit-uations. But it is also a ratherinteresting problem to include airresistance and then inviii1WITIUtrajectory. Air resistance is notneglegible if a student throws acrumpled wad of paper at a waste canor if a naval cruiser fires a shellat a target ten kilometers away.
It isan easy matter to return tothe original computer program and modifyit to handle a ballistic trajectory withlinear air resistance:100 LET Fl = -C * V1105 LET F2 = -C * V2 N * 0110 LET Al Fl /N115 LET A2 = F2/I1120 LET VI V1 + Al 4 0125 LET V2 - V2 + A2 * 0130 LET X = X + V1 *0135 LET Y a + V2 * 0140 LET T = T + 0150 PRINT T, X, Y160 GO TO 100C is the air drag coefficient, G is theacceleration due to gravity, the quantitieswith a 1 suffix refer to horizontal orx-components, and quantities with a 2suffix refer to vertical or y-components.This program yields a data table of hor-izontal and vertical positions whichstudents are asked to plot in order to seethe trajectory. Initial conditions ofmuzzle velocity and firing angle are
varied; the air drag coefficient is alsovaried.
CENTRAL FORCE ORBITSGravitational forces, electrostatic
forces, and forces on an isotropicharmonic oscillator are all examples ofcentral forces, forces which depend onlyupon an object's distance from a certainorigin. Central forces occur throughoutnature. Any course in classical mechanicswill spend a considerable amount of timeand effort investigating central forcesin general and the inverse-square law ofNewton's Law of Universal Gravitation inparticular. The orbits of interplanetaryspace probes, comets and planets all areboth important andinteresting.
Again, the computer can enhance thisaspect of classical mechanics. Forexample after the analytical solutionsare derived and discussed, e desktopmini-computer with attached x-y plottercan be brought into class. As the classwatches, orbits are drawn for variousinitial conditions. The conditions fora circular orbit become obvious, andthe real meaning of escape velocity ismade entirely clear.
This computer and plotter also allowthe students a glimpse at how planetswould move if the universe had beendesigned differently. Orbits for a forcethat has a radial dependence of
1 or 1 orrz.9
or whatever one likes can be handled justas readily as the 1 of the real force.
This always sparks'students' interest andleaves them with an experience somewhatmore tangible than just the discussionof an equation.
CONCLUSIONClassical mechanics forms a very
important foundation in the education ofphysicists and engineers. And almostInvariably, it is considered difficult bystudents. Both student-written andinstructor-written computer programs offeran additional tool to aid students inunderstanding mechanics. Students findthe computer enjoyable (even those withlittle or no background who also findits use challenging). They report thatits use greatly aids them in gaining athorough understanding of the analyticalsolutions, for as they see how a system
99
actually responds, the mathematics even-tually derived to describe that motionbecomes much more meaningful.
REFERENCES(1) A F. Vierling. "Harmonic Notion".olosics: An
Pulcolomifommision on CollegePoysits, 1969, Page 35.
(2) R. Eisberg, Applied MathematicalPhysics ienc ngtfa cu a ors, c raw- 9
i 0
Science and Engineering gg
90 NECC 1980
COMPUTER. AUGMENTED VIDEO EDUCATION INELECTRICAL ENGINEERING AT THE U. S. NAVAL ACADEMY
Michael W. HageeTian S. Lim
Richard A. PollakUnited States Naval AcademyAnnapolis, Maryland 21402
(301) 267-3492
ABSTRACTThis paper describes a computer-
augmented video education (CAVE) projectbeing undertaken in the Electrical Engi-neering Department at the United StatesNaval Academy. The project is designedto produce a series of modules involvingcomputer graphics display and integratedcomputer-controlled television to helpengineering students (non-electrical) aswell as non-engineering students learnthe essentials of electrical engineering.Each module contains a quick recap of thetheory and basic sample problems,followed by an exercise to test students'proficiency.
I. INTRODUCTIONAlthough there are different areas
in which a midshipman at the UnitedStates Naval Academy can major (fromnuclear physics to English literature),all midshipmen regardless'of major arerequired to take a two-semester surveycourse in electrical engineering duringtheir second class (junior) year. Twodifferent two-semester courses areoffered. One is the core course programfor non-engineering majors and the otheris the engineering core course programfor engineering majors. Both coursescover basically the same material atslightly different levels. Both coursesare a mixture of theory and practicalwork and cover just about every majorelectrical engineering area.
This unique requirement for allStudents to successfully complete atwo - semester college level electricalengineering course has presented sometime-consuming problems. In the past,one-onone extra instruction (El)sessions have been the main tool inhelping those non-technically orientedstudents.
El sessions involving an instructorand one or two students have always been
an important part of the learning processin the Department of Electrical Engineer-:ng and is in part a philosophy of teach-ing at the Naval Academy. The somewhatslower or weaker students find this typeof aid essential in any course in whichapplying basic principles to solve prob-lems is a primary objective and also usedas a measure of achievement. These EIsessions were found, not surprisingly, tobe remarkably similar. They begin with aquick recap of pertinent theory followedby the solution of a basic problem apply-ing the principles just reviewed. As time_permits, the problem solving is repeatedwith variations in the hope that thestudent builds both,experience and con-fidence. In fact, the confidence of thestudents that they can succeed by apply-ing the different. principles td slightlydifferent situations is of paramountimportance in such a course. A typicalRI sessions is also characterized byusually covering only one or two of thestated learning objectives of the program.Sessions covering the same material arerepeated with different students. In thissense they are inefficient and tax theinstructor not necessarily just for thetime involved but also because even askillful and patient instructor can findit difficult to maintain enthusiasmthrough constant repetition.
II. DESCRIPTION OF COMPUTER-SUPPORTEDINSTRUCTION SYSTEMThe Academic Computing Center at the
U. S. Naval Academy has developed acomputer-supported instructional system(CSIS). It is intended for use as adrill/practice/tutorial tool to aid USNAinstructors in computer presentation oftheir course material. The same learningmaterial can also be used as a test. Onefeature allows the instructor to build anexercise in basic building blocks calledframes. A frame may be in the form of a
101
question, a comment, or a help sequencesor it may be a means of letting thestudent control branching. The frame-type designator determines how the com-puter is to analyze the student's response.if such a response is necessary. Anotherimportant feature is that frames may c'ntainnumeric and string variables and functionexpressions -- all under the control ofthe author/instructor. This featureallows the instructor to present the samefunctions (question types) to allstudents, but each studenk. gets a dif-ferent set of values for the variablesand functions (questions) used in theframe. Furthermore, the instructor cancontrol such things as the number oftries a student is allowed for answeringa question correctly, whether the framecontains graphics, the number of times aframe is to be presented with differentvariables substituted, and allowablecorrect answers.
A team of eight instructors from theDepartment of Electrical Engineering iscurrently working on a project using CSIS.The current project will incorporatethose phases of aid to students intopackages or modules of material availableat any time in the Academy's audio visualcenters.
The project team is producing aseries of modules of video and computermaterials which could be used by studentsas a self-help, extra-instruction tech-nique. Bach module is keyed to one ofthe recurrent stumbling blocks encoun-tered each year by non-engineeringmajors in their first course in electri-cal engineering. These modules arebeing designed to replace some of theregular extra-instruction tutored/taughtby the faculty.
A study of past exams, studentevaluationst and faculty questionnairesindicated that all students, from thesuperior to the weaker, experiencedcommon stumbling blocks. Interestinglythe problem areas were common to. boththe engineering and now.engineeringstudents. These areas covered suchtopics as network reduction, Thevenin'sTheorem, Norton's Theorem, voltage/current division, RC and RL transientanalysis.
It is our conviction that the com-puter is most significantly exploitedas an educational tool when it is usedin conjunction with other activities.All module design work is approachedwith the ide.,41 of complementing classwork with practical drill and practiceexercises which reinforce the majorlearning objectives.
Science and Engineering 91
All extra-instruction modules aredesigned to give initially a brief (oneor two frame) review of the objective inquestion. However, the heart of themodules is the drill and practice frames.These frames can select problems andcircuits at random and through the useof predictable wrong answers (PWA) guidethe student through the necessary stepsto a solution. The instructor/programmercan incorporate video anywhere in theprogram, from using educational televisionfor the initial introduction to havingcertain segments serve as help sequencesfor the PWAs.
the
both the graphicsterminal and the video cassette playerare completely controlled by the computer,the student is relieved of all coordina-tion tasks but is still an active part ina multi-media presentation. A controlinterface was designed at the NavalAcademy to take signals from the author'sprogram and physically control the videocassette player. There are commands suchas "GO TO" and "PLAY." These commandstransport the video tape under author/instructor or student control. Thecomputer really acts as a coordinator.The student, not the computer, can controlthe instructional flow. This controlresults both from answers to presentedproblems and from input to periodicdecision frames. Figure 1 shows anarrangement of the color televisionmonitor, color video cassette player, andthe computer graphics display terminal.
In what follows, we describe some ofthe computer-aided NI programs.
A. Voltage/Current DivisionFigure j2 depicts a typical question
frame block diagram. Figure 3 reflects aspecific question frame in the voltagedivision module. As the student entersthis particular frame, the program assignsrandom values from previously stored filesto all circuit variables. The componentacross which the voltage drop is to becalculated is also randomly selected.
Based upon the values assigned andcomponent selected, the program calculatesthe correct answer, six predictable wronganswers (PWA), and one unpredictable wronganswer. The student then enters hisanswer.
In Figure 4 the user has entered avalue of 65, and the computer has re-sponded with the unpredictable wronganswer message. In Figure 5 the studententered the correct magnitude but had thewrong polarity which caused one of the PWAmessages to be displayed. Finally inFigure 6 the student has answered correct-ly and received an appropriati congratula-tory response.
102
92 NECC 1980
FIGURE 1
f4;r,
7.1 -4 -tt
The PMAs are .sot limited to a one-or two-line message but can be expandedinto several supplementary frames or avideo presentation. Upon completion ofthese help sequences the user is usuallygiven the same or similar questions toteat his understanding.
After a successful answer, theprogram checks to see whether thestudent has received the required numberof presentations. If not, a new circuitwith different components and valueswould be randomly selected and presented.
If the number of presentationcriteria is satisfied, the programdetermines whether the user has answereda predetermined number of questionssuccessfully to continue to the nextobjective. If not the student is sentto another series of help sequences:otherwise he is directed to the nextobjective.
All variables, circuits, com-ponents, number of presentations, andnumber of attempts are under the author'scontrol.
0. RL Transient ModuleThis program consists of 28 frames.
As an example, suppose the student hasreached Frame 8 and is presented with the
circuit as shown in Figure 7. NO isasked to answer the questions and thepicture remains on the screen until hehas provided an answer. If the answer iscorrect, he moves on to the next topic.If the answer is wrong the first time,he is given a second chance. If hissecond answer is still wrong, he isgiven three choices: (1) mare review,(2) take the test again, or (3) stop.This information is given in Frame 13 asshown in the Appendix. All he has to doto make the choice is to type in REVIEW,CONTINUE, or STOP. If he chooses REVIEW,or CONTINUE, an appropriate picture willappear on the screen to provide the infor-mation he has asked for. If he choosesto STOP, he will be asked if he wishes tocomment on the program. He can eithertype in "NO" (no comment) or make acomment and then sign off.
III. SUMMARYComputer-augmented video education is
a highly interactive process that is dif-ficult to describe verbally. The bestunderstanding can really only be achisovedby sitting at a carousel/terminal andrunning the module. A good computer-basedinstruction system must satisfy severaldifferent communities: the educator-
103
programmer, the instructor, and mostimportantly the student.
Student response to the first set ofmcdules has been heartening, as a largelumber indicated that they found themodules to be helpful and enjoyable. Manystudents tried the programs after hearingthat they were of help: on the nightbefore each major exam approximately 258of the class worked one or more modules.W. feel it is important to develop pro-grams that will be challenging and com-pelling and that will encourage studentsnot only to try other topics, but alsoto recommend the programs to otherstudents.
The twelve modules that were usedcovered material of about 1/4 of thecourse, and approximately 758 of thestudents in the course used one or moreprograms. It appears that the availabili-ty of additional topics would have re-sulted in a greater percentage of thestudents using the modules, since othertopics were suggested by both users andnon-users. The advisability of provid-ing a complete set of supplementaryprograms is being studied at this time.
A total of six sections out of 45were randomly selected to authenticateeach module. Three of these sectionswere designated control groups while theremaining three were the test groups.Although data are still being analyzed,initial indications point to a highcorrelation between the test groups andthe higher test averages. (The completedata analysis will be available by thetime of the conference.)
There were two major problem areaspointed out by the students. The first,not surprisingly. was delays caused byslow computer response. The chainingprocess used in the display of graphicscaused some fairly large delays especial-ly when the number of users on the systemwas in excess of 100. This problemmust be dealt with (it will be studiedthis summer ), since these delays mayhave prevented some students from usingthe programs and could dampen enthusiasmfor any interactive computer materials.
The second most often recordedcomment was a pedagogical one. Thestudents had a strong desire to see theschool solution to a problem. This wastrue even though they may have answeredthe problem correctly. There was alsoa strong preference for seeing thespecific problem missed worked in thehelp sequence rather than seeing a,general review of the theory involved.
Through the use of computergraphics display terminals, wellengineered and flexible software, and
Science and Engineering 93
a system with a demonstrated high reli-ability, we believe we have developed apedagogically sound and interesting setof modules that will complement theregular classroom instruction. Thestudent can use the system to increasebasic electrical engineering skillsthrough drill and practice. The in-structor can anticipate to spend lesstime on Er sessions and therefore candevote more time to students with seriousdifficulties.
This type of customised instructionoffers the added benefit of diagnosingstudents' specific problems in givenconcept areas as well as measuring vari-ous teaching methods, since one canmonitor the responses/comments ofstudents to the given material.
REFERENCES1. R. Pollak and J. Schwab, "uSNA -
Computer Supported InstructionSystems Sophisticated, Generativeand Easy to Author," Proceedings ofthe 16th Annual Meeting of Associa-tion for Educational Data Systems -Higher Education. 1978.
2. M. B. Sousa. "Computer AugmentedVideo Education." EducationalTechnology. FebruaiiWW7W. 46-48.
FIGURE 2
IndividualHelp
Sequence
N
Fail(Help)Frame
NextSuccessFrame
94 NECC 1980
FIGURE 3
QUESTION #1 FRAME 51
You will be given three chances to answerall questions correctly. If at any timeyou feel you would like to review thevoltage divider technique concept, type"00" when a question is asked.
(The calculator mode may be entered bytyping "$CS ".) --Wait for the circuit.- -
You will be given three chances to answerall questions correctly. If at any timeyou feel you would like to review thevoltage divider technique concept, type"00" when a question is asked.
(The calculator mode may be entered by.typing "$CA".) --Wait for the circuit.- -
I'm sorry butyouranswers don't agree.I think your approach is incorrect.Please TRY AGAIN or type 00 to review.
V1/4A.
R3CIRCUIT 7
FIGURE 5
QUESTION #1 FRAME 51
You will be given three chances to answerall questions correctly. If at any timeyou feel you would like to review thevoltage divider technique concept, type"30" when a question is asked.
(The calculator mode may be entered 1typing "#QA".) --Wait for the circuit --
I'm sorry but our answers don't agree.I think your approach is incorrect.Please TRY AGAIN or type 00 to review.ANSWER? 4.9
You have the correct magnitudewbut thewrong polarity. Try Again.
1 05
CIRCUIT 7
FIGURE 6
QUESTION #1 FRAME 51
You will be given three chances to answerall questions correctly. If at any timeyou feel you would like to review thevoltage divider technique concept, type"00" when a question is asked.
(The calculator mode may be entered bytyping "SCA".) --Wait for the circuit.- -
I'm sorry but our answers don't agree. Ithink your approach is incorrect. PleaseTRY AGAIN or type 00 to review. ANSWER-1.91
You have the correct magnitude but thewrong polarity. Try Again. LAST CHANCEON THIS QUESTION. 21.9 GROOVY(
it R3
R471.
1
FIGURE 7
QUESTION #5 FRAME 8
TEST 1
Consider the following
1
R2
E-=
circuit.
2
CIRCUIT 7
Let E*10 VOLTS R1=3 OHMSL=30 mH R2=6 OHMS
The switch has been in position 1 fora long time (at least 5 time constants).What is the current through theinductor?
ANSWER?
Science and Engineering 95
APPENDIX
You have missed the problem.
You have 3 choices as follows:
1. Type REVIEW for more drill.
2. Type CONTINUE if you would like totry the test problem again.
3. Type STOP to stop.
ANSWER?
1U
Structured Programming
USE OF PROGRAMMING METMCDOLOGYIN INTRODUCTORY COMPUTER SCIENCE COURSES
Elizabeth AlpertHartnell College156 Homestead Ave.Salinas, CA 93901408 758-8211 x 431
INTRODUCTION"Twenty-five years ago our job was toinstruct the machine; now it is themachine's job to execute our program."
E.W. Dijkstra
In August 1979 I participated in theNinth Annual Computer Science Instituteat the University of California, SantaCruz. The institute was dedicated to Pro-gramming Methodologya one-week intro-ductory course followed by a two-weeklecture series. The speakers were animpressive group - -many of the members ofIPIP Working Group 2.3 plus several otherprominnt computer scientists. I expectedto receive information I could use in theclassroom, but I did not anticipate thatthe material would lead to a total re-evaluation of what and how I was teaching.
The nature of computing is changingrapidly. The proliferation of microproces-sors has spread the availability of com-puting and created an increased need forwhat Wirth (1) calls "programming know-how." In addition, advancements in micro-processors and LSI technology have pavedthe way for the development of newprogramming techniques. Economizing onmemory space or eliminating an instructionor two are antiquated programming goals.The design and implementation of correct,
96
readable, and understandable programsshould be the goal of modern programmers.
Enormous contributions have been madetoward reaching this goal. It is the re-sponsibility of computer science educatorsnot only to lecture about methodology butto incorporate the new methodologiesthoughout the computer science curriculum.The purpose of this paper is to discussthose aspects of programming methodologythat have particular relevance for computerscience education and to make some sugges-tions for their incorporation early inthe curriculum.
THE SCOPE OF PROGRAMMING METHODOLOGYThe major objective of programming
methodology is to increase the program-mers* ability to design and implementprograms. Programming methodologyaddresses problem solving techniques,program reliability and adaptability, pro-gram correctness and structure, guidelinesfor partitioning large program tasks, andsoftware tools. The computer sciencestudent, whether in a four- or two-yearprogram, must develop an appreciation ofwhat methodology is and how to use it.
PROGRAM DESIGN
Modularization....a separation of concerns."
E.W. Dijkstra
Most programming problems deal withmore than one issue. Thinking about allthe issues at once is confusing anddifficult to do. By separating the issuesand dealing with the complexities one byone, a program becomes naturally parti-tioned into modules. Programs written ina modular structure are easier to thinkabout, modify, maintain, and understand.
How to partition a problem is notnecessarily obvious, and therefore it isdifficult to teach. What should be taughtis that partitioning a problem is thefirst step in problem solving and programdesign, and is a worthy goal in itself.
Most introductory textbooks discussmodularization, but do not include enoughexercises to provide students with anopportunity to apply it. It is up to theinstructor to develop sophisticatedproblems
problemsimple to solve
after the has been partitioned.The whole cycle of changing, adding,deleting, and exchanging modules should bemade a part of all programming projects.
SpecificationSpecification is a statement of speci-
fic requirements. It is a process that isindependent of composing the program andis not part of the solution. Specifica-tion provides a means of communication,and in fact should minimize what theprogrammer needs to know. To be useful,specifications must be precise and well-structured. Parnas (2) believes thatspecifications stated in terms of exter-nally observable phenomena are preferableto specifications given as another pro-gram because they are user-oriented andunderstandable even by the non-programmer.They do not leave room for guessing aboutthe requirements nor do they suggestparticular implementations. There arevarious techniques of formal specificationthat appear in the programming methodologyliterature, but regardless of the methodthe common guidelines, as suggested byParnas, are:
state everything that is requiredstate nothing that is not requiredleave no room for doubt.
It is essential that students learnto appreciate the necessity of formalspecifications and learn a well-structuredspecification methodology. Practicalexperience in writing formal specifica-tions should be required of some program-
Structured Programming 97
ming projects. Most introductory textbooksprovide completely specific problems. Thestudents are even given the input/outputformats. It would be a worthwhile exer-cise for instructors to assign practicalproblems with which the student hasfamiliarity--e.g., registration process-ing, payroll- -and require the student todo a complete specification of a problem.
ALGORITHM DESIGN AND DESCRIPTION"Nice descriptions of neat algorithmsare not gifts from heaven; they areformally designed."
Dijkstra and Gries
Once a program has been formallyspecified, algorithms can be designed foreach of the parts. Before the algorithmis designed, however, it is necessary toselect a notation to describe the algo-rithm. The notation must be concise,understandable, and formal.
An Introduction to Formal DescripticnsOne of the most powerful tools for
algorithmic design is the notation (orlanguage) developed by Dijkstra (2). Itis essentially a calculus for the deriva-tion of programs using predicate trans-formers for each statement. A predicatetransformer is a rule for deriving from aresult predicate (post - condition] a predi-cate corresponding to the set of statesthat ensure that the statements will termi-nate and establish the truth of the post -condition. The statementlin Dijkstra'slanguage include: Ale, abort, assignment,and the guarded construcii-IT---fi, do---od. A guard is a Boolean ail-reel-Ion at theVied of a statement.list; a list may beexecuted only if the corresponding guardis true.
The statements are defined by theirpredicate transformers. The mathematicalsemantics of a statement S and any post-condition Rare given by the weakest pre-condition such that S establishes R. Thisis denoted by wp($,R).
For the assignment statement thepredicate transformer is
wp( "x := E", R) = R-x:=Ewhere R._
ag:=Edenotes the predicate obtained
by substituting all'frae occurrences ofx in R by some expression E. A defini-tion of the predicate transformers for -
all the statements can be found in (3).A description of the c-erational semanticsfor each statement V" Lows.
The "skip" stay sent is executed bydoing nothing. The "abort" statementsignals failure. An assignment statementis denoted as "x : E" where x is a
1 u
98 NECC 1980
variable and E an expression. The semi-colon denotes sequencing; the execution of"Si; S2" will result in Sl being executedbefore S2. The "TF" statement is denotedas
if B1+ SL
111 B2 + SL
2
Bn
-to SLnfi
where the Bi denote guards and the SLi arestatement lists. In the execution of IFany one of the SLi will be executed if thecorresponding Si is true.* The Q "bar"acts as a separator between otherwise un-ordered alternatives. The "DO" statement,denoted by
do B. SL1
B2+ SL
20
0 Bn
SLnod
continues to executeany one of the alterna-tives whose guard is true until all of theguards are false.
Some Examples of Formal DescriptionsDescribing algorithms using this nota-
tion can be accomplished through somefaifly simple formal methods. Let us con-sider a situation where it is desirable toestablish the result as a = 7 where a is aprogram variable. The program a := 7 willestablish the truth of the result sincethe initial state is true (i.e., 7 = 7 isT, and w( "a := 7", a = 7) = T). The pro-gram a := 6 will not establish the truthof a = 7 since there is no initial statethat will guarantee it (i.e., 7 = 6 is F)and wp( "a := 6, a = 7) = F).
Let us next consider the problem ofestablishing, for a given x and y, a resultwhere m, a program variable, contains thelarger of x and y and either x or y when
= y. For the result to be true whenm x, x must be > y. For the result tobe true whennt= y, y must be > x. Theconditions x > y, y x become the guardsfor the program statements.
if x2 y + m := x
0y,x+m:= yfi
*The IF aborts if no guard is true.
Since one of the two conditions must betrue the prbgram will never abort. Notethat if x = y both conditions will betrue, and it is indeterminate which state-ment will be chosen; nor does it matter(3).
Programs using the do---od iterativeconstruct can be developed in much thesame way. Let us consider developing aprogram to establish a result R thatstates that S is the sum of the elementsin an array A of n elements where n > 0.Obviously, the establishment of the truthof R requires a loop structure. What mustbe found is a generalization of R, i.e.,a relation P, that can be easily made trueinitially and that can be held true duringthe iterative process. In generalizing,the variable i replaces the constant n.Thus, P can be defined to be the relationthat S is the sum of all elements in thearray up to A (i) and 0 < i < n. Usingthis generalization, the result, R, isestablished if P is true and i = n.
Since P must be true before executingthe loop, conditions must be found thatestablish this state. Setting i to 0 andS to 0 will establish P - -the sum of noe lements is zero. During the iterativeprocess, computational progress must bemade to ensure termination. Stepping itowards n by increments of 1 will do that.Keeping P invariant then requires addingA(i) to S after incrementing i. All thatremains is the choice of the guard forthe loop. Since i = n is the conditionfor termination, i 0 n must be the condi-tion for remaining in the loop. Theprogram has now been derived:
S to 0; i := 0;
do i 0 n +
i pal i + 1; S := 5 + A(i)
od
A more robust program results whenthe guard is weaker. A guard i 0 n isweaker than i < n. If, for whateverreason, n were less than zero, the loopguard i < n would cause an immediate skip,and termination of the loop with nonotification of error.
A common solution to this same prob-lem is the following
S im A(1), i tm 2
do i 0 n + 1 +
S ;NB S + A(i); i i + 1
od
100
The second solution is more efficient--onepass through the loop has been eliminated.However, for n <:1 the second program isnot correct. The program could be modi-fied to the following equivalent program:
S A(I); i 1
do i 0 n
i + 1; S S + A(i)
od
but both will result in an infinite loopfor n < 1.
Reasons for Using Formal DescriptionsBy this time the reader might be
thinking that the examples have not shownanything new. That is exactly the point.The solutions may be the same but theyhave been arrived at through a totallydifferent process. It is the methodologythat is important as a program design tool.
These solutions could probably havebeen written Without the analysis, but formore sophisticated problems the programsolution may not be as obvious or intui-tive. And even if one were to arriveintuitively at the same program, that thedevelopment of the algorithm can beformally verified is significant.
By following the methods outlined byDijkstra, solutions to problems can bedevised and described in a notation that isquickly understood. The algorithms can beeasily discussed because of the simplicityof the language. Initial formulations ofalgorithms can often be *massaged" pro»ducing simpler or more efficient solutionswithout changing the assertion about theirvalidity.
Additional examples of this approachoan be found in the appendix. It issuggested that the reader solve the prob-lems first and then compare the solutionswith the algorithms.
The greatest challenge in computerscience education is teaching students howto design algorithms. The intuitive approachMay=work for some students but it certainlycannot be depended on. Illustrating thedesign of a few basic algorithms is-usually helpful, because the studentscan transfer the techniques to relatedproblems. Po- example, different algo-rithms for manipulating the elements in alist may share many of the sameproperties.
An initial reaction to the Dijkstranotation and the methods that are describedslush more formally in his book is that themethods are too mathematical and beyond thecomprehension of most beginning students.That charge may be true for the underlyingtheory and the Mere sophisticated problems
Structured Programming gg
whose elegant solutions can only be under-stood when derived formally, but for thetypes of problems introduced at beginninglevels the pre- and post-condition asser-tions as well as the iterative structurescan be defined in a less formal manner.Dijkstra, himself in fact, often doesjust that.
What is important is that students betaught to design algorithms and beencouraged to express those designs in anotation that is independent of the pro-gramming language used for implementation.When algorithms are designed with aparticular language in mind they areusually overly complicated and often in-correct. There is no great accomplishmentintraining students to think in ?OMAN(or any other programming language). Thereal goal should be to train students inhow to think analytically and providethem with tools that can assist them inthat task.
Other FormalismsAt this point the reader might be
wondering why there has been no referenceto flow charts, structured programming,or Stepwise refinement. The designmethodology that has been presented isactually an historical outgrowth of thesetechniques.
Twenty-five years ago, when program-mers were concerned with the details ofmachine execution, flow charting mighthave been a useful technique. Now,however, it is antiquated as a design tool.ThiVoit appropriate use of the flow chartis as documentation--a description of theexecution of the program. It is unreason-WarariXpect students to try to designalgorithms using flow charts. NOreover,they hardly ever draw the flow chartsbefore coding a program even wheninstructed to do so. They usually drawflow charts after the program has beencoded and is running. Instead of beingchastised for this, they should be praised.
Structured programming is more animplementation methodology than a designone. The constraints of structured pro-gramming for design purposes are fairlyprimitive. The Dijkstra methodology hasequivalent basic constructs, and strongerguidelines for combining them. Studentsshould certainly be taught structuredimplementation, but will find that imple-menting well-designed algorithms requiresvery little additional work.
Stepwise refinement still plays animportant role in design, especially forlarger problems. Dijkstra uses refinementin his notation when describing complexproblems. Hehner C41 has suggested addingto the Dijkstra notation a formal methodof refinement that eliminates the need
110
100 NECC 1980
for the do--cd structure; and may furthersimplify the notation.
Students usually are successful whenusing refinement techniques. That anEnglish phrase or descriptive term can beincluded in the design process(and itsprecise definition done separately), simplfies thinking about a problem. The prac-tice of writing down simple concise step*is one that instructors should follow inthe classroom whenever a problem is beingworked out. Too often students attempt tosolve problems by thinking in the syntax ofthe programming language when just a simpledescription suffices as a first step.
PROGRAM IMPLEMENTATION"Software is undoubtedly the majorsource of unreliability in mostcomputer systeks today."
J.J. Horning
Unfortunately, the design of a correctalgorithm does not ensure that its imple-mentation will be reliable. Whereas thecosts of hardware are plummeting, thecosts of sortware keep rising. Seventypercent of programming effort is spenton maintenance and much of that main-tenance is repair to unreliable or incor-rect software. In view of this situation,one of the important goals of implementa-tion should be reliability. This sectionof the paper presents some suggestionstaken from Horning (5) for creating morereliable programs.
LanguageThe goals of simplicity, under-
standability, and maintainability are asimportant for program implementation asthey are for program design. A simple,elegant algos.hm is easier to implementthan one that .ias not been carefullyplanned, but an effort must be made tomaintain the simplicity. This effect canbe achieved by choosing programminglanguage constructs that are simple andeasy to understand. Well-designedalgorithms can be implemented using onlya subset of a total languages
Another adVantage of choosing a sub-set is that the programmer can becomemore familiar with the constructs andtruly understand how to use them effec-tively. The use of a language structurethat is only vaguely understood can bedangerous.
Very often in beginning programminglanguage classes instructors cover as manyof the language structures as possible.This is actually a disservice to thestudents. A more successful approach wouldbe to throw out all the complex structuresand consider only the very basic ones.
The programmer who truly understands thebasiez will be more valuable than onewho thinks he/she knows all thecomplexities.
Documentation and StyleDocumentation and style are ath.r.
areas that can improve reliability. Docu-mentation should include not only adescription of what the program issupposed to do but also information re-lated to the design. Self - documentingtechniques such as explicit declarationsand choice of meaningful variable namesare helpful.
Developing a style of programmingshould go right along with developingformalisms. Indentation and formattingof the program text is a relativelysimple matter, yet it greatly enhances thereader's comprehension.
Most current textbooks provideexamples of documentation and style. With-out constant pressure from the instructor,however, there is little transfer of whatthe student sees in the text to theprogramming assignmerts.
RobustnessThe concept of robustness was Men-
tioned briefly in dealing with programdesign. It is relevant as well to programimplementation. P program is more robustif it functions properly when given awider range of input values. The inclu-sion of a few precautionary measures tohandle the events that aren't supposed tooccur can contribute much to the relia-bility of a program. These measures canbe language, machine, or data directed.At present there are no exception handlingschemes that are infallible, but neverthe-less, it is vital that beginning program-mers realize the importanca of suchconsiderations and develop an appreciationfor the need to make programs robust.
CONCLUSION"Well it Iprograsq works, it mustbe right."
Anonymous
This paper has presented certenconcerns of methodology in program designand implementation. These concerns shouldalso be the concerns of computer scienceeducators. The need for computer literacycourses is being stressed by college anduniversity administrations across thenation. Soon every college student,whether in a two- or four-year curriculumwill be enrolling in a computer scienceclass. The question is--what and how arewe going to teach these people?
For years, the philosophy has beenthat students need to get on the machinequickly. It didn't matter much what theydid there. If we are teaching people howto program, it does matter. Many studentstake only one course in programming anduse that experience in their futureendeavors. With the increase in smallbusiness and home computers this patternwill increase. The decision must be madethen whether these people should merelylearnt programming language or reallylearn how to program.
The methodologies described in thispaper are not difficult to understand orteach. They can be included in all elemen-tary courses now being taught. If they aretaught as part of the programming processso that students initially learn to dothings correctly, the quality of program-ming is more likely to improve.
Teaching all the details of any pro-gramming language is of transitory value.Languages change, machines change, newversions of software are released, and notwo manufacturers do everything the same.But teaching methodology has lasting value.
Just because a program works doesn'tmean it's right. What must be conveyed tostudents is that there is more to program-ming than just coding. They need todevelop an appreciation for. programs thatare simple, elegant, readable and robust- -the kind of program of which Dijkstrasays, "Ain't it a beauty!"
AcknowledgementsI would like to thank Professor
William M. McKeeman and the University ofCalifornia Extension for giving me theopportunity to participate in the Instituteof Computer Science at OC Santa Cruz. I amgrateful to Kenneth Friedenbach for hincontributions to the development of thispaper. I would also like to thankProfessor Fred Schneider of CornellUniversity for his comments on an earlierversion of this paper.
*2. Parnas, David L. The Role ofProgram Specifications. Universityof North Carolina at Chapel Hill.
*These were distributed at the Program-ming Methodology Lecture Series, Ninth AnnualComputer Science Institute, Universityof California, Santa cruz, August 1979.
Structured Programming 101
3. Dijkstra, Edsger W. A Discipline ofPr ramming. Prentice-Hall, Inc.1976.
*4. Hehner, Eric C.R. do Considered odA Contribution to the ProgrammingCalculus. University of Toronto.
*5. !WOW J. J. Effects of Program-ming Languages on Reliability.Xerox Palo Alto Research Center.
APPENDIXThe following are examples of algo-
rithms developed by formal methods pre-sented by David Gries during a five-dayIntroduction to Programming Methodologyclass at the University of California,Santa Cruz, August 1979*.
Example 1. Coincidence problemGiven a function f with M valueswhere
f(0) < < f(2) < < f(14-1)
and a function g when N values where
g(0) < g(1) < g(2) < < g(N1)
coast the pairs that: are the same.For example, for
f = 3,g = 1,
5,
3,
8,4,
12,7,
15,8, 11
18
the count c would be 2.
Solution 1. m := 0; n vs 0; c gm' 0;
do n 0 M and m 0 M
if f(s) < g(n) + m tr. m + 1
f > g(n1 * n vs n + 1
f (21) = g(u) -0 n := n + 1;
m g= M + 1;
C t= c + 1
fiod
The loop invariant is that c contains acount of the number of CO = g(i) where0 < i < m, 0 < j < n. Termination isttri condition-M = M, provided f(m-l) <g(n) or n = N, provided g(n-1) < f(n).
112
102 NECC 1980
Example 2. Different ValuesGiven a function f where
f(1) < f(2) < f(3)< <f(M), M > 1,
count the number of different valuesin f(1 :M).
Solution. 2. m 1= 1; c := 1
do m # M
if f (m) = f(m 1) 1- m 1m m + 1
f(m) # f(m + 1) +c := c + 1;m 1= m + 1
fiod
The loop invariant is that c contains 1 +the number. of f (i 1) # f (i) for1< i< a and 1< m e m.
The program may be simplified to:
a so c tat 1
do m # M
m 1m m 1
if f (m) = f (m + 1) 'skip"I f (m) # f (m # 1) y c := c +1
fiod
1'3
Structured Programming 103
FORTRAN 77: Impact on IntroductoryCourses in Programming Using FORTRAN
by
Frauk L. FriedmanDepartment of Computer and Information Sciences
Room 381 SpedumnslellTemple University
Philadelphia, Pennsylvania 19122(215) 787-1912
ABSTRACTA first course in teaching problem solving and
structured progtamming using FORTRAN is brieflydescribed. The features of the new FORTRAN 77standard which in the author's view impact mostsignificantly upon the course are summarised, andthe effect of those features upon the structureand content of the course is discussed.
INTRODUCTIONOa the third of April, 1978, The American
National Standards Institute (ANSI) approved anew American National Standard for the FORTRANprogramming language (ANSI 78]. This standard,designated as FORTRAN 77, is a revision of the1966 American National Standard FORTRAN. It isexpected to have a significant impact upon theuse of the FORTRAN language in a wide variety ofapplications areas. The new language should alsohave a considerable effect upon the use ofFORTRAN as a convenient language for teachingintroductory programming, since it provides anumber of significant pedagogic advantages overits predecessor.
The features of FORTRAN 77 that supportthese advantages are the subject of this paper.A summary description, with examples, of eachfeature is described, and the effect of the
feature upon the structure and content of anintroductory course is discussed. Before pro-ceeding to these topics, however, an outlinedescription of such a course is presented(see also EFK 77 and FK 78]).
COURSE STRUCTUREFigure 1 contains an outline of the topics
covered in the course, a time scale for thesetopics, and a list of problems that areassociated with each topic. The course isoriented around a set of two -dozen problemswhich illustrate a variety of problem-solvingtechniques. Most of these problems are solvedin their entirety, from the analysis and initialalgorithm outline, through to the final flow-diagram refinements and FORTRAN program.
Each problem is used to illustrate theapplication of a new feature of the FORTRAN
language. The feature Will normally have beenintroduced first, and a brief description of Itssyntactic form provided. The problem providesadditional motivation for mastering this new fea-ture since the problem solution would be much moredifficult without it.
Some of the problems shown in Fig. 1 emnhonifnskills that are fundamental to programming, such asfinding the largest value in a data collection,searching an array for a specified item, and sort-ing. Other problems relate to a variety of appli-cation areas: business-oriented problems (checkingaccount transactions and inventory control), games(bowling score computation and Tic -Tac -Toe),
By concentrating on problems that require theintroduction of additional features of the FORTRANlanguage for a reasonable solution, the motivationfor these features becomes readily apparent. Oncethe essentials of the festures needed to solve aparticular problem are introduced, class dis-cussion focuses on data description and algorithmdesign. It is occasionally the case that otherFORTRAN features are introduced as the algorithmi developed, refined, and finally implemented.The essentials of these features are described, andexamples of their use are given, usually within thecontext of the problem at hand.
SUMMARY OF NEW FEATURESThe new FORTRAN standard describes two levels
of the languages FORTRAN (sometimes referred toas Full FORTRAN), and Subset FORTRAN. Whereas theFORTRAN subset was previously described in aseparate standard (American National Standard BasicFORTRAN, ANSI 13.10-1966), the description ofSubset FORTRAN is now included in the descriptionof the full language, and the old standard hasbeen withdrawn.
The guiding criteria used in the developmentof the standard were CBRAI 78]:
1. the inclusion of only those new featuresproven through actual usage
2. the inclusion of new features that en-hance the portability of programs
11 4
104 NECC 1900
Week
1-2 Introduction to computers,programs and prorcam-ming languages
FORTRAN and-the basiccomputer operations
3
4
5
Problem AnalysisAlgorithm.development
and refinementFlow diagrams
One and two alternativedecisions
WHILE loopsCompiler role in trans-
relating structures
Data types in FORTRANList-directed formatting
DO loopsArithmetic expressions and
functions
6-7 Lists and subscriptedvariables
Searching a listIndex computation
8 Block-IF decisions struc-ture
Generalised DO loopNext iteration and loop
exit controlNested Structures
9-10 Functions and subroutinesArgument lists and global
Checking account program, primenumber identification, Centi-grade to Fahrenheit conversion
Computation of table of fac-torials, finding an item in alist, frequency distribution of
exam scores. Table lookup viadirect computation and searcharrays
Scoring a bowling game, drawinga bar graph, sorting an array
Simple statistical packageSort/merge package
Mortgage interest tables
General search subroutine,finding parameters of aDO loop header, text editorprogram system
Matrix inversion, status of aTic -Tac -Toe game, printing
block-letter patterns, schedu-ling class rooms
Course Outline and Assigned Problems
3. minimal increase in language or processorcomplexity
4. avoidance of features that conflict withthe previous (1966) standard t
5. elimination of features in the 1966standard only under clearly demonstrated circum-stances
6. production of a more precise descriptionof the language.
The new standard describes programs writtenin FORTRAN 77, and not the processors (such as acomplier or interpreter) of these programs. Theimplementation of a standarconforminn processor
1 s
Is to be inferred (rose the standard. The standardis to be interpreted as specifying only the minimumreggirementm of the language. Thus a standard -con -
fotaing processor for the language vast be able tohandle all standard - conforming programs accordingto the rules of the standard. It may, in addition,however, have extensions for features such as bitmanipulation or errs- processing that are notspecified in the language. It is then the decisionof the user whether or not to conform to the
standard when writing a program. Of course,standard-conforming programs usually will beportable to all machines supporting a standard -conforming compiler; non-standard programs maynot be as portable.
This section contains a list and descriptionof the new features of Tull FORTRAN which affectmost significantly the introductory course justdew:tibed. Those features discussed that have notbeen included in the subset are so designated.Users of systems not supporting the full languageshould sake adjustments as appropriate in thecurriculum changes suggested in the last sectionof the paper.
The relevant features of Full FORTRAN arelisted in Table 1 and summarized in the remainderof this section. The material presented is in noway intended as a complete description of thesefeatures; in fact, it barely scratches the surface.Additional examples and detail may be found in theBrainerd paper and in FORTRAN 77 introductoryprogramming texts (see [DR 78], and D0 79]).
TABLE I
NEW FORTRAN 77 FEATURES MOST RELEVANTTO INSTRUCTION IN INTRODUCTORY PROGRAMMING
1. The character data type2. List-directed formatting (not in the
subset3. The block-IF (if-then-else) decision
structure4. Generalized form of the DO
Statement
5. Arrays6. Expressions: the PARAMETER Statement
and mixed -mode arithmetic
7. The SAVE statement.
1-22&SheraeLar_Data_Tga&Perhaps the most significant change in the
standard is the addition of the character datatype, which now replaces the Hollerith type.It was the use of the Hollerith type that mademany FORTRAN programs difficult to understand, tocheck out, and to transport'from one computer to
another having a different sine storage cell.(Although the Hollerith type is no longer in-cluded in the standard, it is expected that mostmajor manufacturers will continue to support thisfeature in their FORTRAN 77 processors).
Some examples of the declarations of charactervariables and arrays are:
The length of each characi :r variable and array isfixed by the declaration statement. In thisexample, BUFFER is declared as a character stringof length 120, while CARD is taken to be an 80element array of character strings of length 1.
Character constants consist of strings ofcharacters enclosed in apostrophes. Charactervariables and array elements may be given values inthe same manner as other typed elements in FORTRAN:through the use of READ, assignment, and DATAstatements. For example, the statements shownbelow all have the effect of assigning the stringIVORY JOE to the variable FNAME, as previouslydeclared.
READ (FMTID10, UNIT" S) FRAME10 FORMAT (A)
Input Card IVORY JUE1234567890
FRAME a 'IVORY' //
FRAME 'IVORY JOE'DATA MAME / 'IVORY JOE' /
The READ statement illustrates the use ofcifiers for format and unit numbers. Other
input ou;put specifiers (for end-of-file, errorconditions, etc.) are also allowed (see (NO 80]).The old form
READ (10, 5) FRAME
is still permitted with the expected caveat thatthe first item listed specifies the unit number andthe second, the format number. The use of PMT andUNIT have the added advantage of allowing order-independent list specifiers in an input/outputstatement.
The use of the A descriptor by itself in afotnat is also new to FORTRAN. When the length ofthe element to be transmitted is not specified inthe format, it is taken from the declared length.
The statement
FNAME 'IVORY' // 'JOE' 'Illustrates the use of the character string concate-nation operation. Substring and string comparisonoperations are also permitted in FORTRAN 77, aswell as intrinsic function opurations for determin-
ing string length and for character.to -integer andinteger.to.character conversion. User-definedcharacter functions are also permitted in FORTRAN77. (The FORTRAN 77 Subset does not supportconcantenation, substrings, or character functions).
2- List - Directed Formatting (Not available in the
Subset)The new FORTRAN standard permits the uae of a
1sG
11
106 NECC 1960
statement label, an integer variable that has beenassigned a label, a cbatactet exptession, or anasterisk for the designation of a format. Potexample, statement lists i), 11), ill) and iv) ateallowed in FORTRAN 77 and produce identicalresults.
I) WRITE (6, 150) 'SUM N (N + / 2150 FORMAT (A, I6)
An astatisk is used to specify list-directedformatting, as determined by the input outputlist, the ptocessor, and the fora of the data. Forinput, the type of each data item is determined bythe FORTRAN system (rather than a user formatdescription). Additional information, such as theposition of the decimal point in a real number isalso determined in this manner. Data items maybe separated from one another using blanks orcomas.
REAL RATECHARACTER SSNO*11, NAME*24INTEGER FOURSREAD*, SSNO, NAME, HOURS, RATE
would produce the following result:
WEE=riODZI.V4 1414-Wi55-,
11121111 RATE1.35
As shown, character strings read via list-directedformatting must be enclosed in apostrophes; thefour data items shown lathe card are separatedfrom one another by one or mote blanks.
In list-directed output, the widths of thefields used to print the listed elements aredetermined by the type of the element and theprocessor. The width of each character string tobe printed is determined by the declared lengthof the string. However, real and integer'dataare printed in fixed-width fields regardless ofthe magnitude of the value to be printed; thefield widths ate pre-determined by the processorand are normally not alterable by the programmer.
3-The Block -IF Decision StructureThe block-IF structure will considerably
reduce the reliance upon the GOTO and the labelin FORTRAN ptogramaIng. This structure, withappropriate code indentation, can greatly enhancethe readibility of programs written in FORTRAN 77.
The logical function PRIMA shown in Fig. 2 (see(BRAT 78] provides a good illustration of the useand advantages of the block-IF.
LOGICAL FUNCTION PRIMEINTEGER N, DIVISRIF (N .LE. 1) THEN
PRIME = .FALSE.ELSE IF ('1 .!Q. 2) THEN
PRIME = .TRUE.ELSE IF (110D(N, 2) .EQ. 0) THEN
PRIME = .FALSE.ELSE
DO 10 DIVISR 0 3, INT(SQRT(REAL(N))), 2IF (MOD (N, D1VISR) .EQ. 0) THEN
PIER .FALSE.
RETURNEND1F
10 CONTINUEPRIME 0 .TRUE.
ENDIFRETURNEND
Fisz_k An Illustration of the Block-If Structure
This function shows the use of the Block-IFin implementing a decision structure with fouralternatives, and a decision structure having onealternative (inside loop 10). A two-alternativeBlock-If may be implemented in the form
IF (condition) THEN
ELSE
.1
ENDIF
If condition is true statementsequence executed
If condition is false statementsequence executed
4-Generalized Form of the DO 100 Statement_The previous sample also illustrates the more
general DO loop feature provided in FORTRAN 77.The form of the FORTRAN 77 DO loop header state-ment is:
DO sn loopvar = expl, exp2, exp3
whereI) expl, exp2, and exp3 represent the
initial, terminal, and step value expressionsrespectively
II) expl, exp2, and exp3 may be any integer,real, or double precision expression having positivenegative, or (except for exp3) zero values (the useof expressions is not permitted in the Subset)
ill) Loopvar may be any integer, real, ordouble precision variable
iv) an optimal comma is permitted after theterminal statement label, an.
1 2 7
The number of times * DO loop is executedis called the trio count. The trip count specifiedby the previously described loop header is computedas
VAX(iNT(v2-v1 + v3)/v3),0)
where vl, v2, and v3 are the values of the expres-sions expl, exp2, and exp3 respectively, and INTtruncates the result of the expression argumentif it is not an integer value. The followingrelationships must bold between v,l, v,2 and v3:
vl < v2 and v3 > 0
or14 > v2 and v3 < 0
If the value of the trip count is not positive, theloop will not execute at all. It is important tonote that this is contrary to the conventionadopted by many processors based upon the previousstandard: that loops were executed at least onceregsrdisss of the values of the loop parameters.The previous standard did not specify what was tobe done for loops written with an initial value
exceeding the terminal value at loop entry. Thechange will not effect the execution of programsthat were written in accordance with the previousstandard, ant it way affect those that werewritten in violation of the standard.
5-AII!1!The Full FORTRAN language defined in the
standard provides three major changes from theprevious standard:
i) arrays smyhave up to seven dimensionsii) the specification of a lower subscript
bound is allowediii) subscripts in an array reference may be
any integer expression (see also, section 6).
Items i) and ii) above, are not permitted inthe Subset.
The specification of the lower subscriptbound is indicated through the use of a colon, asIn
REAL X (-315)
which defines a nine-element array X with elementsx(-3), X(- 2)...X(0)...X(3). If the lower boundis omitted, it is assumed to bs one.
sere are places in the nunr1ORTBANin which expressions ore permitted where only
constants, variables,' or restricted forms of ex-Freestone Were previously allowed. For example,expressions are now allowed in output statements(see i. below), as subscripts (ii), as arraydimension bounds (ii), and es indexed -DO pare
meters (see section 4). Sons examples arei) WRITE (6, 130) 'SUM ', N * (N + 1) / 2
ii) INTEGER SIZE, I, J, lCPARMISTER (SIZE 10)
Sbuctured Programmkp 107
REAL X(2*SIXS)DirtA MI), I 1, SIZE) I MAO IWRITE (6, *) X(6*X-J)
Example 11) illustrates the use of the PARANITIRstatement, which is used to attach a symbolic sameto a program constant or parameter. (The UNARM'totem/et is net included in the Subset). If thesymbolic name is not of the defeat implied type,its type scat be specified before its appearance ise PARAMETER statement. Expressives are permittedto the right of the equal symbol, but they map
contain only constants or previously definedsymbolic constants.
Symbolic constants may be used in expressionsin the same way es variables. They Imp, in addi-tion, be used in specificities and DATA stateliest*
in places when only constants bed been previouslyallowed. Such uses are illustrated is limo 3 and4 of Example ii), abets the symbolic constant SIZEis used where only constants were previouslyallowed.
Example ii) also illustrates the use of theimplied -DO in DATA statemeut (not allowed in theSubset) and the moralisation of the fore of sub-scripts allowed is /OSMAN 77. Any integer ex-passion soy be used to specify a subscript,provided the value is within the bounds specifiedfor the corresponding dimension in the arraydeclaration. (The ell standard limited subscriptspecification to expressions of the form
Cl * V t C2
where C1,
C2were integer comstants, mud Ewes an
Integer variable).With regard to the formation of expressions,
the major difference between the 1966 standard andthe new NORMAN is the latitudes of slued-modearithmetic. Integer, real, double precision, andcomplex operands map appear in arithmetic expres-sions except that double precision sad complexoperands may not appear in tbe'esme expression.
The type of an expressive is determined byexmainlegoperand-operstor-operand triples --thetype of the result of each triple is determined bythe typo of the two operands involved. For ex-ample, if X, 1, and J are integers, and 14, J-2,then the result of the evelostive of the StitIMOSt
X - 2.3 I/J
is 4, es determined es follows:i) divide I/J; slims I'S and Jet are both
integers, the result of this divisive is aninteger, 2.
11) add the rod value 2.3 and the integer2; since 2.3 is real, the reedit of the divisionis first converted to real prior to addition whichthen yields a result of 4.3.
ill) the red result 4.3 is assigned for thelateen wettable E, resulting in the wrist nuiSn'seat of the truacatosivalue 4.
108 NECC 1980
7-The SAVE Statement
Contrary to what a number of FORTRAN users havebecome accustomed to, there is no requirement ineither the 1966 standard or the new standard toretain the values of local variables in subprograms
from one execution to the next. Programs writtenunder the assumption that local variables weresaved were non-standard and would not executecorrectly on all.proccssors.
While saving the value of local variables insubprograms used to be rather cumbersome, theFORTRAN 77 SAVE statement can be used to simplifythis process. For example, a small subprogram toincrement a counter and check for overflow nightappear as follows:
SUBROUTINE EMITINTEGER MAXVALPARANETER (MAXVAL st 100)INTEGER COUNTDATA COUNT /0/NAVE COUNT.COUNT w COUNT + 1IF (COUNT .GE. MAXVAL) THEN
PRINT*, 'COUNTER EXCEEDS MAX VALUE OF ', MAXPal,PRINT*, 'COUNTER RESET TO 1.'COUNT 1. 1
RETURNENDIFRETURNEND
The SAVE statement causes the value of COUNT to besaved between calls to the subprogram BUMPIT.
IMPACT OF NEW FEATURES UPON INSTRUCTIONOnly a few of the new features found in
FORTRAN 77 have been described in the previous
section. As indicated, the list is restricted tothose features which are felt to have the greatestimpact upon instruction in introductory programmingcourses using FORTRAN. Yet of the features dis-cussed, only three, the character data type, theBlock -rF structure, and list-directed formattinghave substantially contributed to major changes inthe structure and content of the course. Of theremaining features, the generalized DO loop(introduced in week 8), implied-DO in a DATAstatement (weeks 6-7), additional flexibility inthe use of expressions (spread throughout thecourse), and the additional array features (week 8)are creatures of increased convenience having onlya minor impact upon the course. The SAVE state-ment (weeks 9-10) is a feature that Mould beunderstood by all students working with subprograms,but it has no other impact upon the course.
The PARAMETER etatement (introduced in week 4)provides a vehicle for discussing the notion of aprogram parameter (as opposed to an in -line con-stant). This statement provides a convenient means
for a Programmer to attach a WOO to &constantvalue (programmer parameter) that has special pro-
gram significance. A value representing the maxi-mum size of an array is one example of a program
parameter. Other examples are illustrated in thesample program shown at the end of the paper.
In the view of this author, mixed-modearithmetic offers no advantage to the student in anintroductory course. On the contrary, the use ofmixed-mode expressions requires additional care anda more sophisticated understanding of expressionevaluation than is desirable or even necessaryat this level. The automatic type con-version required by the mixed-mode expressionprovides very little programming convenience andhas the added disadvantage of being hidden from theuser. The use of the type conversion functionssuch as INT (real-to-integer with truncation), AIRE(real-to-nearest integer), and REAL (integer-to-real) in avoiding mixed-mode. arithmetic should beencouraged.
It is perhaps not too surprising that thecharacter data type, Bloc{ -IF, and list-directedformatting have had the most influence upon thecourse. For, as an analysis of the course struc-ture outline indicates, the major emphasis in thecourse is upon problem solving, rather than thedetails of the FORTRAN language. These three fea-tures make it even easier to concentrate onproblem- solving techniques and algorithm develop-ment, with less emphasis upon the languageimplementation considerations. Yet each contrib-utes to this effort in a different way.
The character data type makes it possible tointroduce the concept of a character string, andthe reading, printing, and comparison of stringsat a very early stage of the course. For example,character string constants are used in list -directed output statements as early as week 1, inorder to provide descriptive labels or headers forthe values printed by the first program run bya student. By week 4, the student is reading andprinting character strings using list-directedformatting and simple character string comparison.During week 4, data types are discussed in detail,and a more formal presentation of the chsractertype is Oren. Finally, by weeks 12 and 13,students are writing programs requiring the use ofsome simple but fundamental string operations suchas aubstring extraction, comparison, insertion,deletion, and replacement. Most important, allwork is now done in a totally machine-independentfashion, without the previous requirement of havingto store data of one type (character) in a memorycell of a different type (real or integer). Thislatter feature has the added advantage of providingadditional compile-time checking for data type andoperator consistency.
The Block-IF structure eliminates most of theneed for COTOs and labels in the implementation ofalgorithms in FORTRAN 77. The only remaining COTOneeds now can easily be restricted to implementa-tion of the WHILE loop structure and loop exit andnext iteration steps. The Block -IF makes it
possible for instructors and students to concen-trate on the specification of decision steps interms of the tasks to be accomplished and thecondition(s) of selection, without concern for thedetails of using labels and COTOs. The resulting
.1r9
implementation is not only easier to write, but alsoeasier to understand and far less subject to error.This point is perhaps but illustrated via onepossible rewrite using COTOs and labels of theexecutable portion of the function PRIME (see Pig.3).
IP (N .GT. 1) GO TO 1PRIME .FALSE.
RETURNIP (N .NE. 2) co TO 2
PRIMERETURN
2 IP (MOD(N, 2) .NE. 0) GO TO 3PRIME .FALSE.RETURN
Even with indentation, the logical structureof this code is obfuscated considerably because ofthe COTOS and labels.
The single and double alternative forms ofthe block-IF are introduced during week 3 of thecourse, and the general form is presented duringweek B. By this time, however, students have been
doing considerable programetag, using the COTOonly for the implementation of the MILE loopstructure. (Even this use could have been avoidedhad the FORTRAN 77 standard contained a VSILE-likeloop structure in which the repetition conditioncould be specified in the structure header).
Another advantage of the block-IF is increasedsimilarity in the structure of student programs.Students (and instructors, too) are now better ableto understand the programs of others Aid to assistin checking out and correcting programs that fail.
Finally, there is the matter of list-directedformatting. List - directed formatting allows theinstructor to delay considerably any discussion ofone of the most detailed, unpleasant features ofthe FORTRAN language -- formats. While it is truethat formers are an important feature of theFORTRAN language, and should not he overlooked, thestudy of formats contributes little to studentmastery of the fundamentals of choosing datastructures and designing and Implementingalgorithms. It is for this reason that formatehave been delayed until week 11 of the course,after the presentation of subroutines and functionsis complete. It is indeed a shame that the list -directed formatting feature is not included in theSubset. It is hoped that most Subset implementorswill view it as top priority for support in theirprocessors.
In the course, list-directed formatting isintroduced in the first program given to students
Structured Programming 109
to prepare and submit for computer entry. Thecapabilities of the list-directed feature are thenslowly expanded informally throughout weeks 2through 4 until finally, in week 5, it isformally described,and a brief overview of how itworks is presented. Exclusive use of list-directed formatting in all problems studied andassigned is then continued until week 11.
CONCLUDING COMMINTSOf all the new features of FORTRAN 77, list -
directed formatting provides the greatest addedconvenience to instructors of introductoryprogramming courses (be they in FORTRAN, BASIC,PASCAL, or any other language). The block -IP andcharacter features are not far behind in thisrespect. But list-directed formatting allowsstudents to begin to do input and outputimmediately, and to continue to do even slightlysophisticated input /output throughout the firsttea weeks of the course (including the study ofsubprograms) without the added complicationintroduced through the use of formats.
The following program (Pig. 4) illustratesthe use of list-directed formatting, the characterdata type, the block -IP structure, and thePARAMETER statement. Arrays are used in the pro-gram only for illustration, not because they arerequired. The program was run on a Control DataCorporation Cyber 174 using the University ofMinnesota FORTRAN compiler, 1177. For the exampleinput entries show below
THE NUMBER OP OUTSTANDING SCORES IS 1TEE NUMBER OF SATISFACTORY SCORES IS 3/RENUMBER OP UNSATISFACTORY SCORES IS 1
THE NUMBER OF INVALID SCORES IS 1
REFERENCES(ANSI 783 American National Standard Programming
Lansuage FORTRAN, American NationalStandards Institute, New York, 1978.
[BRAT 783 Brainard, Walter S. et. 81., "FORTRAN77", CACI (21, 10), October 1978, pp.806-20.
110 NECC 1980
CC PROCESS EACH RECORD. DETERMINE AND PRINT RATING ALONG WITH
C PRINT COUNTS
'
C PROGRAM TO PROCESS A SET OF EXAM SCORES RANGING BETWEEN 0 AND 10C AND CLASSIFY ACCORDING TO OUTSTANDING (8-10), SATISFACTORY (4-7),C OR POOR (0-3). COMPUTE AND !RENT FREQUENCY COUNTS.
CC PRANK L. PRIEDMAN 12-9-79C
CC READ AND VALIDATE N
CC
PRIM*, 'PROGRAM TERMINATED.'
C NAME AND SCORE. INCREMENT APPROPRIATE COUNTER.
.GT. MAXSCR) THEN
. nine*,
PRINT*, THE NUMBER OP OUTSTANDING SCORES IS ", OUTNRPRINT*, ' THE NUMBER OF SATISFACTORY SCORES IS ', SATNR
plaw, "PRINT*, THE NUMBER OF INVALID SCORES IS ', ERRCNT
C
YAMMER oamouT 8, MINSAT - 4, MAWR s 0, MAXSCR 10)
PARAMETER (MAXCNT 100)CRARACTER*24 NAME (MAXCNT)
READ*, N
' NUMBER OF EXAM SCORES IS NOT VALID, MAX IS 1, MAXCNT
SATNR * 0UNSNR * 0
DO 40 I 1, N
mum '
PRIME*, 9 THE NUMBER OF UNSATISFACTORY SCORES IS ', UNSNR
INTEGER MINOUT, MINSAT, MINSCR, MAXSCR
INTEGER MAXCNT
INTEGER SCORE (MAXCNT), N, Ix, ICRARACTER*20 RATING
IF (N .GT. 0 .AND. N .L14MAXCAT) THENmu:RINT*, 'NUMBER OP EXAM SCORES IS ', N
OUTNR * 0
BRUNT - 0
STOPEND
INTEGER OUTNR, SATNR, UNSNR, ERRCNT
20 CONTINUE
DO 20 II * N
PRINT*,
ELSE:1P (SCORE (I) .GE. mum, THEN
READ*, NAME(II), SCORE (II)
1
LENIts Somas' FORTRAN 77 Program
NAME
', RATING
SCORE
1 21
RATING'
Structured Programming 111
ERN 783 Davis, Cordon B., and Thomas R. Hoffman,TORTRANI A Structured, DisciplinedStyle, McGrew-Rill, 1978.
CFK 773 Priedaaa, frank L. and Elliot B. Roffman,Problem Solving and Structured Program:Lag in FORTRAN, Addism4esley, 1977.
CPR 783 Friedman, Frank L. and Elliot R. Koffman,"Teaching Problem Solving and StructuredProgramming in FORTRAN," Computers andEducation (2. 3), Perganon Press,January 1978, Pp. 235-45.
(35 793 Rune, J.N.P., and R.C. Volt, ProarennimRFORTRAN 77: A Structured Approach,-711*W1779.
CND 801 Meissner, Loren P. and Elliot T. Organick,FORTRAN 77 Featuring Structure Programming,Addison - Wesley, 1980.
112 NECC 1980
USING MODEL-BASEDINSTRUCTION TO TEACH
. . PASCAL
Bogdan Czejdo
Warsaw Technical Univ.Brigham Young Univ.
I. INTRODUCTION177:1iiiiWer we introduce the concept
of model-based instruction, first defining'model' and then presenting a simplifiedview of the modeling process. A way ofclassifying models is presented, and aprocess for preparing a set of complemen-tary models for use in model-based Jastruc-tion along with the characteristics ofthis type of instruction. The paper thenconcludes with the application of model-based instruction to computer-assistedinstruction.
II. THE ROLE OF MODELS IN TEACHINGA model is defined in Webster's Bic-
tionaryw to be a*:1. copy, image2. pattern of something to be me&3. archetype4. description or analogy used to
help visualize something thatcannot be directly visualized
S. system of postulates, data, andinferences presented as a mathe-matical description of an entityor state of affairs
For the purposes of this paper, a modelis defined to be any diagram, table, pic-ture, or figure which helps the studentto understand and remember a'concept orperform an action which is a part of aset of teaching objectives. As shown byresearch in a variety of teaching areas04,models play a very important role in thelearning process.
A simplified view of the modeling processin learning is represented in Figure 1.
We have chosen a subset of meaningwhich corresponds with the meaningof model in our paper.
l'Iten 1. 711, °mattes of Not .16
The internal models students created intheir %mon minds Allow them to understandand remember concepts and perform pre-scribed actions. The students may createthese models on their own (transformationIl) based on descriptions provided in text-1'books and lectures. Alternatively, theinternal models used by the student may bebased on external models (transformation 13)already created by the textbook author orlecturer (transformation 12).
The creation and use of internal modelsis a complex psychological process whichwill be mentioned only briefly here. Itshould be recognized that the creation of .
models by the student.is a difficult andtime-consuming task. During lecture, there .
is usually insufficient time for the studentto formulate suitable models on his own.The fuzzy and incomplete models that maycome to mind are soon forgotten because theyare not strongly impressed on him throughreinforcement by the teacher or by his ownperformance.
Indeed, transformation Ii is difficultfor the student for a varie y of reasons.The student may be untrained in the crea-tion of models or have a weak imagination.Because of an incomplete understanding ofthe full subject, the models he createsmay turn out to be unsuitable. UnlearningIn initial model and relearning a morea4Aquate model may be very difficult:
123
For all of these reasons, it is importantfor the teacher, and in particular the com-puter science teacher, to pay careful atten-tion to the subject of models and to providethe student with useful external models.(In the rest of this paper, the term *model"will refer. to external model.) - -
III. TYPES OF MODELSA wide variety of models have been
examined and evaluated for potential usein teaching PASCAL programming. Thesemodels have been drawn from several text-books(3,5) and have also been created bythe authors of this paper. What constitutesthe model and the object to be modeleddepends on the point of view of teacher andstudent. A computer program, for example,is often a model of some real-world activityor process. However, in the context of thispaper, the computer program will be consideredto be the object to be modeled.
One of the best ways to classify modelsis by their structure. Six basic structureshave been identified:
Most of the models we use are gra phicstructures, four of which are leviilic-tfees(hierarchies), networks, and domains. Levelsand tree structures can be described usingthe example of an interactive language. Onthe microcomputer systems used by the authorsin their introductory PASCAL course, there isan interactive command language which is inter-preted by a system monitor or operating system.Each user starts at the command level. Bytyping an "E", the user can go to the editlevel. To return to the command level fromthe edit level, the user can type a *QVsequence. Figure 2 shows simple level struc-ture.
Command level
Edit level
1QU
Figure 2. Simple Level Structure
Levels are represented by horizontal lines.The transitions are represented by labeledvertical lines. Figure 3 shows a simpletree structure.
Structured Programming 113
Figure 3. Simple Tree Structure
Each circle represents being in a parti-cular state (at a particular level in thiscase).
It should now be obvious how one can com-bibe elementary submodels into a more com-plex structure. In Figures 4 and 5, amore complete model of an interactive lan-guage is drawn using the level and treestructures of Figures 2 and 3.
Command level
Edit level
Insert level
Figure 4. Level Structure
Figure 5. Tree Structure
The root or first level is drawn at thetop of the figure but, of course, thereverse option is also possible. In the
I 0 44. 41
114 NECC 1900
tree model, each state has at most one"parent" state making it a strictly hier-archical structure. If this restriction isrelaxed, a network structure results. Anexample of a network structure is the syn-tax diagram in Figure 6 for a BEGIN-ENDblock;
Figure 6.
Syntax Diagram for BEGINEND Block
Another *Wimple of network structure isthe model of the MAK computer shown inFigure 7.
CPU Disk Memory
Figure 7.
Model of Flow of Information inMinicomputer flan
The tree and network structures consistof nodes [states) connected with directedlines. The final graphic structure to bediscussed in this paper, the domain, con-sists of a set of'areas which either over-lap or are disjoint. Figure 8 illustratesa simple example of a domain structure thatis a model for teaching the Boolean opera-tions OR, NOT, and AND.
There are, of course, a wide variety ofother graphic structures that might beuseful in other fields of learning.
A second type of structure is the arraystructure. A simple one-dimensional arrayIIMEcalled a list, a two-dimensionalarray is a matrix. Array structures findtheir greatest utility in modeling computermemories (core, disk, etc.) and in repre-senting data. See Figure 9 for an exampleof an array structure which describes thename, type, and value of PASCAL variables.
PASCAL Variables
Name MS .Value
count
surname
Pi
alpha
condition
integer
string
real
character
boolean
43
Jones
3.14
G
true
Figure 9.
Name/Type/Value Model
Mathematical structures constitute athilrEYfroriodel structure consisting ofa string of symbols (letters, numbers,mathematical operators, etc.). Figure 10Illustrates one type of mathematical struc-ture, the Backus Normal Form model. It isan alternative model for the same BEGIN-ENDblock as described by Figura 6, but with amore mathematical flavor.
For completeness, we add a fourth type,the textual structure, which is nothing buttext711173ifairiEgiage. An article,letter, or book all contain such structures.We introduce this type of model here, notbecause we want to study books as textualstructures, but because we want to alio',textual structures as small components oflarger models.
Having introduced various types ofstructures, we are now in a position todescribe compound structures. A compound
12
structure is one that is composed of severaltypes of substructures For example, it ispossible to combine two types of model struc-tures (array and graphic) into one compoundmodel as shown in Figure 11.
l- 000e0Ao1 Neon) b) isrtrage
no SI MUM) 8 SIRING01 IOTWOO)
ZOMBA Ps mum'tvoCibelei Mt 101 intabol)Menet= Mit CU) ImOseldb)
N - mune
el
fooest404 NM) MIMI)
d)
noun U.01144444446, User Streuture 414 owl GeephIesi Steuctego
Iete Cempound MK 40.
The last, and probably the most signifi-cant, model structure to be described hereis the rallel structure. A parallelstructure s similar to the compound struc-ture in that it consists of a collection ofseveral more elementary models. However, aparallel structure differs from the com-pound structure, as the sub-models withinthe parallel structure remain distinct andphysically separate, related only in thatthey represent complementary aspects orfeatures of the same real-world object.The advantage of several parallel modelsover one compound model lies in the sim-plicity that can be retained in each com-plementary sub -model while at the same timerepresenting all the necessary features ofthe system.
For example, the models in both Figures5 and 7 represent certain features of theTERM microcomputer. Figure 5 illustratesthe steps necessary to have the computerperform certain actions, while Figure 7shows Ale flow of information within thesystem. Thus, these two models form aparallel structure. Another example ofparallel structure is shown in Figures6 and 11. Figure 6 depicts the syntaxof the PASCAL language (in a sense, thesteps or actions in programming) whileFigure 11 represents the computer memory(information flow) associated with a
Structured Programming 115
PASCAL program. Thus, a parallel structureconsisting of one model to represent theinformation flow and another to representthe human performance (operator actions)seems especially appropriate and of generalapplicability.
IV. EVOLVING PARALLEL STRUCTURESEffective teaching of a comgix Model or
of a set of parallel models involves aseries of steps on lessons. The modelsupon which these lessons are based must becarefully selected and designed so that thestudent's understanding grows in a smoothand orderly way. This process suggests afamily of closely inter-related modelsculminating in the final parallel structurewhich has been selected to represent thesystem under study. The model chosen fora particular micro-lesson may be relatedto the models used in previous micro-lessons by;
(1) incorporating some new feature orcomponent into a previous model.
(2) synthesis of several previous modelsinto a new model.
(3) providing a more exact descriptionof some part of a previous model.
As an example of this process, considerthe model of a TERAK computer in Figure 7.One can introduce the concept of a workfile by expanding this model as shown inFigure 12.
rMOnitor Ic
IKeyboarci
CPU Disk Memory
I
file ABC
workfile
Figure 12. TERAK Information Flow
Another example of model evolution canbe seen from a comparison of Figures 5 and13. Here again a new feature is added tothe model to create a more extensive model.
126
116 NECC 1980
Figure 13. Tree Structure
It should be noted that the models ofFigures 5, 13, 7, and 12 are inter-relatedand constitute an evolving parallel struc-ture. Figure 14 depicts such a structurewhere M1 and Mi' represent the models ofFigures 5 and 13 and M2 and M2' correspondto Figures 7 and 12. Each connectingline (Ri) represents the relationshipbetween two models or the relationship ofa model to reality.
Microteaching MicroteachingUnit 1
1
Unit 21
Figure 14.
Evolving Parallel Structure
V. TEACHING OBJECTIVES IN MODEL -BASEDINSTRUCTIONOnce an evolving parallel structure of
models has been designed, it becomes fairlystraightforward to define the teachingobjectives. The objectives can be des-cribed in terms of the models, the rela-tionships between the models, and rela-tionships of the models to reality.
The purpose of each micro- teaching unitwould be to solve a practical problem insome particular area. The process of pro-blem solving using models is shown inFigure 15.
Model 2
Model 1
Reality
Figure 15. Problem Solving Using Models
The problem-solving process illustratedin Figure 15 consists of 6 basic points:
1. PR-Practical problem to be solved2. Model M23. MS2- specified model M44. Model M15. MS1- specified model M16. PP-Practical performance
By specification, we mean the process ofapplying the model to a specific situationby selecting one state or path of the model.
The choice of teaching objective foreach micro-teaching unit can now be speci-fied by the triplet of transformations:
T1,
T2'
T3
where each T is defined as follows:
MS2 T1 2(PR MM)
2
MSM = T2 (MS2, 1(MSM, MM)
PP = T3(MS
1 )
These transformations represent:
T - The transformation of the Model1 M, for a given practical problem,
PR.
T2- The transformation of the model
M1
for a given specified model,2.
T3- Practical perfoimance for the
specified model MS1.
Models enable us to describe teachingobjectives more precisely, which is, ofcourse, a very important condition for thesuccess of teaching. Another advantage ofthis way of designing teaching objectivesis connected with the analysis of problemsolving, as one can concentrate on under-standing and performance more than remem-bering.
V/. MODEL -BASED INSTRUCTIONHaving prepared specific teaching ob-
jectives, we are in a position to designthe micro-teaching lessons in detail. Twophases .re involved: _first we teach the
127
models and the relationship (the generali-zation processes), and second we teach howto apply them (the specification processes).An example of this process is shown inFigure 16.
Figure 16.
Diagram of First Phaseof Microteaching Lesson
A description of the steps in Figure 16is as follows:
1. General description of a device(TERM computer).
2. Introducing the basic elements of theModel, M2, and relations (Keyboard,Monitor, CPU, and disk drive).
3. Basic elements of the model, MI (modes).4. Description of the first task to be
performed (transfer from keyboard toCPU).
5-6. Generalization of the model, M2, sothat it represents task 1 (adding in-lormation flow from keyboard to CPU).
7-8. Generalization of the model, M1, sothat it represents performance con-nected with task 1 (adding actions Eand I).
9-15. Repetition of steps 4-9 but for task2.
Thus, the first phase is the process ofgeneralization of a model. It includes theintroduction of new aspects to the modelrelated to each elementary task. The secondphase is the process of specification and isbased on model specification as shown inFigure 15. Again, we repeat this phaseseveral times with increasing student activ-ity until students can solve any problem inthis area.
Now we are ready to describe the basiccharacteristics of model-based instruction:
1. Parallel structures2. An evolving sequence of models3. Teaching objectives which are based
on the evolving-parallel structure4. A method of teaching based on moving
from a practical situation to a modeland then back to practical performance.
Structured Programming 117
In various practical situations, it maybe appropriate to relax one or more of theabove requirements for model-based instruc-tion. In addition, the diagram of teaching(Figures 15 and 16) can be modified depot:dinson the student's role in the creation anduse of the model. When_the model is an-algorithmic description of a process ofperformance, then the process of specifi-cation of the model is impossible, and wehave algorithm-based instruction ratherthan model-based instruction.
Some models require that certain para-meters be supplied by the user to completethe model. If this is the case, then thetransformations 2-3 and 4-5 of Figure 15will consist of a series of steps insteadof just one step.
Other models require some work to con-struct using simpler models. If the pro-cess of construction is given by some algo-rithmic rules, then the diagram of teachingwill not be substantially changed. However,if the creation of one model is describedin terms of other models, then in the pro-cess of problem solving we will have twophases in place of one. The first will beconnected with model construction for thepractical situation, and the second willbe connected with applying this model tosome practical action.
VII. THE APPLICATION OF MODEL-BASEDINSTRUCTION TO COMPUTER-ASSISTEDINSTRUCTION
On the basis of the principles of model-based instruction, a CAI facility has beenimplemented which teaches the PASCAL lan-guage on the TERAX computer. Plans areunderway to also implement this facility onthe TICCIT system.
It appears that MEI is a very usefulmethod in CAI because:
1. Computer graphics provide an effec-tive way of displaying models andtheir transformations. Model-basedlanguages are much easier to imple-ment than natural languages.
2. The strict definition of teachingobjectives makes it very convenientto evaluate user responses. This isvery important in tracing each stepin the process of problem solving.
Three types of conversation blocks havebeen designed as shown in Figures 17-19.
Figure 17. Simple Conversation Block
118 NECC 1980
Figure 17 shows a simple conversation block.It consists oft
P 4. the problem to be solvedC 4 correction block+ 4 right answer- 4. wrong answer
ri,nrs IS. Teaching CobvOtaatiest 'lock
In Figure 18, several simple conversationblocks are combined into a teaching sequence.In each simple conversation block, one ofthe transformations, tl, t2, t3, performedby the student, is evaluated.
Figure it. svaleatioft COOVO,SSUOR Sleek
Figure 19, in comparison with Fiaure 18,shows the process of evaluating two students'solution to a problem. The students' possibleresponses are represented by Al, A2, A3e andA4. Al is the correct answer. A2 is thewrong answer, but the answer is covered bythe rules of model M1. A3 is the wronganswer not covered by the rules of M1. A4is the exit path followed when a wrong answeris repeated. There are as usual correctionblocks, C2, C3, and C4, which differ depend-ing on the response of the user.
The conversation blocks described aboveare, together with the information blocks,the basic elements of a CAI micro- lesson.The diagram in Figure 20 is an example ofthis type of CAI lesson.
Mara 20. Disven Of CAISIteteleMpos
In Figure 20, IB represents an informa-tion block to be displayed on the monitor.B1 is a simple conversation block, 82 isa teaching conversation block, and B3 isan evaluation conversation block. It is,of course, possible to repeat conversationblocks B2 and B3 until the lesson ismastered by the student.
VIII. SUMMARYIn tE717751r, we described the use of
model-based instruction to teach thePASCAL language. In the classified set ofmodels, we found and defined a parallelstructure. We next showed how to transformthis structure into a sequence of evolvingmodels that constitute an evolving parallelstructure, the basis for defining teachingobjectives. The characteristics of model-based instruction were then described.Finally, the application of this teachingmethodology to computer-aided instructionwas shown, using as an example the teachingof the PASCAL language.
REFERENCES
1. Webster's Seventh New Collegiate Dic-ti.Taly, 1961 ed. G 4 C Merriam Co.Z--Yakowski, Ludwik. Nauczanie Problemowew Szkole Zawodowej. Warszawa: WSIP, 1974.3. BowIes, Kenneth L. Microcomputer Pro-blem Solving Using PASCAL. New YorksSpringer-Verlag, 1977.4. Jensen, Kathleen and With, Niklaus.Pascal User Manual and Report. New YorksSpringer-Verlag, 1970.S. Schneider, G.M., Weiengart, S.W., andPetit:me, D.M. An Introduction to Pro-gram:deg and Problem Solving with Pascal.New Yorks John Wiley, 1970.6. Niemierko, Boleslaw. Testy Osiagniec.Ilholnxch. Warszawa: WSIP, 19/37.
(MINI -MICRO COMPUTER) PASCALVERSION 11.0. Institute for InformationSystems, La Jolla CA, 1979.8. Czejdo, B., Kozinski, W., Kwiatkowski,S., Kipinski, E. "Zastosowanie MaszynCyfrowych do Automatyzacji Przetwarzania iPrzekazywania Informacji." Prace NaukowePolitechniki Warszawskiej Elektryka,47 (December 3977).9. Chanon, R.M. "A Course in Programmingand Practice: Toward Small Systems."SIGCSE BULLETIN, 1 (February 1978): 224-228.
1 9
Structured Programming 119
STRUCTURED MACHINE-LANGUAGE:AN INTRODUCTION TO BOTH LOW- AND HIGH-LEVEL PROGRAMMING
David G. HannayDepartment of Electrical Engineering and Computer Science
Union College, Schenectady, NY 12308(518) 370-6273
INTRODUCTIONStudents in introductory computer
courses are often frustrated by the un-familiar processes involved in solvingbewildering algebraic problems, particular-ly if their background in mathematics isweak. But these same students can learnthe fundamentals of structured programmingquickly and painlessly by directing theactivities of a simulated robot, R1-D1.
RI-DI is an indirect descendent of theturtle used by the LOGO group at M.I.T. tointroduce elementary school children to thecomputer. The immediate reward of seeinga turtle move across the floor or a robotacross the screen makes even non-mathe-matical students excited about programming.This pedagogical principle of immediate re-enforcement is as valid for college studentsas it is for young children.
But this visible movement is one ofthe least significant of Rl -Dl's abilities,for it has the arithmetic capabilities toperform simple computations and the logiccapabilities to illustrate the fundamentalsof structured programming.
RI-D1 was also designed for ease ofprogramming; each command consists of asingle letter or digit. A program con-sists of a free-form string of commandswith virtually none of the syntax rulesinvolved in the typical programming lang-uages; hence a student is free to concen-trate on such programming techniques asdecision making and loop control.
This particular robot has been usedsuccessfully in introductory programmingclasses by means of a simulator writtenin PASCAL. This simulator traces Rl -Dl's-movements and calculations on a comand-by-command basis. Once students havebeen introduced to programming via therobot, the concepts of programming in ahigher-level language become easier toexplain, and the concepts of computerorganization can be taught from a more
interesting perspective.
CAPABILITIESRl -Dl is a robot that has not outgrown
his training wheels. It can move back andforth in a straight line and position it-self over one of ten squares (memory cells)numbered 0 to 9:
forward
0 1 1^11MISIMMINI11 9
(- backwardIt can be directed to move to a specificsquare or to take a position relative tothe current square, moving forward orbackward one square at a time.
R1-D1 can also do a limited amount ofarithmetic. Each of the ten cells canstore an integer number. It can also addor subtract from an on-board accumulator.
Decision making is accomplished bycomparing the contents of the accumulatorwith the contents of the current cell thenexecuting a single command (or group ofcommands enclosed in parentheses) based onthe result of that comparison. For example,you could have the robot execute a commandonly if the accumulator was less than thecontents of cell 15.
Finally, R1-D1 can be programmed to re-peat one command (or a series of commands)a specified number of times. This struc-ture can be enriched by the use of thedecision making commands to exit from theloop prematurely based on certain conditions.For example, it could add a number to itselfseven times, or until the sum exceeded 1000,whichever comes first.
Input and output commands do not existas such. When the simulator is run, youspecify the initial contents of each of theten cells; their contents are displayedautomatically after each command is executed.However, if you wish to teach about I/O
13
120 NECC 1980instructions as such, reading from andwriting to the cells can be considered asinput and output respectively. Or, whilediscussing computer organization, you maytreat these instructions as memory fetchand store.
SUMMARY OF COMMANDSMiscellaneous Commands:"H" Balt." " Temporary Pause.(Allows entry of
immediate mode commands.)"N" Make a Noise."@" Execute commands from an indirect
command file.
Movement Commands:"F" Move forward one square."13" Move Backward one square."0" Move to square O."1" Move to square T."2" Move to square B."3" Move to square T."4" Move to square 1%"5" Move to square 14"6" Move to square B."7" Move to square T."8" Move to square T."9" Move to square I
Data Transfer Commands:"W" Write contents of accumulator
into current cell."R" Read contents of current cell
Into accumulator.
Arithmetic Commands:"8" Zero out accumulator."I" Increment accumulator by 1."D" Decrement accumulator by 1."A" Add contents of current cell
TO accumulator."S" Subtract contents of current
cell from accumulator.
Skip Commands:"E" Execute next command only if
accumulator is Equal to contentsof current cell.
"G" Execute next command only ifaccumulator is Greater thancontents of current cell.
"L" Execute 'nett command only ifaccumulator is Less thancontents of current cell.
Program Control Commands:"P* Perform a command as many times
MODES OF OPERATIONThe R1-Dl can be run in any one of
three different modes: immediate, direct,or programmed. In immediate mode, com-mands are executed as soon as the key isstruck on the keyboard. This immediatefeedback has proved very beneficial ineliminating some of the students' fearand has made the robot understood and
enjoyed even by children in elementaryschool. (While in this mode tl-e last twogroups of instructions, skip and programcontrol commands, are not allowed, as they,by definition, involve multi-charactercommand sequences.)
In direct mode, a complete commandstring is given directly to the simulatorwhich then begins execution. This allowsthe students to use multi-character commandsequences with minimal knowledge of thehost operating system. Students communi-cate directly with this one program withoutlearning to use a text editor, compiler, orother system programs.
Finally, as the system text editor isintroduced, students can edit and save pro-grams for RI-DI and modify them as they seethe results generated. This approach notOnly helps in teaching such concepts as de-bugging by tracing, but also gives thestudents a chance to use a subset of,thesystem text editor on a manageable pieceof text since programs are typically onlyone line long.
APPLICATION TO VARIED AGE LEVELSFirst graders enjoy watching the robot
move back and forth across the screen asthey enter commands in immediate mode.High schoolers can be shown graphically therelation of mult":)lication to addition anddivision to sub' *.ction as illustrated bythe sample program in the next section.College freshmen gain a glimpse of bothmachine language and PASCAL programmingconcepts.
SAMPLE PROGRAMA program which would take the initial
contents of cells 1 and 2, and put theirsum in cell 3, their product in cell 4,their quotient in cell 5, and their remain-der in cell 6 is given below:
1R2A3W1RCZ2PA4WZ5WIRC6WP(6R2LXS6W5RIW)11
This program not only illustrates severalimportant programming concepts such asloading and storing, looping, decisionmaking, etc,, but also serves to remindstudents that multiplication can be per-formed by successive additions, and thatdivision can be performed by successivesubtractions.
The first section of the program readsthe value from cell 1, adds the value incell 2, then stores the result in cell 3.
The second section of the program picksup the first number and loads it into therepeat count register. The accumulator iscleared, then the second number is repeated-ly added to it. 'The product is then storedin cell 4.
The third section of the program firstclears the quotient to zero and sets upthe dividend as the remainder. Then, eachtime through the "P" loop the divisor issubtracted from the dividend, and the re-mainder and quotient are updated. As soon
131
as the dividend is less than the divisor,the loop exit is taken and the programhalts.
One of the major disadvantages of thistype of programming is the obvious lack ofon-line documentation. But since the pro-grams are so short and simple, this lackof documentation has not proven to be ahandicap to students.
CONCLUSIONDecision making and loop control on
R1-D1 follow the general outlines of thePASCAL IF..THEN and REPEAT statementsrespectively. The equivalent of a GO TOstatement has been purposely omitted,thereby teaching students the concepts ofstructured programming from the very be-ginning. Parentheses can be used, whennecessary, to treat a group of commandsas a single command, just as BEGIN..ENDare used in PASCAL.
Students in introductory computercourses learn the fundamentals of struc-tured programming quickly and painlesslyby directing the activities of the robot.A few sessions with R1-D1 at the beginningof the computer science curriculum leavestudents both more knowledgeable and moreenthusiastic about programming than theywere in the pre-robot period.
SIMULATOR PROGRAM LISTING
PROGRAM R1D1(INPUT,OUTPUT);( THE WELL STRUCTURED ROBOT )
( BY; DAVID G. HANNAY )
VARCOMMAND
CHAR;ACC,COUNT,LEVEL,LOC,PC,PTR
t INTEGER;CELL
ARRAY(0..9) OFPROG
PACKED ARRAY[1.DOUBLE,VALID
t SET OF CHAR;
INTEGER;( PROGRAM STORAGE )
.00) OP CHAR;f SET OF PREFIX COMMANDS )
f SET OF VALID COMMANDS )
Structured Programming 121
( DIRECT EXECUTION VERSION )
{ CURRENT COMMAND )
( ACCUMULATOR )
f REPEAT COUNT REGISTER )
{ DEPTH OF *PERFORM" LOOPS )
{ CURRENT LOCATION OF R1D1 )
( PROGRAM COUNTER )
( TEMPORARY POINTER )
{ MEMORY CELLS )
PROCEDURE PERFORM(FIRST, LAST, TIMES: INTEGER);FORWARD/
David Moursund*Dept. of Computer &Information ScienceUniversity of OregonEugene, Oregon 97403Phone: (603) 686-4408
ABSTRACT.Although the instructional use'of
computers at the Precollese level isgrowing by leaps and bounds, it is stillin its infancy. Over the short run manyof the major problems are correctly per-ceived to be due to a lack of adequateor appropriate hardware, software, andcourseware. But there are two othermajor problems that will be the domi-nant long-term considerations. Theseare the problems of securing widespreadagreement upon the ultimate goal(s) forinstructional use gf computers and theproblem of teacher training.
The Association for Computing Ma-chinery's Elementary and SecondarySchools Subcommittee has been workingfor the past two years to identify someof the major problems of instructionaluse of computers and to help lay a foun-dation for progress towards their solu-tion. Some two dozen taskproups havebeen established. This paper is theintroduction to a NECC/2 session in
11.,:oursund is Chairman of the ACMElementary and Secondary Schools Sub-committee. He is also editor of TheComputing Teacher, a professional jour-nal aimed mainly at elementary andsecondary school teachers interestedin instructional use of computers.
which several of the taskgroup leaderswill discuss their progress.
In June 1978 the ACM Elementary andSecondary Schools Subcommittee was formedby merger of a Secondary Schools Subcom-mittee and a Teacher Certification Sub-conmittee. Since that time ES J has work-ed diligently to identify some of themajor Problems related to instructionaluse of computers at the precollege leveland to help solve them. Dug to very lim-ited financial resources ES in mostcases can only hone to provide some lead-ership and to lay a foundation which mayhelp lead to long term solutions.
This same two-year time span has.witnessed a massive influx of computerfacility into the schools. Reliable dataon how much hardware has become availablehas not been collected. !le know. however,that total sales of microcomputers byApple, Commodore, Radio Shack, and a num-ber of other companies total in the hun-dreds of thousands. The state of Minne-sota has long been recognized r itsleading role in the instructiot. use ofcomputers. Through the Minnesota Educa-tional Computing Consortium it has madetime-shared computing available to almostevery school in ehe states Precollegepublic education system. During the pasttwo years these same schools have added
125136
128 NECC 1980
approximately 1,000 microcomouters, whilecontinuing to increase their use of thetime-shared system. Minnesota has a pon-ulation of about 4 million, or a littleless than 2% of the total US population.
One can get additional insight intothis massive growth of microcomputer usein education by talking to a variety ofteachers and by attending local and na-tional professional meetings, ss it hasbeen my privilege to do. For example,each year the northern California affil-iate of the National Council of Teachersof Mathematics holds a meeting at Asil-omar, near Monterey, California. Usualattendance is about 1,500 educators. BobAlbrecht, well-known author and leader inthe computer education field, has attendedthis meeting for many years. Accordingto Bob the 1978 meeting included 4 modestamount of computer activity, and therewere 10 microcomputers available for dem-onstrations and hands-on use.
At the most recent 1979 meetingthere were computer-related talks sched-uled in parallel with every math session.There were some 66 microcomputers avail-able for hands-on experience, and therewas a well-organized software swap. Thesoftware swap was set up by the Computer-Uaing Educators (CUE), a northern Cali-fornia group that is lass than two yearsold. By the end of the Asilomar meetingCUE had well over 200 members. (1)
At the Asilomar meeting I talkedwith dozen of teachers who are just get-ting started in the instructional use ofcomputers. The most typical questionwent approximately like this: "Ourschool has $X to spend for computers.What should we do ?' The typical figurementioned ranged from a thousand dollarsor so up to many times that amount.
Two conclusions seem evident.First, many people with very little for-mal training or experience in the in-structional use of computers are now be-coming involved; indeed, they are beingasked to take major decisions that willaffect how computers are used in theirschools. Second, while the primary con-cern is still hardware, there is growingawareness on the part cf novice usersthat software is els° a major issue.
On the average, the novice computereducators know little about work done byothers nor of the higher level problemsfaced by the field of instructional useof computers.
WHERE ARE WE HEADED?In my opinion one of the major pro-
blema of instructional use of computersis there is little agreement as to where
we are headed. The overall long-termgoals are neither clearly understood norwidely accepted by the people who will beinvolved in implementing them.
Over the past decade computer liter-acy for all students has emerged as themajor goal in instru "tional use of com-puters at the nr -" ege level. Ini-tially comput4 acy tended to meancomputer aware It was thought bymany to be adeqa students couldcomt to understano ..ome of the capabili-ties and limitations of computers andthus gain some insight into how computerswere affecting the world and their lives.There was relatively little mention ofgiving simdents substantial training inuse of computers or inter iting theiruse into the curriculum. Lnitially, atleast, it was clear that resources to doso were not available within the.foresee-able future.
But the future proved difficult toforesee accurately, and large-scale in-tegrated circuitry became commonplace.As the price of computers dropped andavailability increased, the meaning ofcomputer literacy changed. Now the ex-pression tends to mean a functional,working knowledge of computers.
The analogy with reading and writingliteracy expressed by Art Luehrmann (2)is highly instructive. We can imaginebeing involved in education at the uriwnof the invention of reading and writingand entering into discussion as to theirrole in education. We can imagine edu-cators getting bogged down on issues suchas the brand of pencils and paper thatare best or lamenting the poor oualityand small, quantity of books that areavailable, But these turn out to beshort-term issues, and are certainly notthe major long-term problem. Over a timespan of a few thousand years, and aidedby the invention of movable iype and high-speed printing presses, reading and writ-ing come to be integrated into every as-'pect of human intellectual activity.
What have proved to be continuingProblems in reading and writing are il-lustrated by looking at examples, such asin mathematics and music. In each fieldthere he been a need to develop appro-priate symbols, notations, and vocabularyin order to represent the key ideas. Keyideas had to be developed and preserved.As this knowledge accumulated, educatoraare faced with the problem of hc.4 toteach it and what students should learn.
Now shift your attention back to theproblem of inatructional tse of computers.If our technologically oriented aocietycontinues to prosper then we can easily
137
ACM Elemental), and Secondary Schools Subcommittee 127
imagine that eventually computers will beas readily and conveniently available asbooks, pencils, and paper are today.Within that framework a single clear-cutgoal for instructional use of computersseems eiident to me. We want to integrateuse of computers into every aspect of hu-man intellectual activity in the samemanner that we have done for reading andwriting. The computer is a universaltool, a new aid to problem solving, andwe want all students to develop a highlevel of functional knowledge and skillin its use.
This ultimate goer. will take manydecades, centuries, or even milieu's toaccomplish. Progress will be made byidentifying problems that face currenteducators and working to solVe these pro-blems. The ACM ES has identified abouttwo dozen problems and established a likenumber of task groups to work on them.The next several sections of this paperdiscuss some of these problems.
SHOULD COMPUTERS BE USED IN EDUCATION?There is growing support prom tea-
chers, school administrators, schoolboard members, parents, and others forthe instructional use of computers ineducation. But much of this support isstill tentative, and much of it is basedupon incorrect insights as to how com-puters will affect education. RobertTaylor of Columbia Teachers Collegeheads the ES J Arguments taskgroup. Hisreview of the literature led him to as-semble a book of readings (3) on thearguments in favor of instructional useof computers that have been so well ex-pressed by various people over the past15 years. Perhaps the main conclusionto draw from Taylor's work is that manypeople are busy reinventing the wheel.Most people currently entering the com-puters in education field are not takingthe time to learn what ib already knownand thus are expending considerable ef-fort redoing what has already been done.
HARDWARE, SOFTWARE, AND COURSEWAREPROBLEMS
Several of the ES3 taskgroups areconcerned with the problems of hardware,software, and courseware; as mentionedearlier, many people view them as themajor issues that need to be resolved.The hardware and software problems areclosely related. We have a growing num-ber of options in educational b vdware,with the quality and capabilit thesemachines continuing to improve -utthere is a lack of software compatibil-ity between different manufacturer's
machines, and we are just beginning tosee how massively difficult the softwareproblem really is.
We can better understand the soft-ware/courseware problem by examining oneof the major directions of expanded com-puter activity, computer-assisted in-struction. The idea is for the computerto take over some of a teacher's func-tions, interacting with students to en-hance student learning. Material isneeded at every grade level and in everydiscipline, and the past 20 years of ex-perience of the PLATO project, as well asmany other computer-assisted instructionpro3ects, points out the difficulty ofdeveloping good quality instructionalsoftware and cvircware. Developing goodsoftware and computer-oriented coursewareis more difficult than developing themore traditional materials currently inuse. nreover, the current market issmall, the number of people involved indeveloping materials is modest, and themachinery for coordinated national ef-forts in development or distribution isnot yet well established.
It is clear, then, that regardlessof hardware progress, software and course-ware will continue to be major problemsfor many years to come. Some federalfunding is being made available, however.For example, Judy Edwards (4) heads athree year, $200,000 per year project towork on educational software for micro-computers.
Another approach is being stronglyencouraged by ES J and has also receivedthe backing of the International Councilfor Computers in Education (5). Local orstatewide groups of computer-using educa-tors who can readily participate in soft-ware and courseware exchange are being formed.Such local groups bring people togetherfor personal interaction and are provingto be a highly effective mode of infor-mation dissemination. Those interestedin starting such a group please contactthe author of this paper.
CURRICULUM CONTENT QUESTIONSComputers are a general aid to pro-
blem solving in every academic discipline,although they currently are much moreused in some areas than in other;. EShas taskgroups concerned with computeruses in mathematics, the sciences, thesocial sciences, the humanities (includ-ing art and music), business, and voca-tional education. The task in each caseis fairly similar. Eventually use of acomputer as an aid to knowing the subjectmatter and to solving problems from thesedisciplines will be routine. So far,
1 3
128 NECC 1980
however, relatively little progress hasoccurred.
We find the most progress in mathe-matics, Calculator use in secondary schoolmathematics is now common, especially inthe more advanced courses; strangely (inthe author's opinion) use of calculatorsin the remedial math courses has been moreslow to gain acceptance. The use of com-puters is taught in some math courses,but we seldom find a math course wherethe math content has changed to reflectwhat computers can do or how one usesthem in problem solving. Almost alwaysthe computer is an add-on topic and isused in conjunction with learning thetraditional topics in the traditional man-ner.
In very few secondary schools one canfind a significant change in the scienceor business curriculum due to commuters.As in math, computer use tends to be add-on in nature, to help students learn thetraditional content. Thus it is a'veryrare student who emerges from high schoolwith a functional knowledge of the com-puter as an aid to problem solving in avariety of disciplines.
The problem here is immense. Thecontent and coursework for every disci-pline needs to be rethought and redonein the light of computers and their ca-pabilities.
ELEMENTARY SCHOOLA single taskgroup is attempting to
bring order out of chaos in elementaryschool education, which represents abouthalf of all of precollege education. Al-though this group is not expending muchenergy on the issue of calculators, thatparticular issue gives good insight in-to some of the overall difficultiesfaced by the field of instructional useof computers.
Calculators are now ouite inexpen-sive and reliable. Even at the retaillevel one can buy a 4-function machinewith 4-key memory and liouid crystaldisplay for under $10. Since such amachine will last for years, an elem-entary school could provide its stu-dents with essentially unlimited accessto calculators at a cost of perhaps S2per student per year. The use of cal-culators at this level has been studiedin many research projects and has re-ceived the backing o such organiza-tions as the National Council of Teachersof Mathematics and the National Councilof Supervisors of Mathematics. Manybooks of calculator materials are nowavailable for use in the elementaryschool.
But calculators have hid essentiallyno impact upon the elementary school math-ematics curriculumt Speculations as towhy would fill at least an entire paper;however, it seems that a major factor isthat teachers lack appropriate knowledgeand skills.
What, then, can we expect to happenwith, computers in the elementary schOol?One answer is computer-assisted instruc-tion, with the main emphasis being on thecomputer taking over some of the instruc-tional processes currently handled inother ways. Proponents talk about tea-cher-proof materials and point out inad-eouacies in the current instructional pro-gram. Opponents point out the inadeoua-cies of commuterized instructional mat-erials and discuss the wisdom of thistype of change in our instructional de-livery system,
Very few elementary schools current-ly attemot to teach studenta about com-puters 'We have only modest insight asto what is appropriate. Again, the majordrawback is teacher knowledge.
TEACHER KNOWLEDGEEvery teacher knows how to read and
write and makes use of this knowledge inteaching. Every teacher has substantialinsight into the use of reading andwriting as a tool to learning the disci-plines s/he teaches, and in solving theproblems of these disciplines. The greatbulk of instructional materials buildsumon students' abilities to read andwrite
Contrast this with teacher knowledgeof computers! It is a truly rare achoolthat has even one teacher who uses com-puters as an everyday tool in cooing withthe yrobleme of his/her disciplihe. Thia,then, is the major problem. We are.ask-ing computer illiterate teachers to helpstudents to become computer literate ata functional level.
This problem is being attacked inmany ways. Among these are teacher cer.tification requirements, pre-servicecourses, in-service courses, self-directedstudy, and so on. Many teacher organiza-tions recognize the problem and are in-cluding computer talks and computer ta-torials in their professional meetings.All of these things are necessary, andall are helping. But progress seems slowrelative to the magnitude of the task.The educational world bas yet to acceptthe use of computers in precollege educa-tion as a major goal, and thus to beginto devote the resources necessary forrapid progress.
13
ACM Elementary and Secondary Schools Subcommittee 129
REFERENCESI. Computer-Using Educators. Founder
and president of this organization isDr. William (Sandy) Wagner, MountainVtew High School, Mountain View, CA94041.
2. Lenhrmann, Arthur. "Should the Com-puter Teach the Student, or Vice-versa?" Proceedings of the SpringJoint Conference, 1972. Reprintedin Creative Computing, vol. 2, no.6, Nov-Dec 1976.
3. Taylor, Robert. Tutor, Tool Tutee:The Computer in the School. leachersCollege Press, scheduled Tor publica-
5. International Council for Computersin Educations This non-profit pro-fessional organization publishes TheComputing Teacher, a journal for i37ucators. Seven issues are plannedfor academic year 1980-81, and theprice for US subscribers is $10.Write to ICCE, %Computing Center,Eastern Opegon College, La Grande,OR 97850,
140
130 NECC 1960
Robert P. TaylorTowbars CollegeColumbia University(212) 6784484
COMpUTING CONPITINCIIS POR BCSOOL MOMS
James L. PoirotNorth Texas University(817) 788-2521
IN/ROPUCTI014On December 4-5, 1978, the Elementary and
Secondary Schools Subcommittee (313) of the ACKCurriculum Committee met in Washington, DC tobegin formally laying out curricular and teachertraining guidelines for the integration of com-puting into the elementary and secondary schoolsof the country. Building upon a lengthy initialdiscussion and a review of earlier papers anddocuments dealing with the ease problem, the sub-committee outlined what it would attempt to dc.This process identified several tasks thattogether would constitute the overall work of thesubcommittee. finally, on the basis of interestand expertise, each participant in the l$3 meetingwas assigned to a working task group to furtherdefine and carry out one of the identified tasks.
Among the tanks identified was one dealingwith teacher training: to define the scope andunbutance of teacher training needed to integrate
coeputing into the schools. Accordingly, ateacher !reining task group was formed.
This paper is a product of that task group.It deals only with a subset of the issues andareas related to designing overall, computer -
literate teacher training. To appreciate itsfocus and accept some of its omissions, one shouldbe aware of the following conctrsinta that taskgroup, placed on itself. first, it was unanimouslyagreed that definitions should be in terse of ore.potencies to be achieved rather than in terms ofprograms or courses to transmit those competan*
cies. Socund,beesuse the computing coepetenciesneeded by the teacher who must teach computing asa subject are more extensive than those needed byother teachers, the competencies needed only bythe computing teacher should be treated es a sepa-rate module. Third, though integrally related toeach other, the competencies needed by the teacherare quite different from those needed by theteacher's teachers, the staffs of institutionsactually doing the teacher training. It wasagreed, there:ors, that specifying the compete*.cies needed by the teacher's teacher would beseparate module of work. It was also agreed thatit should only be undertaken after the competen-cies needed by the teacher had been specified.
The task group saw the coerstencies needed
James P. PowellNorth Carolina State University(919) 737-2858
by teachers at the school level as belonging toone of three sets. The Met set encompassesthose basic universal computing oompatenciesrequired for any school teaching, regardless oflevel or subject. The second encompasss thoseadditional computing competencies needed only bythe teacher **must teach computing es a subjectin its own right. The third encompasses' addi-tional computing related, subject - specific com-petencies needed by teachers of reinjects otherthan computing.
This paper outlines the competencies in allthree sets. It Incorporates critical suggestionsreceived as a result of wide circulation of twoearlier papers on the topic (1,2). We also trustit will stimulate useful discussion and criti-cism. We hope it provides some guidance to thosewondering abet teachers should know about com-puting.
COMPUTING COMMITUNCIsS Halpin BY TENCaiRsThree sets of computing competencies
The Wet (1.0) includes those which allteachers must haft, regardless of their level ordiscipline, even if that discipline is theteaching of computing itself. The second set(2.0) includes those needed only by the teacherof computing as a subject. It should be notedthis second set presupposes the Met. The thirdset (3.0) includes additional competencies forteachers who use computing to support or enhanceinstruction in subjects other than computing.leery teacher should acquire the competencieslisted in the first set (1.0) and the competen-cies listed in either the second set (2.0) or thethird set (3.0).
1.0 s Universal computing competencies needed 1well teachers
These are computing competencies which ellschool teachers should have to teach effectivelyin a society permeated by computes. They caststo either cc both Of two goals: (1) to undergated computing and (C) to use computing. Theycan be stated partially in terms of competencieslisted in ACM's "Curriculum '78r (3) and pertially in termed different competencies derivedfrom other sources. >t number of such other
14
ACM Elementary and Secondary Schools Subcommittee 131
1
sources are listed in the references at the enactthis paper. They refloat the abundance of diversework that has taken place in the past decaderelating computing to education.
1.1 ComfttenciesIn terms of these universal competencies,
_Ames teacher should:C1.1 be able to read and write simple pro-
grams that work correctly and to under-stand bow progress and subprograms fittogether into systems;
C1.2 have experience using educationalapplication software and documentation;
C1.3 have a working knowledge of computerterminology, particularly as it relatesto hardware;
C1.4 know by example, particularly in usingcomputers in education, some types ofproblems that ate and some general typesof problems that are not currentlyamenable to computer solution;
C1.5 be able to identify and use currentinformation on computing as it relatesto education;
C1.6 be able to discuss at the level of anintelligent layperson some of the-his--tory of computing, particularly as itrelates to education; and
Cl.? be able to discuss mural or human-impactissues of computing as they relate tosocietal use of computers generally andeducational use particularly.
The ahoy* competencies should be transmittedwithin the general preparation programs for allteachers by having those progress include thetopics listed below (11.1 through 11.5). For
those being trained to teach computer science,those topics will represent only a small subset ofwhet mat be learned about computing and its uses(see Section 2.0). For all other teachers, how-ever (apart from the subject-specific competen-cies covered in Section 3.0), these topics covermuch of what must be learned about computing bythe teacher who is to be minimally literate.
1.2 s Topics of StudyT1.1 Programming "Topics: Includes develop -
sent of simple algorithms and theirimplementation in a programing lan-guage, programming style, debugging andverification, end task-specific pro-gramming for educational applications.
11.3 Classic Applications of Cosentino inmoons Includes representativeexperience with problem solving and
text manipulation; simulation, drilland practice, and complex tutorial
systems including complete studentprogress record keeping; and educa-tional administrative systems.
11.4 Wuman/Nachine Relationships: Includesartificial intelligence, robotics,computer assistance in decision making(e.g., medical, legal, business),simulations, and computers in fiction.
11.5 Information on computers, in educationsIncludes periodicals, important books;on-line inquiry sources such as ERIC;professional societies such as acts,
ARCS, NCTRI time-sharing networks,networks of compuuter users; and hard-ware vendor groups.
In 11.1 procedures or algorithm are at theheart of computing so teachers should learn whatthey are by implementing simple examples such as:a procedure to average a class's grades, a gameto guess what number the computer is thinking of,or a procedure to display a large box on thescreen and then mate it shrink to disappearance.
Teachers need to be able to implement suchprocedures in only one language but should beable to read at the same, or a greater level ofdifficulty in a second so that the idea of alanguage, its strength and limitations, springsfrom personally experiencing functional dif-ferences between two languages. Within thelimitations of simple programs,teachers should betaught to write well-structured code, easilyreadable by others, and to document their code inacceptable fashion.
11.2 should be integrated throughout thecourse of study. In order to successfully usecomputing the vocabulary of the field must beunderstood. What is the difference betweentape, a floppy disk, and a disk? Why use one overthe other? These are general ideas but someminimal understanding of thesis necessary to usecomputers.
11.3 should certainly familiarise theteacher with several of the existing well-developed CAI systems cited in Section, 3.0 below.
?etchers will not develop new ideas about whatcould be done in their areas unless they see thebest of what has been done; neither will they geta full understanding of what can not bereasonably well dons by computer without suchexposure. Tor example, acquaintance with thephysics system designed by Bork (4), with CCCdrill and practice systems based on Suppes's work(5), or with PLATO work (6), should serve toacquaint the teacher with the CAI issues.
Since many teachers end up in administrationand since administrative uses of computers affectthe teacher, WOO introduction to one or morerepresentative administrative systems should beincluded. A student record system would probablybe a reasonable choice for illustrative study; itdeals with information familiar to the teacherwithout using financial details some might finddifficult to understand.
Teachers should also be familiar with super
1 2
132 NECC 1980
calculator modes of computer use as a classicapplication. As home computers become morecommon, perhaps little formal work in this areawill be necessary. Clearly,word processing mustbe covered= every teacher does so such word pro-cessing manually that none should be left ignorantof how much word processing help the computer cangive.
With respect to T1.4, the long range and theimmediate implications of computing as a form ofartifical intelligence should be taught. Theexcitement of learning to think about thinking andof contemplating the powers and limits of humanintelligence are so significantly linked with com-puting experience that this aspect must bestudied; the opportunity is too great to pass up.Artificial intelligence pay best convey both thepower and limitation of computing in education.Acquaintance with any of several perspectives onthis experience is essential and can be taughtusing such projects as the WOO work (7) or theSOLO work (I). A growing body of fiction aboutcomputing can also contribute effectively to theteacher's insight into the emerging world (9,10).
With respect to T1.5, teachers must knowwhere to look to keip abreast in this rapidlychanging field. Course work and instructionshould therefore routinely call attention to andrequire that the trudge, use a range of sourcesabout computing and eddcation.
The ideas presented above represent a minimalset of competencies which every teacher shouldobtain. The topics Oreeented provide a frameworkto achieve this minimal level of competency. Inaddition to these competencies, every teachershould also acquire the competencies listed ineither Section 2.0 or Section 3.0.
2.0 s Competencies needed for the teacher ofcomputing
While every classroom teacher should have thegeneral set of computing competencies suggestedabove, the teacher of computing needs more. Thelikelihood that he or she will, in addition toteaching, be forced to function as a generalresource to faculty, administration, and studentsonly increases the need for more extensive compe-tency in computing.
Since much of the knowledge required for sucha teacher is similar to that required of anyonedesiring to be a computer professional, many ofthe computer competencies defined in the recentACT4 Curriculum Committee report, "Curriculum '70"(3), apply to the teacher as well. This sectortherefore relies extensively upon that report.
2.1 : Competencies
The core material recommended for teachers ofcomputer science represents essential elementarymaterial, as well as material especially designedfor educators. Computer science teachers ehoulds
C2.1 be able to write and document readable,well-structured programs and linked
systems of two or more programs,C2.2 be able to determine whether they have
written a reasonably efficient andwell-organized programs
C2.3 understand basic computer architectonal
C2.4 understand the range of computingtopics that are suitable to be taught,as well as the justification for
teaching these topics,C2.5 know what educational tools can be
uniquely employed in computer scienceeducation.
The first three competencies are of the sortcommonly needed by all computing professionals,and are listed in "Curriculum '711" as among thoseto be covered by the undergraduate computerscience degree program. Competencies C2.4 andC2.5 are not orAmsonly needed by all computingprofessionals. They are essential only in thepreparation of computer science teachers.
lormal training in mathematics is
considered crucial for the teachers of computerscience. While the specification of this meths -matical content is considered beyond the scope ofthis report, it should be noted that this contentmay include topics which are frequently not partof formal mathematics programs (e.g. finitemathematics, statistics).
For individuals who are to serve as a com-puter resource person for their school or schoolsystem, two additional competencies have beenidentified.
C2.6 Develop the ability to assist in theselection, acquisition, and use ofcomputers, interactive terminals, andcomputer services suitable to enhanceinstruction.
C2.7 Ds able to assist teachers in evalua-tion, selection, and/Or development ofappropriate instructional materialsthat use computing facilities.
2.2 : Topics of StudyThese competencies should be gained through
a series of courses and other vehicles developedthrough joint efforts of teacher education pro-grams and the computer science program. We pre-sent below a list of topics that should beincluded in the program.
T2.2 Software Croanizationt Includes cos-pUter structure and machine language,data representation, eymbolic codingand assembly systems, addressing tech-niques, macros, program segmentationand linkage, linkers and loaders,
systems and utility programs.
1 el 3
ACM Elementary and Secondary Schools Subcommittee 133
T2.3 aardware Organizations includescomputer systems organisation, logicdesign, data representation andtransfer, digital arithmetic, digitelstorage and accessing control,reliability.
T2.4 Data Structures and illing_Processingsincludes data structures, sorting andsearching, trees, file terminology,sequential access, random access, fileI/0.
T2.$ Computers in Society: . Computers andtheir effects on governments, careers,thought, law, personal behavior)privacy and its protections informationsecurity and its preservation.
12.6 TeachilMi_CbmPuter Sciences includes(1) knowledge of learning theories asthey apply to learning about computers.(2) knowledge of several appropriatecurricular scope and sequences for avariety of program goals (e.g., liter-acy, careers, college preparation. per-sonal problem solving).
'Curriculum '70," along with avast amount ofresearch in computer education, supports theinclusion of topics 22.1 through T2.5. Knowledgeof programming topics, software organisation,etc., are essential for the computer professionalof today.
The teaching of computing is a unique cox»pater profession. Knowing how to program,however, does not, in itself, qualify a teacherfor teaching computer science. Materials on whyand how to teach computer topics included in T2.6are invaluable to the teacher of computing andshould be included within a program of studytraining such teachers.
Competencies C2.6 and C2.7 of the previoussection are required for those serving as computerresource personnel for a school system. Thesecompetencies should be gained through the computerscience program and the teacher preparation pro-gross. The following topics will assist in devel-oping the required competencies.
22.7 Winced Tunics, in computer, Sciencesincludes advanced topics in computerorganisation, operating system, archi-tecture) .datebase systems, computercommunications.
12.0 Computers, in Education: includesdetailed knowledge of learning/teachingresearch as it has implications foraffective design of institutional com-puting systems and administrative usesof computing in our educationalsetting.
including study of computers in educationwill increase the teacher's ability to as:Me arole of leadership in providing direction toeschool system in integrating computing into itscurriculum. This additional computer backgroundshould allow the computing teacher to act as aresource person to assist in fostering development
and implementation of cosputing throughout theschool, even when the other teachers know nothingof computing.
3.0 Subject specific Computing competenciesneeded by teachers
In addition to the set of universalcompetencies needed by all school teachers, thereare additional level- and subject-specificcompetencies which teachers should have. Anyteacher will require at least one of these, butno one of them will be universally appropriatefor all teachers. The definition! of thosecompetencies spring entirely from the vast andhighly diverse body of experience with usingcomputing in education over the last decade.Seeress representing SOW of this work are listedin the bibliography. The competencies can bestated generally, irrespective of the teacher'seventual level of subjects the topics, though,will vary considerably, depending on both.
3.1 ComPotenoteeIn terms of these subject-specific
amp:tingles, the teacher should:C3.1 be able to use and evaulate the general
capabilities of the computer as a toolfor pursuing various discipline.. orlevel-specific educational tasks;
C3.2 be able to use and evaluate alternativehardware and software systems designedto function as tutors or teacher aids;
C3.3 be familiar with information and quan-titative techniques of study in the(teacher's) subject.
These competencies should be developed bythe teacher preparation program, tailored to suitthe trainee's intended teaching leiel and
subject. Vs will not present an exhaustive listof topics corresponding to these subject-specific and level - specific uosPeunotas.Instead, we will present model topics for a fewselected subjects end levels.
3.1.3 -Libidos of Stu& for Teachers of EarlyChildhood arisary_grasea_imel-- TICTEC3.1 Computerises wimplh.gt opens -
gap Experience with a wide rends ofcomputerises but computing-relatedactivities that children van partici-pate in to enhance their readiness tounderstand and work with cosputers.
21C3.2 games and Simulations: Experiencewith a wide range aflame end simula-tions that stimulate children toexplore and better understand funda-mental concepts and strategies oflearning.
T1C3.3 Tutorial patens: Soperience withsimple and complex tutorial systemsfocusing upon mathematics, spelling,reading, and other elementory topics,including bi- lingual variations ofsuch systems.
1 4
134 NECC 1980
=3.4 exploratory, programming systems:
Experience with well-developed explor-atory systems where child-appropriateEA) subsystems such as robots are pro-grasmatioally manipulated by the childin a discovery or problem- solving
approach.Less work has taken place in the area covered
by 14E3.1 than one might expect. Despite thelikely wide-spread availability of microproces-sors in the immediate future, computer less com-puting activities can still be very useful, Sycontrast with heavily machine-dependent activi-ties, they provide a more contemplative, lessinvolved opportunity to examine some of the funds -mental ideas connected with computing. They thusallow those using them to deepen their under-standing even if titet have access to computers.Typical examples may be found in some of Papert'swork (ll) and in Taylor (12).
Vast quantities of games and simulations areavailable for TOC3.2, but careful choices shouldbe made in selecting them. Many are not well -written, either from a programming point of viewor from a child-user point of view, and no game orsimulation should be selected unless it succeedsin both areas. Some of the best work in this areaat this age level came out of the People's Com-puter Company under the initial stimulus ofAlbrecht (13).
With respect to TEC3.3, though,many tutorialsystems have been developed, not all of them aregood. The experience of the teacher shouldcertainly include at least one good system andsome discussion of what lies behind it. The workof the CCC group under Suppes is certainly aworthy example in this category (14).
Finally, new exploratory systems relevantfor 11C3.4 are smearing, but the pioneering workis still only for illustration. In particular,the =MCC work which produced SHALLTALe (15) andthe LOGO work, particularly as Seymour Papert hasadvertised and sustsined it (16), is outstanding.
Microprocessors can be the primary machineused, but not so much that the trainee is leftignorant of the advantages of larger systems.
3.2.2 : Tbeics of Study for Teachers of ForeignLanguage --SW.TP53.1 Genes and Simulations: Experience
with a wide range of games and simula-tions designed to provide culturalbackground and informal languagelearning, using the culture as contextand the language as the medium of com-munication in the gems or simulotion.
4114.2 Tutorial Systems: Experience withtutorial systems designed to enhancethe learning of a foreign languagethrough a carefully arranged body ofinteractive experiences driven by com-
petency-based, computer-administeredtesting.
TP53.3 perste:: language, Text Editing: Expe-rience with a powerful text editor
used to create and manipulate texts ina foreign language.
Under TFL3.1, simulations based on relevantactivities and situations in the language-culture can provide insight into the languagsdifficult to obtain in other ways. these, andmany popular computer games, should also be used
provide a more informal language practice forlearners. This practice can take two forms: (1)
translatinq and (2) informally using thelanguage. Teachers should practice translatingall user text, of appropriate games and simula-tion, into the target language, thus preparing tohave their students do such translation.Teachers should also have wide experience withplaying games whoa, text is entirely in the tar-get foreign language and which expect all playerresponses in that language. Such playing in aforeign language can be a valuable informalenhancement. Experience with a wide range ofsuch games and slwulaticns may else suggest new,more appropriate ones which the teacher canCreate or have others, including students,create. Some attempt to create such new material(or new variations of old material) should bepart of every foreign language teacher'straining. Naturally, where audio is available,it should be appropriately used.
There are many examples of language drilland practice suitable for use under TFL3.2. Worksuch as that done by Suppes (17) at Stanfordshould certainly be familiar -to languageteachers, though alternatives certainly exist(18). Work in this area should rely es heavilyupon audio and graphics as possible, thus cuttingdown on the automatic tendency to always trans-late from the native language and to developcompetence only in reading' and writing theforeign language.
TFL3.3 should ensure that teachers use asuitable computer text editor to manipulate textin the target foreign language. With appropriateaccent mark capabilities, such editors canencourage language learners to practice much moreprose writing and thus enhance their overall com-mend of the language.
3.2.2 : 'lbpics of Study for Teachers of PhysicalScience -- TPSTPS3.1 Exploratory programming Systems:
Experience with well-defined explora-tory systems through structured,discipline-appropriate languages.Systems meet include graphic capa-bilities, be programmahly control-lable by the student, and be orientedto discovery through problem-solvingactivities relevant to the physicalsciences.
TPS3.2 Tutorial Systems: Experience withtutorial systems designed to enhancelearning of the physical sciencesthrough a carefully arranged body ofinteractive experiences driven by
1'45
ACM Elementary and Secondary Schools Subcommittee 135
7253.3 Wass and Insulations Brperience withetand-om gams and simulationsdesigned to enhance understanding ofspecific physical phenomena or signi-ficant,past experinents.
7293.4 Clammog/Laboraberyexperience with automated sanagesentof people, learning, time, andresources including automated inven-torying, laboratoryinfornatim/referance system; ingeneral, uses of the cosputer toprovide the science teacher with moretime to west with individuals.
1123.5 Date Collection and Analysis: expe-rience with :systems which collect and
analyse date including on-linegathering and analysis of experimentaldata and process control systems whichcollect, analyse and provide feedbackto the system.
Tape possibilities under all three topics forPhysical science teechtcs have been extensivelyexplored already by Bork (19) and others (20).their careful work and well-documented analysisshould be extended and incorporated in thetrailing of physical science teachers. Suchtraining should include experience with relevantexamples which illustrate the three topics; itshould also require each trainee to constructselected, similar, small modules of computer-supported instruction as a normal part of teachertraining.
=MAWThis paper has addressed the computing compe-
tencies needed by pre-college teachers. Thesecompetencise are listed in three groupings basedon the involvement of the teacher in computer-related activities. Because of the variation ofsubject matter, implementation magpies have beengiven for teachers of- grades 1-4 (early
childhood), for teachers of foreign languages andfor teachers of physical science. These examplesare not meant to be In inclusive but to indicatethe level of necessary background knowledge.
Before graduating from a teacher trainingprogram, all teachers should be required toacquire the first sat of competencies and either,the second or third set. This requirement willprepare each teacher to use computing in theclassroom.
No aotempt has been made to package the com-petencies into specific courses. It is felt thateach environment will possibly require a differenttechnique for the introduction of the material.
Another group of individuals that need to beconsidered are the currant in-service teachers.The competencies described above are as importantfor them as for our future teachers. In-eervice
courses must be developed to provide the indicatedbackground for bi..8011,100 teachers.
List of ContributorsVivian Coon, University of Missouri -RollaRichard Dinnis, University of IllinoisDan Isaacson, University of OregonDavid Musson., University of OregonJohn W. Hamblen, University of Missouri-RollaAtone R. Kamorewskii Bremen High School,
Midlothian, IllinoisJames Lockard, Buena Vista CollegeDick Ricketts, Multnomah County Education
Service District, Portland, OriginStan Troitman, Wheelock College
Cladd References1) Poriot, James, James Powell, John Romblon,
*Mart Taylor. "Computing Competencies forSchool Teachers -- A Preliminary Projectionfor the Teacher of Computing." NationalEducational computing ConferenceProceedings, Iowa City, 1979.
2) Taylor, R.P., John Hamblin, James L.Poriot, and James D. Powell. * ComputingCompetencies for Teachers -- A PreliminaryProjection for All but the Teacher ofComputing." National educational CosoutiftConference Proceedings, Iowa City, 1979.
3) Curriculum Committee on Computer Science(C3$). °Curriculmi701 Recommendations forthe Undergraduate Program in ComputerScience." (Preliginary Draft, dated Autumn1976).
4) Mork, Alfred. "Learning to Teach ViaTeaching theCOmputer to leech.' Journal ofComoutereased Instruction, November 1275,vol 2.
5) Sums, Patrick. "Computer- Assisted
Instruction at Stanford" in Man andComputer, larger, 1970.
6) Smith,' Stanley and Bruce Arne Sherood."Educational Uses of the PLATO CmpUterSystem." SCINC4* April 1976 , vol 192.
7) Papert, Seymour. "Teaching ChildrenThinking," Logo Memo 2, NUT AI LAO, October1971.
S) Dwyer, Thames. °The Act of Ideation*Blueprint for a Renaissance' Creative,
Information: Computers is fiction.Reading, Mass.* Addison4Iesley, 1976.
10) Taylor, lobes* P. (editor). Tam of -the.Marvelous pocking: Thirty -Five Stories ofCommuting. lbrrimtowa, 1447.: CreativeComputing Press, 1980.
11) Papert, Seymour. "Teaching Children to Beathematic:Ian* VS. Teaching Aboutleatbesetica.0 Logout,* 4, MIS AX LAS, July1971.
12) Taylor, H. P. 0Computerlase Competing forYoung Children or What to Do till theCoyotes Cones" in Proceeding of the lidWorld Conference on Omuta' in Education.Morth Melland/Amsterdam, 1975.
116
136 NECC 1980
13) IOC (collective authorship). What to DoMK You It t Return. Menlo Park, Calif.:People's Computes Company, 1975.
14) Stipp's, Patrick and Elisabeth decker. "Seel-Mitten Studies of CCC Elementety School Cur-ti:Mums 1971-1975." Computes CurriculumCorporation, Palo Alto, 1976.
1$) fey, Alan. "A Personal *mates far Childrenof All Ages" in itOriesdines of KR NationalConference, Soften, August 1972.
16) Papert, Seymour. Versonal Competing and theImpact on Education." Edited transcriptionof a talk delivered at the Gerold P. *legNemeelel Confatencs, as printed in the pro-ceeding', Camputinii in College, and Oftivet-Iljr 197$ and muga University of Bova,1978.
17) E4F'ss, Petrick, Sober! Salt h, Marian Beard."University-Level Competes - Assisted Instruc-
21) Conference on Isaiiithematicel Skills andlearning. U. S. Department of Beath. Edu-cation, and Welfare, Euclid, Ohio, 1975.
25) Lennie, J. Richard. "Undergraduate Programato Increase Instructional Competing inSchwas. ProoesdingeofIdathConferenceonCote sting in the Undergraduate Curricula,last ignBilbge Web* 1977.
26) Dennis, Z. Richard, Dillhung, C. andMitisnieks, 3. "Computes AotiVities in Secon-dary Schools in Illinois." Illinois Serieson Slucaticeig Applications of
24, Univ. of Illinois, 1977.27) 110.00014 D.M. "Problems of Implementation:
Courses for Pre- and In- Service education ".for on and fps In Secondary,
Schools. Northololland Publishing Co.,
1978.
28) IF/P Technical Committee for Education,Working Group on Secondary SchoolEducation: Computer Education for Teachersin Secondary Schools: Aims and Objectivesin Teacher Training AFIPS, Montvale, NJ,1972.
29) Zeuhrmann, Arthur. "Reading, Writing,Arithmetic, and Computing" in ;sprayingInstructional Productivity in Higher Educa-tion. Educational Technology Publications,1975.
30) minas, Stephen. "Computer Literacy: TheNext Great Crisis in American Education.".Oregon Computing,' Teacher, vol 6, no 1, Sept1971.
31) Noussund, David. "Report of the ACK TeacherCertification Subcommittee." SIGCSE Bulle-tin, vol 9, no 1, Dec 1977. PP-8-18.
32) Poirot, James L. "A Course Description forTeacher Education in Computer Science."sIGCsn Bulletin, vol. 8, February 1976, pp.39-48.
33) Poirot, J.L. and Groves, D.P. BeginningComputer, Science. Menchaca, Texas:Sterling Swift Publishing, 1978.
31) Poirot, J. L. and Groves, D.N. computeScience for the Teacher. ,amcjaca. Texas"Sterling Swift Publishing, 1976.
3$) Recommendations Regarding Computer k ughSchool Education. Conference Board of theMathematical Sciences, Washington, O. C.,April 1972.
36) Taylor, 'Robert P. "Graduate RemedialTraining in Computing for Educators."SIMMS Training Symposium Proceedings,Dayton, February 1979.
37) Taylor, Robert P. (editor). The Computes inthe School: Tutor, 221, Tutee. New Fork:Teachers College Press, 1980.
38) Technology in Science Education: The NextTen Yeats. National science Foundation,Science Education Directorate, Washington,D.C., July 1979.
39) Topics, k Instructional Computing, Aspecial publication of SIGCUE, 1975.
Invited Session
CAUSE PROGRAM AND PROJECTS
Chaired By Lawrence OliverProgram Manager CAUSE
National Science FoundationWashington, D.C. 20550
(202) 282-7736
ABSTRACTDirnation\about the Comprehensive
Assistance to Undergraduate Science Educa-tion (CAUSE) program and a brief histor-ical overview will be presented withemphasis on projects involving the use ofcomputers in education. Exemplary CAUSEprojects will be presented focusing on:1) individualized testing using micro-processors; 2) minicomputers in under-graduate laboratory science education; and3) computer generation of instructionalmaterials.
PARTICIPANTS (Listed in order of Presen-EWE:GU--
A Brief Historical Overview of CAUSEComputer-Related Projects
Larry OliverNational Science Foundation
Individualized Testing Using Micro-processors
Douglas MoCohmUniversity of California, Davis
Minicomputers in Undergraduate LaboratoryScience Education
Elisha R. HugginsDartmouth College
Computer Generation of InstructionalMaterials
Michael'P. BarnettBrooklyn College of CUNT
1 I c`137
Tutorial
DATA BASES - WHAT ARE THEY?
Arlan DeXockUniversity of Missouri-Rolla325 Math-Computer Science Bldg.
Rolla, Missouri 65401(314) 341-4491
ABSTRACT--WiTs7vorkshop will cover the three majorapproaches to ourrent data base hystems=hierarachica4vSODASYL, and relational.The relative advantages and disadvantagesof each approach will be compared. Inparticular IBM's Database System IRS, _-Cullinane System IDMS, and Cincom's SystemTOTAL will be discussed.
149
138
Computer. Science Education
REQUIRED FRESHMAN COMPUTER EDUCATION IN A LIBERAL ARTS COLLEGE
David E. WetmoreSt. Andrews Presbyterian College
Laurinburg, N.C. 28352
(919)-27636!2x367
Computer education requires a defini-tion of the group to be educated* the fa-cilities (hardware and software) withwhich the education is to be accomplished*the goals of the educational process: andthe level and rigor of that process. Inpractice* the first two factors often de-termine the last two.
On the undergraduate level, there arethree fundamentally different groups to beeducated: potential computer scientists*other science students* and everybodyelse. There is some agreement on theneeds of the first two groups. The needsof the last group the educated members ofsociety who are not mathematically or sci-entifically oriented* are not so welldefined. This paper describes the attempts* spanning a decade of one institu-tion to educate *everybody else.*
St. Andrews Presbyterian College iva small (600 student) liberal arts col-lege. In 3.969 we instituted a sciencecourse required of all freshmen* *SelectedTopics in Modern Science." This course isboth the terminal course for non-sciencemaims and the introductory course for allScience PRJOVS. We felt that the ever in-creasing role of the computer in societyrequired that we give all of our studentsa familiarity with computers. We decidedthat this could best be done by teachingthem the rudiments of a programminglanguage. At that time we had no interac-
139
tive capabilities and all dour work wasdone in batch mode. Software constraintslimited our choice of languages to PL/1PLC* or FORTRAN. Most of our students arenot mathematically inclined* and so wechose PLC because of its string handlingcapabilities.
Over the next seven years* until theacquisition of interactive hardware* ourcomputer education program evolved into afairly stable system. As our present sys-tem is a direct response to our earlierexperiences* a relatively detailed des-cription of our old system follows.
The program consisted of eightthree-hour sessions held weekly. Theaverage section size was about 25* andthere were usually six to eight class sec-tions. During the first week there was ashort introductory lecture* followed bythe ideas of algorithms and flow diagram-ming. Students were given a simple bil-ling problem its flow diagram and itsprogram and were asked to punch up theprogram* run it* and bring their output tothe next session.
At the second session the printoutswere discussed* with emphasis on the de-bugging process. Then non-formatted 1/0assignment* and unconditional branchingstatements were introduced. The assign-ment for outside work was to modify thebilling program to handle multiple ous
150
140 NECC 1980
tamers by a loop and to write ane run aprogram to calculate averages.
At the third session we introducedarrays and conditional branching state-ments. The students were to modify theiraverage programs to use arrays and also tocalculate the average deviation of theirset of values. In the fourth session weheld a review and then discussed commentlines and simple output formatting. Thestudents' assignment was to rewrite theiraveraging program using comment lines andformatted output to increase the under-standability of the program.
In the fifth week character stringswere introduced. The assignment was toproduce a program calculating the averageword length in a text. The sixth week wasa continuation of character strings. Thehomework assignment was to write and run aprogram converting clear text into pigLatin.
During the seventh and eighth weeks,each student was expected to design,write, and debug a program which he feltmight be useful to him in his academiccareer. Most students wrote such thingsas checkbook balancing programs, grade
.point average calculators, or simple sta-tictical packages.
This educational venture was success-ful in teaching students the rudiments ofprogramming, the strengths and wealraessesof the computer and something of the po-tential role of the computer in society.We were not successful in convticing thestudents that the computer vas a usefulteal, however, as most of etiom never usedit during the rest of their academiccareer.
In 1977 we ev:.:hased a DigitalPDP-11/60 computer with eight video termi-nals for stuetl-. use. During our firstyear wit% tails syst:....1 we did nothing morethan our former system by switchingto BASIC and using the "MR' series toteach the language. The homework assign-ments remained unchanged. Our work con-vinced us that there are certain attri-butes of an ideal curriculum "or teachingeverybody else about 1--vputera. I believethat there are six sue" attributes:
First, what is being done must be ofobvious applicability to the student.Many college freshmen are not mathemati-cally inclined and do not foresee any useto them of the computer as number crunch-er.
Second, the system must obviously becomputer dependent. In our old system,none of the assigned prograps would repaythe effort involved in writing and debug-ging the programs.
Third, the material must be capableof ,gradual entry. New knowledge should berequired in small increments. Althoughour ,original sequence was modified quitefrequently, we were never able to avoidone session in which the stuaents wereoverloaded with information.
Fourth, the system should beword-oriented instead of number-oriented.Most students are more comfortable withwords than numbers. For example, we foundthat interest in the computer increaseddramatically when we started to work withcharacter strings.
Fifth, the system must have under-standable error messages. For beginningstudents a cryptic error message is prob-ably worse than a simple statement that anerror has been committed.
Sixth, the system must possess theattributes of any good tool: versatility,usefulness, resistance to breakdown, obvi-ous efficiency, and feedback to the user.
We decided that the system whichwould best combine the attributes listedabove was word processing. Word process-ing includes programs to write text, altertext, move text within or between files,and to format text. The systems are notonly obviously computer dependent, theyare impossible without a computer. It isapparent that they will work with textualmaterial of any length or complexity.
The major advantage of word process-ing is its applicability. A freshman at aliberal arts college knows that L4 willwrite many papers over the course of thenext four years. Any tool that will makethat writint easier is welcomed.
Word processing systems are usuallydesigned for use by non-technical peopleand are structured so that much can bedone with a few commands. Additional ma-terial can be learned when its need is ob-vious to the student. There are no quan-tum leaps in knowledge requirements.Furthermore, because of VI,: users forwhich most word processing systems are de-signed, they are quite fail-safe, andtheir error messages are usually clear andself-explanatory.
1 51
We chose UOT, a line-oriented editoras our editing system. EDT is not aspowerful as other available editors, butit is characterised by simplicity and easeof use. jinn was chosen as our formattingprogram. It is capable of straightforwardformatting with a very few commands butcan be used for extremely complex format-ting jobs.
The textual material used in thecourse is a 25-page manual consisting ofthe following information:
IntroductionLogging onAccount securityOverview of the programsLogging offFiles and file naming
Editing.ProgramEntering the programCreating filesWriting filesAppending tc filesError messagesEditing, an overviewText displayFinding textAdding lines within textDeleting lines from textChanges within lines
Getting Hard Copy
Text FormattingIntroductionCase controlMargin controlCentering text
Displaying Files
Account Directories
Spelling Checker Program
Summary Table of Programs
Each new student at St. Andrews isassigned a computer account. To use thisaccount !*:. is necessary only tcego to thecomputer center and obtain the accountnueber and password. Each May, accountsare deleted unless the student requeststhat it be kept. At present 230 students(37% of the student body) have active ac-counts. About 22% of the students nottaking a coaputer science course or thefreshman course have active accounts.
Computer Science Education 141
The computer block lasts four weekswith four assignments. During the firstsession the introduction of the manual,editing program through appending tofiles, and getting hard oopy are covered.The students* first assignment is to acti-vate their accounts and write a file of atleast 100 lines in two different sessions.The student must hand in a hard copy ofthe file the next week.
For the second week the reading ma-terial is the remainder of the editingprogram section and the text formattingmaterial. This material is discussed andquestions are answered. The homework as-signment for the third week is to edit thefile produced during the first week sothat it will include formatting commandsfor case and margin control and for para-graphing. A copy of the edited file is tobe handed in.
During the third meeting the rema-inder of the manual is covered, and eachstudent is asked to bring to the fourthsession an output file (formatted text)from the file produced during the secondweek.
The fourth and final session is de-voted to questions from the class and workwith individuals. The homework assignment(which is weighted twice as heavily as anyof the others) is to hand in, within threeweeks, a copy of a computer - produced papersubmitted for grading in another class.
An advanced word processing documentis available in the college bookstore.The advanced document deals with addition-al features of the editing and formattingprograms. This last year 174 studentsbought the required introductory text, andabout 50 copies of the advanced text weresold. (*)
Student evaluation of the computerblock has been favorable. In our usualterm-end evaluation it received an averagerating of very good. Xi past years, priorto word processing, the computer block wasusually rated as good to fair.Furthermore, conversations with facultymembers outside the sciences indicate thatan increasing number of papers in upperlevel courses are being computer produced.Thus, although we have no hard data, wefeel that we are having some success atour primary goal, that of showing all stu-dents the utility of the computer as atool.
152
142 NECC IWO
In addition, we feel that word pro-cessing is an excellent computer applica-tion with which to show students the use-fulness, strengths, and weaknesses of com-puter systems.
(0 Copies of the introductory hand-out may be obtained by sending a self -addressed, self - stamped ($0.70) large man-illa envelope to the author.
153
Computer Silence Education 143
DEVELOPMENT OF COMMUNICATIONS SKILLSIN SOFTWARE ENGINEERING
John A. SeidlerJohn G. Reinke
Department of Mathematics/Computer Scienceunieers'4 of ScrantonScranton, PA 18510
(717) 9S1 -7428
INTRODUCTIONNow such will an individual's
programming skills be enhanced by anability to present ideas orally and inwriting? There is such more toprogramming tnan just writing programs.Yet the emphasis in computing andinformation science curricula is onprogramming witn little emphasis on manyof the other essential tasks that surroundand are included in the programmingprocess. Since 1975, we have been teachinga course- that provides a broader view ofsoftware engineering. Tnis course,entitled Computer Projects, is anundergraduate degree requirement that istaken in the senior year. AS part of thiscourse, students develop and document anapproved project of their choice and giveseveral ors presentations as well asseveral written reports.
Normally, these projects begin duringthe junior year. The projects can be donein a variety of areas but must have areculty\sponsor. Students are encouragedto seek out faculty from other departmentswho are interested in using computingresources in some way to expand onexisting projects. In any case, allprojects must have faculty sponsors whoare interested in the projects.
A key element in this course is itsorganization. During the semester, eachstudent must give two lectures about hisproject and write seven reports. Thecombination of the oral presentations withthe written reports produces a totaldescription of the project and makes thestudents aware of a more global picture ofthe programming process as well as givinghim a better understanding of the typesof things that programmers and systemsanalysts do besides grinding out code.
Subsequent sections of this paperdescribe the course and the rationale forvarious items_ that are included and the
reasons for the way he reports areorganized.
COURSE ORGANIZATIONThe course is organized around the
presentation of student projects. For thefirst two weeks the students are given ageneral orientation regerding what isexpected from them. Daring the next threeweeks the students make their first oralpresentations which moat be between 15 and25 minutes long and provide appropriategeneral. descriptions 6( the Protectareas. The presentation Should bedirected toward convincing the audience ofthe importance of the project And theimpact it has on its environment.
The middle three weeks of the courseare spent describing the various methodsthat can be used in the remaining reportsand the second oral presentation. Also,the students ate shown theinterconnections that should exist betweenthe oral presentations and their variousreports. The second half of the semesteris devoted to the second oralpresentations. These are 45 to SS minutepresentations on the technical details ofeach project. This talk emphasizes theappropriateness of the technique selectedto solve the problem.
While the above sequence is occurringin class, the students Submit reportsabout every two weeks. As a collection,these reports are supposed to form acomplete description of the project.
THE ORAL PRESENTATIONSBoth oral presentations are evaluated
by both students and faculty member'present. Students are greded on a varietyof factors such as content, the use ofvisual aides, and the appropriateness ofthe level of presentation for theaudience. The first presentation is a 21minute survey of the area of the project.
_I 'F
144 NECC 1980
rue purpose nere is to acquaint tne
audience wits tne problem area andconvince tnem of tne importance of tne
project.Tee second presentation is a 40 to 52
minute presentation on technical aspects,hardware, data organization, andalgorithms. Part of this presentationmust InclUde, when appropriate,structured walkthrough of a part of tire'project.
REPORTSme seven reports, when taken as a
unit, should form a complete descriptionof the project. Content of tnese reportsis a compromise between an ideal and whatis realistic witnin the time frame of a
semester. In all cases, the reports arejudged not just on their technical meritsbut also on the organization, grammar, andspelling.
Report Ono. lids report is an
exten ed abstract of the project andshould be about throe to five typeddouble-spaced pages long. the studentsare told to relate this report to theirfirst presentation. Specifically, thefirst talk should elaborate on the generalproblem area while tea extended abstractshould only briefly mention it.
Report Two. This tnree page report isa requirements analysis. Here the studentmust describe the hardware needs of hisproject and describe now he came to hisconclusions. To wqp the students, a listof general questions is distributed. Thislist includes:
i. Wnat resources does the projectrequire? (Primary memory, disk,tapes, etc.)
2. Time requirements - What effectdo various variables have on thetime requirements?
3. Terminals - CRT, hardcopy,graphics, hign quality hardcopy-Which is ideal?-Which are satisfactory?-Baud rates?-Synchronous, asynchronous?-Intelligent terminals, etc.
4. Program space requirements?5. Backup
Report Three. This lip a justificationdocument. The' students assume that theyare writing a document in response to a
request by corpo ate management forjustification of continued funding of the
project. Here the students are requiredto defend the value of their project overthe status quo. Explanations of details
must be given in a way which would beunderstood by a semi-technical manager orexecutive. As a suggestion, students aretold that they should produce figures thatdescribe their project versus the statusquo, then emphasize how these data arearrived at legitimately.
One purpose of these first three lbt reports is to give students a chance tocollect their thoughts .during the firstfive weeks of the semester. The rest ofthe reports demand a considerable efforton their part and require a thoroughanalysis of their projects.
Report Four. This report is acollection or-Teur analysis and designdocuments. The first is $ systems chartin the form of a tree structure thatdescribes the top-down organization oftheir system. Included with this chart is- brief description of the purpose of eachmodule in the System.
The second part of this report is adata chart. The data chart is in theseparts--input, process, and output. Theinput section describes the files used asinput to the system along with the recordformats of inpbt files. Each item in arecord is defined in terms of its type andwhat role it plays in the system. Theoutput section describes both the outputfile and update files. By an update filewe mean s file used both as an input andoutput. As in the input section, allrecord formats and items in the recordmust be completely described.
The process section describes thetype, bound, and purpose of every major
'variable used. This section is groupedfirst by describing global variables,then common variables, then a break downof other variables by module. By majorvariables, we mean that, for the sake ofbrevity, the students can selectivelydecide to limit the size of the data chartby not including all variables, forexample, temporary storage and loopindices.
The third component in report four isa 4Ata flow graph. Again, due to timelimitations, the students are not requiredto create a data flow graph of the entiresystem. Rather, they use severalimportant items and show how they aretransformed and combined with other datato produce the results. The data flowgraph completes the systems chart in thatthe systems chart shows the tree structureof a system.
The fourth item in report four is abibliography. This bibliography does notjust state the references. Instead, abrief statement mast accompany tech ,
.5,5-
reference describing the relevantinformation obtained from that reference.Also,. the bibliography does not list onlytexts and periudicals. It also includesreferences to individuals who haveprovided information or assistance. It isgrouped as follows:
1. Text References2. Popular Periodical References
(Time to Creative Computing)3. PiEBNicef1717raical Re erences
(all other periodicals)4. individual Citations (persons
who assisted)With report four, we have a broad
based view of'the project. This reportacts as a framework upon which the rest ofthe reports can be built and integrated.
Report Five. Report five is a user'smanual, esarr$7 the most difficult reportfor the students to write. Thefundamental problem is that_they know toomuch about their project, and they lack anappreciation of the difficulties thatothers encounter (especially those who arenot experts in computing) in simply usingthe project.
The key to improve this report is anemphasis on well-constructed ,examples.Even this is a problem, however, becausethe students usually create unusualesoteric examples that are not typical ofnormal use of their project. That is,their examples are often like the horn :leexamples we find in manuals that areproduced by the various computermanufactures.
Another problem tnat is evident inthis report is poor grammar. in order tocorrect this with a minimum effort, thestudents are told that the report isgraded using the basics in Strunk andWhite's Elements of Style. Also, some timeis spent on some the asics.
A final emphasis in this report isthe importance of stating everything.That is, many times students assume thatthe readers of their reports have the samebackground and knowledge that the authorshave. As such, 'their report is notcomplete because there is still asubstantial amount of Knowledge that islocket; up in their heads and not stated inblack and white.
As a complement to report four, thisreport explains the details, criticalelements, and various interrelationshipsin the system. Here, students areencouraged to find the weaknesses inherentin the various graphs and charts of reportfour.
Computer Science Education 145
Report Six. This is usually theeasiest report for the students to write.They are so wrapped up in the technicaldetails of their project that the volumeof material that can be includedcontributes significantly to thegeneration of this report. However, thereare still problems they encounter inconstruction of this report.
Their biggest problem is one oforganization. The students are often sowrapped up in the project that they don'tknow where to begin writing.Surprisingly, they find it difficult touse the sections of report four as anoutline for this report. Once theyrealize the usability of report four ingenerating report six, this report becomeswell organized and easy to read.
Re ort Seven. This report containsfour sect ons7-61- first section consistsof annotated and documented programlistings. The documentation shouldexplicitly indicate the connectionsbetween the code generated for the projectand the various reports. Section two is aset of corrections to reports four, five,and six, which allows students to modifythese reports ..o stay in line with the waytheir system has evolved.
Section three of this report isanther difficult section. Here studentsare required to make a critical evaluationof their system, pointing out flaws anddeficiencies both in their overall designand implementation. This is followed bysection four, which describes what waslearned through the project. Included inthis section are recommendations formodifications, enhancements, and ideas forfuture projects.
OSSERVATIOHSThere is no doubt that this course
gives the students a broader appreciationof the entire hardware/softwaredevelopment process. If there is a flaw,kt is not in this course: rather, it isthat preceding courses do not require moreof the development process so that thiscourse will be less of a shock tostudents.
Typical of the variety of projectsthat students have done in this courseare:
1. Developcant of relative andhierarchical data basesimulations
2. General purpose software sciencemeasurements system
3. Programmable calculator use inthe physics laboratory
hidden line eliminationS. Software for forcasting the
university's financial statusBy allowing a broad collection of
projects, wa have obtained an interestingby-product. Faculty from otherdepartments, as well as variousadministrative offices, are more aware ofthe potential of our computing resources.Therefore, despite a stable, or slightlyshrinking student population, we havecreated an increased demand for computingresources. This has given us anopportunity to obtain the variety ofcomputing resources that we need tosupport our computer science program.
Computer Science Education 147
.1'
SYSTEMATIC ASSESSMENT OF PROGRAMING ASSIGNMENTS
Judy H. BishopComputet Science Division
Univetsity of the IlitmeetstandJohannesburg 2001
SOUTH AFRICA(tel. (11)-394011)
SUNHARYThe obvious way to assess a course on
programming is to assign merles to programswritten by the students. However, it is notevident st the outset that systematic assessmentis feasible. Nor is it necessarily ap:4rentwhen features the meeker should look for and theweight to be attached to them. This paperexplains a method,taat has been used for fouryears at two universities to mark course work atsecond year level. The experience gained showsthat systematic assessment is possible and canbe used to provide guidance to students as to
what constitutes a good program.
BACKGROUNDThe second year course in advanced program-
ming (Barron 19751 is the backbone of the
computer science degree at both the Univetsityof Southampton, England, and the University of
the Witwatersrand, South Africa. Since 1974, ithas been taught with the aim of turning out notjust good, but excellent progtammers. LikeNoonan (19791, we felt that the primary studentptoduct of such a course should be a single,
large programming project, split into foursmaller projects. Two such projects metedesigned to be used in alternate years, eachrequiring a total of about 1000 lines Of PascalMullis* 1974, 19751. Although e writtenexamination is required to tett a full knowledgeof data structuting end sorting techniques, thebulk of the final mark for the course is derivedfrom the project. At the outset, therefore, itwas realised that the marking of these projectsbed to be of a higher standard than usual.
Very little guidance in grading can be foundin the literature. Abehire (19781 in hisotherwise excellent advice for teachers, skirtsthe gables, stagily stating "Establish s gradingpolicy ... Erode each program ... Tell the
students your grading policy." In petticular,how does one neutralise the effect of subjectivefactors such es the student's previous perforemeoce, his presentation (nester essay), end thepreference of different markers for one or another
style of program layout?The methods used to matk the assignments foe
lets* first year closes* or small third yearclass,* do not seem approriate for a coitussized second year class (40-80 students).Program ossignmeets in first course can, forthe moat pert, be assessed by checking thee they
work for certain samples of data (Noonan 19791although some teachers ovoid this approach andamass only by s special fora of examination(Rsdue 1976, Troebetta 19791. At the other endof the scale, projects undertaken by third yearand greduste students can be looked stindividually and marked over a wide range ofcriteria (Hell 1979]. Second year assignmentsfell *members in the middle. One needs to givemarks for good ptogtan and data design as wellas for correct results, but the 'umbers involvedpreclude spending hours agonizing over eachprogram. In fact, the projects es set uprequire the first three parts to be handed in attwo weak intervals. Each pert builds on thenext so that it is desirable thee work handed inon Friday be necked and returned by Monday. Soas not to rely on the impossible, end alsobecause typically some 40Z of solutions era nothanded in in final fora, specimen solutions atedistributed and students era free to use theseinstead of their own in the next phase.Nevertheless, speed is a factot.
As final complication, the within may beshined between two or three people and thismeans that a standard must be well defined endunderstood. These thris requirements of quick,accurate, end equal narking led to the design of
the systematic method presented here.
THE METHODThe method quite openly borrows ftom a
technique used in faecal to mike programs morereadable. Instead of working in numbers,marking schema is sat up vein words or phrasesto deactibe the assessment of the solution undervarious criteria. Once all the solutions havebeen marked, weights are assigned to the wordsend 4 final sack calculeted. This mark is then
1 5S
148 NECC 1980
compared to the earlier estimates made onimpression. and the marks might be voderated one
way or the other. The six steps are1. Divide the solutions into five groups
based on evident output.2. Set up s symbolic marking scheme.3. hark all the solutions sccording to the
scheme.
4. Assign weights to the words used inmarking.
5. Calculate a numeric mark.6. Compare the estimated and calculated
marks and moderate if necessary.
1. Rough Groupie.Divide the programs into groups according to
the evident output.
A. Fully correct1. Answers partially wrongC. Incomplete output, perhaps with an execu-
tion errorD. No output - execution error in reading
phaseC. Compilation errors
These groups can be roughly translated intothe traditional classes as follows:
A : First (802 +) (unless layout ishorrible or algorithmsmessy or inefficleit)
B,C : Secondor third (502 -792) (the largest group -
ebould not get a Firstor Fall)
D t Thirdor fail (592-) (tricky group - they
must be marked oncriteria other thanresults)
Fail (492-) (depending on the levelof the course, thesemight get zero)
2. The Marking Scheme (Grading Policy)
A marking scheme is devised by selectingfive or more criteria on which the programsmight differ, for exempla, the algorithm, designof output, schievement, and testing. In initiale asignmente, the criteria mould emphasize pointsS uch IS nest program layout or the correct useof procedures and parameters. In later assign-ments, a quick Ounce st the solutions hooded inwill show whether thee* lessons have beenlearned and ere no longer worth assessingbecsuse the standard is uniformly high.
Each criterion is then assigned a list ofbetween two and five possible values. Theseare meaningful adjectives, spplicsble both tothe criterion and to the expected results. Atypical scheme would bet
answer. (right, some wrong, none,
still reading)output design 0 (as required,
improvement possible, shortcuts,unformatted)
algorithm (elegant, competent, average,messy)
achievement (more than enough, fullfilledrequirements, gaps left,far too little)
If there are no answers, the potential outputdesign will still-be assessed.
2. In a text formatting problem, "shortcuts"would be given if, for example, formattingsuch as centering or right - justifying wasdone by spacing to an absolute column numberinstead of being based on the length of theline. "Unformatted " - would be granted whenthere is no centering or justifying.Possible improvements would be small thingssuch as leaving a line at the bottom of apage before printing the page number. Therating of the output design should not beconfused with that for achievements.
3. The values given here for algorithm havebeen found to be applicable only when themarker is genuinely astonished at thestandard. A competent slgorithm is one inwhich-
- there are no redundant or out of placevariables
- control structures are used correctly (e.g.case, not cascading if's)loope cover short ranges (i.e. if a loop ismore than 20 lines long, part of it shouldbe in a procedure)
- the intention of each line is obvious,with comments only being necessary toindicate yet-to-be-included seztiens ormodifications that might be considered.
- the use of the language is quietly apt,avoiding over- simplification and clevertricks alike.
Most students' algorithms fell down somewhereand are classed Si average. A messy program-can be recognized a mile off!
4. At some stsge, one wishes to emphssize thatprocedures should be written, se far aspossible, es self-contained entities. Thisdoes not just mean that they should havelocal variables, but that they could becelled from different contexts and stillproduce the same results. This disciplineis particularly important in a course whereone assignment leads to the next. Ifprocedures are not self-contained, they willhave to be altered before being lifted out of
1 5 9
one program into another.5. The testing ctiterion is there to get
students out of the habit of handing in only
One tun. Very often instructors Prefer asingle tun because it is leas to mark, butin fact the opposite is true. It is easierto put the onus on the student to illustratealiat his program can do rather than have to
deduce this from the listing.
Careful considetation must be given to theinteraction of individual criteria. A primaexampl is that of testing versus achievement.If a feature is included, such as a test for anunexpected end-of-file, then a tun showing thatit works should be handed in. This can lead toa progtan being given
if the end-of-file was checked for, but only atun with correct data wan handed in. On theother hand, a ptogram that did not make thecheck would get
achievement = too littletesting = complete
because it had actually tested all that it coulddo. The weightings for these criteria (Step 4)would be assigned such that the first programwould get slightly more marks.
3. Harking (grading)The process of marking uses two tables. On
the first. values for the criteria in themarking scheme are filled in for each studentwith the last two columns, "matk" and "calculatedclass ", being left blank for the time being.On the second table, the group is recorded,alongside an "estimated class" and a comment.The estimated class represents one's gutteaction to the stmdents' efforts, and JO as faras nany informal lurking systems go. Thecomment lists any relevant factors that woadnot show up in the marks and could be used at alater date to explain border-line cases. Typicaltable entries would bet
TABLE I
NAME ANSUERS OUTPUT 'ALGORITHM ACHIEVEJ. Mullins right as required' average ful.req
I
Coo t ITEST Iter MARICALC.CLAS1partial yes
TABLE II
'NAM COMMENTheJ. MVIlinsl A
CROUP ESTI
. 'Gooduse of t language I
Two further lists are useful. One recordsthe names of students whose programs at worthdistributing and notes the points that the
Computer Science Education 149
programs illustrate. The other lists the
important points that should be commented on intutorials or in a handout, including moonmisconceptions or interesting approaches.
4. Weights
Only at this stage are numerical Valuesassigned to the atlas used in cIessifyins thestudents' solutions. Each vales in eachcriteria is assigned a weight, so that the left
hand value* add up to the desired maximumlark. Pot nest assignments, a maximum of 20 isadequate. Suitable weights for the samplescheme above are:
3 1 0answers (right, sommons, none, still reeding)
5
output design (as tequited,3
improvement possible,2 1
shortcuts, unformatted)6 5 . 3 1 (2)
algorithm (elesent, competent, average, messy)6
achievement (mote than - enough,
5 3fulfilled requirements, gaps left,
1
fat too little)4 3 1
testing (complete, partial, inadequate)1 0
eelfeontained (yes, no)
The maximum mark in this scheme is actuallymote than 20 because elegant algorithms endachievement note than required should berewarded by additional marks. These scorescould well be cancelled out by the studentfalling down on sons other aspect, such astesting. Note that even the worst program vestget some marks (for effort) so that sato dotsnot appear for every criterion.
5. Classifying
Simple summations give the actual !arks forthe first table. Prom these, classes ateassigned as follows:
CLASS I MARXI 80* 16+11-1 70-79 14-1511-2 60-69 12.43III 50-.59 10-11
P 20-49 5-10PP 0-19 0-4
6. ModeratingIn this final step the effort so far is
1 6
150 NECC
rewarded by comparing the ,three assessments:
- the rough grouping based on visible resultsthe gut reaction class estimate
- the mark calulated on detailed criteria.If there are violent disagreements between
the estimated and calculated classes, the
weightinee should be adjusted first. Adjustmentwas made in the above scheme because a programwith a very messy algorithm got 16 marks andreally should not have got a first. Thereforemessy was given a weighting of 1 instead of 2.
I II-1 11-2 III F
Rough 12 ( 24 combined ) 9
Estimated 11
1
5
.%**.s.qiw
13
Calculated 12 3 10 6. 10
The above table shows a comparison from anactual assignment given to 41 students. Onlyone program improved its rating on thecalculated mark, Ails eight were demoted. Thevalue of the calculation method is evident inthat close examination showed the estimates tobe faulty. Three programs in Croup C had tiny
errors and would have worked soon. They wereestimated at II-2e. Then the scheme showed thatthey had shortcut or unformatted output in
addition and correctly placed them in class
III. On the other hand, another C program had
top marks for output, algorithm, and achievement
which compensated for the slightly wrong answersand moved the class from a II-1 (estimated) to a
I (calculated).
EVALUATION
Although the method takes a while to
describe, in practice it is a stream- linedprocaas ehich can enable 60 assignments of about300 lines each to be accurately assessed in
about twelve hours. /he setting up of the
scheme takes about twenty minutes and assigningthe weights about tan minutes. The grouping,classifying, and moderating take about an hour.Men this investment, the individual solutionscan be marked in ten minutes each. This
compares favourably with the norm of thirtyminutes for an examination script.
An important advantage of the method is that
it almost eliminates prejudice of any sort.The example abtve was taken from the experienceof a marker who is usually quite generous andshowed that some 20Z of the estimates were toohigh. On the other hand, there are some markerswho equate "very good" With 7 out of 10 and
have a mental block against giving full marks.With this scheme, they find that they canjustify "complete" for testing in their minds
and let the calculations take care of themselves.
One of the criticisms of the method may bethat it does not encourage a fine enough net.
In nearly all cases, the weights in the above
example jump from 5 to 3 to 1. This is
intentional. After four years experience, wehave found that one cannot make the fine
distinction between en algorithm that isexcellent, very good, good, fair or poor. Theseterms are too vague and are avoided in favour ofcompetent, average and messy, with which one canidentify a program more readily. Thus the marks
do tend to be discrete and categorized ratherthan continuous.
Although the method does not provide anybuilt -In checks against plagiarism, the fast
pace with which programs can be marked meansthat one's memory is capable of detectingstriking similarities. In four years, one caseof genuine plagiarism was detected and proven.Numerous similar programs were found, but byreferring to the scheme, it was established that
they differed in One or more important details.Using the method has definitely increased
our understanding of what constitutes a good
program. Ass result, our teaching has improvedand certainly the standard of the studentprograms has risen over the years to such anextent that me would disagree witli Dolma andPozefsky (19791 who state that "Student- written
Programs accepted by computer science instructorsare usually inferior to programs which exemplifycurrently-accepted 'good' professional practice."Our experience is that students produce programswhich are just as good, if not better, than thosein an "up-to-date programming shop".
Because the symbolic scheme does not haveactual marks, it can be shown freely to studentsso that they know what to aim for. The tablesof detailed evaluations are also invaluable incountering queries about results because the
words are all there to explain why the studentlost marks (e.g. shortcuts, inadequate testing).
COHCLUSICNThis method of grading is ideal for medium
sized classes where programming excellence isexamined by practical work. It minimizes the
effects of prejudice and provides a balanced
assessment of the factors that are beingparticularly stressed at any one phase of theproject. It is adaptab'e to most types ofprojects and has been expanded and used withsuccess in more senior classes. It increasesones understanding of good programming, andenables one to communicate this to students in a
handy way, Finally, and moat importantly, itestablishes confidence in systematic assessmentfor both staff and students.
ACKNOWLEDGEMENTS
The Advanced Programming course vas devisedby David Marron of the University of Southampton
and it was at his suggestion that this method of
grading was designed, Thanks are due to the
1 62
Computer Science Education 151
classes of 1975 to 1979 who responded with suchenthusiasm to these ideas.
REFERENCESAbshire, Gary, "Techniques for Computer Science
Computer Programs: A Course in AdvancedProgramming", Computer Studies Group,University of Southampton, 1975. Revised1978.
Deimel, Lionel and Pozefsky, Hark, "Requirementsfor Student Programs in UndergraduateCurriculum: Hon Huth is Enough?", SIGCSEBulletin 11 (1), 14-17, February 1979.
Hall, Colleen, "Third Year Project HarkingScheme", Computer Science Report, Universityof the Witwatersrand, July 1979.
Hullins, Judy, "The Family Tree Project for
Advanced Programming", Computer Studies GroupUniversity of Southampton, 1975. Revised at
Computer Science Division, University of theWitwatersrand, 1978.
Hullins, Judy, "The Play Structure Project",Computer Studies Group, University ofSouthampton, 1976. Revised at ComputerScience Division, University of theWitwatersrand, 1979.
Noonan, Robert, "The Second Course in ComputerProgramming: Some Principles andConsequences", SIGCSE Bulletin 11 (1),
187-191, February 1979.Radue, J.E. "On the Teaching and Evaluation of
FORTRAN Service Course" SIGCSE Bulletin
(2), 32-35, June 1976.Trombetts, Michael, "On Testing Programming
Ability", SIGCSE Bulletin 11 (4), 57-60,
December 1979.
152 NECC 1980
DATA STRUCTURES AT THE ASSOCIATE DEGREE LEVEL
Richard P. DempseyComputer Science
The Pennsylvania State UniversityThe Worthington Scranton Campus
Dunmore, Pa. 18512717-961-4757
Recent design methodologies based onthe structure of the data and the develop-ment of data base management systems whichuse a wide variety of data structures addweight to the importance of data storageand processing. The fundamental roleplayed by data structures courses inapplied bachelors degree programs also in-dicates its significance (1, 2, 3).Associate degree programs that trainstudents to be quality entrylevelprogramers with sufficient background toadjust to new trends in data processingmust provide these students with back-ground in both internal and external datastructures (4). This paper describes thephilosophy and content of such a coursetaught as part of the associate degreecomputer science curriculum at theScranton Campus of Pennsylvania StateUniversity.
The objectives of the course are asfollows:
1) To present the structure andfunctional characteristics of externalstorage devices and their impact on fileorganization techniques.
2) To present concepts of internaland external sorting and searchingtechniques.
3) To present the advanced COBOLlanguage elements associated with theabove.
4) To present the concepts of in-ternal data structures.
5) To expose the student to the useof libraries and utility routines.
6) To make the student aware ofconsiderations in the design of datafiles.
This course, titled "Techniques ofOrganization," is offered in the firstterm of the students' second year. It isa three-hour course. Prerequisite
courses from the first year are:1) lst term. "Introduction to
Algorithmic Processes" This courseemphasizes solving problems using thecomputer as a tool. The language iseither WATPIV or PL/C.
2) 2nd term. "Computer Organizationand Programming! This course emphasizesthe binary and decimal instruction set ofthe IBM 360/370 Assembler Language. Theobjective is to get the student to see howthe computer operates at machine level.This provides a solid background to makethe student a better COBOL programmer anddebugger. All programs are run ou ASSIST(described in (5)1.
3) 3rd term. "Introduction to DataProcessing? This is not the usual courseusing this title, but a high-powered be-ginning COBOL course. As the studentshave programmed in two languages already,this course moves fast. It covers suchtopics as table handling, SORT verb,sequential files on disk and tape, andREPORT-WRITER. Emphasis is placed onprogramming style, documentation, andquality of code. Structured programmingis emphasized throughout. The programsare typical data processing problems.
The students now have a solid, basicprogramming background. Through the"Techniques in Organization" course, thisbackground will be reinforced while theylearn to handle data in new ways. Thematerial in this course is presented withan applications orientation and a minimumof theory. The objective is to give thestudent a working knowledge of dataorganizations and experience at using them.Some topics will simply be introduced andreinforced in following courses.
The order of presentation of thetopics in this course is not always thesame. The nature of the programming
163
assignments, shifts in emphasis due to newtrends in data processing, and the makeupof the class itself all affect the arrange-sent of topics. These topics, withapproximate times in parenthesis, are:
Topic 1. Indexed Sequential Files(1.5 weeks) A thorough working knowledgeof ISAM is expected. Some time is spent onthe physical aspects of ISAM files so thatthe students are aware of what is physical-4. happening. The pros and cons of ISAMfiles and the options available within itare discussed. The elements of COBOLneeded to process ISAM files are presented.
Topic 2. Direct access files (1 week)Both relative and direct files are covered.The concept of hashing is introduced. Thepros and cons of direct access files arediscussed in relationship to the other fileorganizations. The COBOL elements neededto process these files are also presented.
Topic 3. VSAM files (1 week) VSAMfiles are discussed, with major emphasis onthe KSDS format. The physical character-istics are presented and compared to thoseof ISAM. Comparison is also made at theCOBOL language level between ISAM and UMW.
Topic 4. Sorting (1 week) Externalsorts are covered briefly. One or twotechniques are looked at so that the stu-dent can see the nature of what takes placein an external :oft. No attempts are medeto teach the student how to write anexternal sort. Internal sorts are alsodiscussed. In the first year the studenthas programmed at least one internal sort,such as the bubble sort. Discussion hereis held to the relative efficiency of suchsorts in relationship to data set size. Asorting technique, such as Quicksort, ispresented as an example of a sort forlarger, internal data sets.
Topic 5. Internal Data Structures (3weeks) Problems are discussed that empha-size some of the limitations and problemsresulting from using only the basic inter-nal storage techniques covered in thelanguage courses. This discussion leadsinto a working level presentation on linearlists, stacks, queues, singly and doublylinked lists, trees, and inverted lists.Algorithms for handling some of thesefactors, as well as ways of manipulatingand allocating storage for them in COBOLare examined.
Internal and external data structuresare compared. For example, a good tie-inof the similarities and differences wouldbe an investigation of how a set of recordsthat need a tree structure relationship canbe stored on external files using variousfile organizations. By this time thestudents are gettihg a good grasp of dataorganization and its impact onprogramingefforts, which should provide them with a
Computer Science Education 153
good basis for what is happening in most ofthe data base management systems they willsurely face. Such knowledge is essentialto making effective use of these system.
Topic 6. Searching (1 week) A shorttime is spent on various searching tech-niques as they relate to the various datastructures. This discussion further en-hances the students' understanding of thesignificance of storing data in differentforms.
Slotted with these six main topics aresome others, basically for programmingpurposes. Partitioned data sets are intro-duced to allow the use of libraries and theCOPY verb. These libraries are created forthe students, but the JCL and controlstatements for IEBUPDTE are explained.. TheJCL and COBOL statements to handle sub-programs are also presented.
The programming assignments for thecourse are all done in COBOL. Students aregiven all needed JCL, but all elements ofit are explained so they see the reason foreach element of each JCL statement. Allprograms must be structured and completelydocumented. They are evaluated on qualityand correctness.'
Typical programing assignments are:Prog. 1. An input edit routine that
creates a sequential file on disk of thevalid records. This file is then sortedin a second step by calling the systemSORT/MERGE package via JCL.
Prog. 2. An ISAM file created fromthe output of Prog. 1. Libraries and theCOPY verb are introduced.
Prog. 3. A report produced by sequen-tially processing the ISAM file from Prog.2. This assignment introduces the use ofthe IEBISAM utility to allow keeping anISAM file as a sequential file. (Studentsare not allowed to keep permanent files onour system).
Prog. 4. It random update, includingadds, deletes, and updates, processedagainst the ISAM file from Prog. 2.
Prog. 5. A three-step program thatcreates, updates, and produces a reportfrom a relative file.
Prog. 6. A program to implement theQuicksort algorithm discussed in class.This assignment requires the use of COBOLsubroutines and stacks.
Prog. 7. A program to build andmanipulate a singly linked list.
Some of the topics, such as directfiles and trees, are not implemented inthis course. The next course, covering ad-vanced assembler, advanced debugging,utilities, and JCL, provides an excellentplace for students to write programs usingthese concepts.
154 NECC 19130
This course covers a lot of materialin a ten-weA term. Care must be taken notto get too seep into topics that don'tmerit the attention in relationship toother topics. Creating more than onecourse for the material is difficult asthere are only six basic computer sciencecourses plus a projects course in ourcurriculum. This limited number ofcourses results from the University'sstrong belief in the total edUcation ofthe student and concnrs with the guide-lines set forth by %he ACM SIGCSE paper onassociate degree programs (6).
A major problem with this course isthe lack of a suitable textbook to coversuch a broad range of topics. Most booksonly cover a subset of these topics endeven then tend to be too elementary or tootheoretical. The topic of internal datastructures has proven to be a significantproblem in this area. Barrodale et al (7)has a nice approach for our applicationslevel but is not detailed enough; mostother data structures books are too de-tailed an theoretical.
This course has kept many computerscience majors offthe streets a night ortwo, but the material and programs haveprovided a soli.: background for ourgraduates. They become productive almostimmediately upon employment in such areaaas application programming and technicalsupport. Several have been able to usethe course as a springboard into data basesystems. This practical applications-level study of the area of data storage andorganization has given them the experienceand confidence to be quality members oi!the data processing field and to adjustquickly to the changing demands of theirfield.
3. Fosdick, Howard and Karen Mackey, "ACot:se in the Pragmatic Tools of theProgramming Environment: Descriptionand Rationale," SIGCSE Bulletin,Vol. 11, No. 3, NUE7-1779, pp. 11 -13.
2. Mackey, Karen and Howard Fosdick, "AnA- lied Computer Science/Systerse __aiming Approach to Teaching DataStructures," S/GCES Bulletin, Vol.11, No. 1, Feb. 1979, pp. 76-78.
3. Beidler, John and John Meinke, "A.Software Tool For Teaching DataStructures," SIGCSE Bulletin, Vol.10, No. 3, Aug. 1978, pp. 120-122.
4. Little, Joyce Currie, "Computer Educa-tion and Community Colleges,"interface, Vol. 1, issue 1, Winter1979, pp. 12-16.
5. Overbeek, R. A. and W. E. Singletany,Assembler Language with ASSIST,Chicago: Science Research Associ-ates, 1976.
6. Little, Joyce Currie et al, "Curricu-lum Recommendation... in ComputerProgramming," SIGCSE Bulletin, Vol.9, No. 2, June-WM-pp. 17-36.
7. Barrodale, et al. Elementary ComputerApplications, Wiley, 1971.
Integrating Computing into K-12 Curriculum
THE SCARSDALE PROJECT
INTEGRATING COMP'ITING INTO TH8 K-12 CURRICULUM
Thomas Sobol, Superintendent of Schools, Scarsdale, N. Y.
Robert Taylor, Teachers College, Columbia University, N. Y., N. Y.
Introduction
Within a few years every child in
America is likely to have at least one
personal computer. The potential impact
upon schools staggers the imagination. At
the least, it is likely to move the focus
of education troll end product to process
and raise visual and auditory forms of
information to a status rivalling that of
written language. Because ideas can be
presented, explored, and expanded by human
interaction with the computer, computing
is certain to transform the schools from
kindergarten upwards, its impact will be
as broad and deep as any intellectual
innovation in recorded history, including
printing. In addition to traditional
communication, teachers and pupils will
communicate through tne computer
immediately and dynamically, in word,
picture, and sound, with each other and
with others throughout the community whom
they will never meet face-to-face. Thus
the nature of the communication will be
transformed and the range of participants
will be startlingly enlarged.
This stage of mass computer use waa
forseen years agot it arrives at last
because computers have been improved,
made smaller, and produced considerably
cheaper over the last five years. Yet
many of the questions raised in the minds
of those who foresaw what was coming
remain unanswered, How will use of
computers affs.ct the child's development?
What are appropriate languages for young
156 NEM 1980
children to use in talking
to computers? What devices other than the
traditional keyboard and screen or
terminal/printer can, when attached to a
computer, maximize its educational utility
for the young child? What should teachers
know about computers? Row expert should a
teacher be in computing to make reasonably
effective use of it in teaching and
learning? What is the teacher's role in a
classroom where every student has free
access to seductively engrossing computer
power? What is the impact of computer
games on the child and what row should
such games have in the school? When
should the computer be used as a tool,
when as a tutor, and when as a tutee?
Little significant research has been
done on most of these questions, even by
pioneers it computing and education. The
research that has been done focuses on a
narrow subsca of the issues such as the
effectiveness of computerized drill and
practice in arithmetic or spelling. There
are few clear answers so far; often what
are advanced as answers are little more
than strongly held hypotheses. There is
no ready-made curriculum, even in outline;
there are po coherent texts; there is
little available teacher training; and
there is little public understanding of
what changes computing might promote in
education.
Given all these unknowns, if a public
school system takes the transforming
impact of computing seriously, where does
it begin? Who does what with whom, at
what cost and with what type of success?
Though little research has been performed,
though many more questions have been
raised than have been answered, and though
nothing like a comprehensive intergration
of computing into the curriculum has been
formulated, the computers are here,
exciting possiblities beckon. To do
nothing is unthinkable. This paper
reports one school system's response.
The Scarsdale% Preipet
Two years ago students and teachers
in Scarsdale, NY, were doing little with
computers. Three or four terminals in an
old office in the High School offered
computer games to a handfotl of computer
freaks, the terminals were down as often
as they were up, and nobody seemed to
understand what the freaks were talking
about. Furthermore, inflatton and
declining enrollment had taken their toll
in the school system: summer school and
driver education had been abandoned, and
other programs were threatened. The time
.
hardly seemed ripe for a brave new
curriculum venture in computing or
anything else.
However the community does enjoy an
enlightened citizenry and teaching
faculty. Many were aware, however dimly,
that computers were transforming the
society and that they ought to be
transforming the schools as well.
Accordingly, in early 1978 the Scarsdale
Board of Education appointed a citizen's
advisory committee to make recommendations
concerning the use of computers and
computing in the schools. The committee's
report, which gave the Board a platform
for action, recommended that computing be
introduced into the curriculum throughout
the school system and that appropriate
steps be taken to purchase necessary
equipment and to train teachers in its
use. The committee also recommended that
the district engage a qualified consultant
to help in the effort. Work began
immediately.
We believed the project should begin
with the education of teachers --es many
of them, system-vide, as possible. Too
many school ventures in the past had
foundered because irre money had been
spent on expensive equipment than in
helping people prepare themselves to use
Integrating Computing into I12 Curriculum 157
it. However, district history suggested
that simply requiring teachers to take a
new in-service course would not accomplish
everything. Grass-roots support, from
both teaching staff and community, would
also be essential if the project were to
succeed.
19713/19744
In the fall the District convened a
steering committee consisting of about a
dozen teachers and principals. Committee
members were charged with keeping their
colleagues informed of developments,
reviewing4progress, drawing up statements
of assumptions and goals, and making
recommendations far the purchase of more
equipment when appropriate. With the
consultant, this committee reviewed the
Advisory Committee report, made plans for
the first in-service education efforts and
decided on immediate hardware purchase.
The committee swiftly acted to
acquire hardware, both to have it
available for the teacher training and to
satisfy some long-standing student demands
in the high school. However, the
committee decided to limit first-year
acquisition to a minimum in the belief
that they could better make such decisions
after they themselves had had a year of
1 6. cs
158 NECC 1980
training and experiencb with computers.
Because it was not clear that any one
microcomputer would be best for
everything, a mixture was acquired: four
Apples, file Pets, and one TRS 80. As
anticipated, the variety provided good
experience in making later acquisition
decisions.
As winter approached, Dr. Taylor met
with all teachers in the school system to
lorovide an overview of contemporary
computer technology and its implications
for teaching and learning. (The teachers
actually met in several groups rather than
in a single blocks English and Foreign
Language, Social Studies, Art, and Musics
Math and Sciences sand Elementary:The
purpose of these overview meetings was to
inform teachers about the long range
prospects and to build their enthusiasm
for beginning training. All teachers were
then invited to participate in the
in-service course scheduled for 8 two-hour
sessions throughout the winter months.
There were three main objectives to
this first courses (1) to introduce the
teachers to the rudimentary concepts of
programming, (2) to get them over their
fear and accustomed to using
sicrocomputers, and (3) to :mtter inform
them of some potential uses of computers
in the classroom. First, FPL. an language
developed at Teachers College to teach
programming, and BASIC were used to teach
the crucial elements of programming.
Second, scheduled use of the ten
microprocessors was built into the course
so that every teacher participating had to
use them to finish the course. Third,
selected articles by pioneers of computing
and education were read and briefly
discussed.
Formally, the course was offered
jointly by the district and by Teachers
College, Columbia University. The
participants could take the course for
Teachers College credit if they paid
tuition, for local school credit if they
paid the nominal teachers institute fee,
or for no credit, without charge. Of a
faculty of 330, 122 enrolled, including
several principals and the superintendent.
By the end of this course in late spring,
one third of the Scarsdale faculty had at
least some familiarity with computer
programming and some hands on experience
with a :ommon microprocessor.
Meanwhile, meetings were held with
parent-teacher groups and with members of
the local press to explain the purposes of
the project and to develop community
support. When time came to prepare a new
school budget, the Board of Education had
no trouble in i*,.:reasing its appropriation
for computer equipment and additional
training.
Determining the next steps was less
easy. It is one thing to provide some
teachers with an exciting introduction to
computing, quite another to modify an
entire curriculum by integrating computing
into it. The vision shared by the authors
and the steering committee was only
general: computing should be incorporated
into the curriculum. The problem was to
refine this general goal into specific,
detailed sub-goals. This task was all the
more difficult because most members of the
steering committee admittedly were not
experienced in current computing
technology let alone expert enough to
predict what computers would be like two
or mere years into the plan when the
curriculum changes might be realized.
Nevertheless, planning proceeded.
To clarify what was underway, three
documents (Appendices A, 8, and C) were
produced: a statement of assumptions, a
set of curriculum goals, and a rough plan
for managing project activity over a
four-year period. Though almost
immediately outdated by experience and
events, these documents and the thinking
Integrating Computing into K-12 Curriculum 159
behind them were essential foundations for
further thinking and for much of the
action that made them obsolete.
19/9/1980
During the summer of 1979 a small
group of teachers was paid to develop
programming exercises for fifth and sixth
grade children. (At this writing, at
mid-year, many pupils have already raced
through this material, despite the fact
that till quite recently it might have
been considered strictly high-school level
material.) Based on the experience gained
with using the first, mixed batch of
microcomputers, the district selected and
acquired 19 more: 15 purchased from
school funds, two donated by an outside
agency, and two purchased by parent
groups. Each school then had a minimum of
at least two machines. More were placed
in the junior and senior high school and
full-time aldes were hired there to staff
a computing center in each.
Th:oughOuc the school year the
Scarsdale Teachers Institute (the
in-service education teacher-run,
teacher-serving program which jointly
sponsored the first formal introductory
computing course) has offered a series of
mini-courses in programming in BASIC;
1 7 .
.1Ms.mmb,
180 NECC 1980
certain junior and senior high school
teachers (primarily in math and science)
have begun exploratory use of computers in
their classes; a concerted effort has
been made to introduce computing and
programming to all fifth and sixth grade
pupils in their classes; a selected group
of K-2 teachers have begun to explore
dynamic graphics applications in beginning
reading and mathematics in their classes;
and formal in-service work of several
kinds has been continuing. At the same
time, planning for subsequent years
continued on various fronts.
Units
What has been accomplished thus far?
To begin with, nearly one third of a
district faculty which two years ago knew
little or nothing about computers now
knows the rudiments of programming, has
lost most of its fear, and is beginning to
try new things. A small cadre of teachers
has become enthusiastic and increasingly
knowledgeable and gives promise of
leadership in the years ahead. Many
pupils from the elementary grades through
the high school have acquired a beginning
knowledge of the computer's capabilities
and of their own capacity to get the
computer to help them think. Computer
hardware exists in all the district's
schools, with more to follow. And there
is a broadening base of understanding and
support of the project throughout the
community.
problems eneauntoreA
Any innovative project encounters
inertia after the first enthusiastic push.
Scarsdale now faces the problem of what to
do about the two thirds of the faculty who
have not learned about computers -- and
what to do about the hundreds of
enthusiastic children who have learned
about them but must, perforce, be assigned
to the classes of teachers who have not.
There is the problem of providing support
to the willing teacher who is learning but-
who needs help. There is the ubiquitous
problem of money and of acquiring hardware
as rapidly as it can be used.
And there is one other, more subtle
problem that must perhaps beset all such
ventures in computer education until the
shape of the future is more clearly
revealed. Computers have the power to
change the ways in which people acquire
and extend knowledge, just as writing
changed such ways millenia ago. But until
we know better how those ways will change
and how a school should be reorganized to
capitalize upon those new ways, computer
17.1
technology is in danger of simply being
harnessed for the pursuit of present goals
within present modes of operation. To
employ computers thus is to miss much of
the revolutionary potential they represent
and to heavily damp the enthusiasm and
insight they might otherwise foster in
children. To finish by using computers
for little more than enhancement within
the traditional curriculum and the
presuppositions implicit in it would be
tragic. The challenge Scarsdale and every
other school or school district faces is
to be sure this doesn't happen.
,The Goal
Using computing to practice
mathematics and language skills in
traditional curricula is helpful but
primitive, almost like using televisor
primarily to display pages from textbooks.
Serious questions must be raised about
mere productivity of educational
approaches which limit visions of computer
application to such narrow horizons. Some
of the skills for which computer
assistance is so easily designed may
actually be of significantly less
importance as computer access continues to
increase. Hampered by ignorance, enticed
by expectation, Scarsdale seeks a grander
goal: to become so comfortable with
Integrating Computing into K-12 Curriculum 161
computing that our understanding of it
will naturally reshape the way we think
about everything. 4
Cencladjer schooldistricts
What might others learn from this
project's successes and problems? We
suggest the following:
1) Begin with people, not equipment.
The eager person who wants the
machine and is ready to employ it is
more valuable and catalytic than the
machine that no one understands.
2) Develop broad support throughout
your community and teaching staff.
3) Engage in on going, system-wide
planning. Do not leave matters to
an individual school or department
or to an outside consultant.
4) Use a consultant or consultants who
can bring not only technical
knowledge of computers, but an
understanding of schools and the way
people use their minds and
imaginations.
5) Concentrate resources of time,
money, and attention on the project.
In this way you can achieve a sense
162 NECC 1980
of purpose and direction despite the
shrinking you may be suffering
elsewhere in the system.
6) Don't wait for the perfect computer; 1)
current machines despite their costs
and limitations are a perfectly good
introductory device for teacher
training and pupil experimentation.
A great deal must be learned now if
better machines are to be well
employed when they do become
available.
7) Don't limit yourself to one type of
machine; get experience with
several. As cheaper, more powerful
ones bow in, you will have a better
experiential background by which to
judge them.
8) Find out what you can do with the
hardware as it stands, then do it.
Don't always wait for expert
software; both teachers and
children can learn powerful things
from exploring a machine's basic
capabilities to plot, edit, speak,
delete, and so forth.
Appendix A
assumptions re coo PutoXfiMAd
C81911th18irL-inaLLIIctiall
The influence of computing on human
life and on the entire planet's
development will continue to
increase exponentially for the
foreseeable future.
2) All pupils should be educated to cope
with this growing influence; as
many pupils as possible should be
educated to master computing and to
use it creatively.
3) Such educating toward computing should
be general rather than specfic
since hardware and software changes
will continue to occur at such a
rapid rate that specific training
will rapidly become obsolete.
4) Any human being of normal intelligence
can understari the fundamental
principles of computers and
computing, can operate computers,
and can learn to write simple
computer programs.
5) Pupils should begin learning about
computers and computing from the
time they first enter school and
should continue such learning
throughout the grades.
6) Pupils should also use computers as an
aid to other studies, in the
humanitits and social sciences and
the fine arts, as well as in
mathematics and natural science.
7) Pupils learn to use computers by using
computers. The use of computers
should be as thorougly integrated
into the school program as the use
of pencils or chalk.
8) Computers are both an object and an
instrument of learning.
Traditionally we study writing to
learn how to write, and we use
writing to refine thought and
express feeling. so too should we
study computers to learn how to use
them and use computers to extena
and refine our thinking.
9) If pupils are to learn about computers
and computing, their teachers must
learn about computers and
computing. If all pupils are to
learn about computers and
computing, all teachers must learn
about computers and computing.
Integrating Computing into K-12 Curriculum 163
10) Computer hardware will become
increasingly less expensive, and
both hardware and software will
become increasingly available.
11) Computers can serve as tutors,
teaching us skills and information
according to programs written by
others. Computers can also serve
as tools, analyzing data,
performing calculations, and so
forth. And they can serve as
tutuees as we teach them to solve
problems, compose music, etc. Bach
use has its place, but the most
important is the last.
12) Programming computers requires logic
and precision. Pupils who learn to
program computers practice logic
and precision.
13) The study of computers and computing
does not run counter to the spirit
of humane studies and the exercise
of free, creative intelligence. On
the contrary, it extends and
deepens them by extending the power
of the human mind and may foster a
much needed dimension to our
intelligence.
164 NECC 1980
Appendix Et
guriculum
1. From kindergarten through high school,
at increasing levels of
sophistication, pupils should learn
the fundamental principles of
computers and computing. Some
Pupils should learn to write simple
computer programs in elementary
schools all pupils should learn to
write simple computer programs
before the end of their junior high
school years. In grades 9 or 10
all pupils should complete a
one-semester course in computing,
computer science, and the social
issues raised by computer
technology. Other high school
computer studies should be oriented
towards computer skills which
students will need in college, such
as computer applications in
numerical analysis and statistics.
2. From kindergarten through high school
pupils should use increasingly
sophisticated calculating and
computing devices as an aid in
studying mathematics, science, and
other subjects.
3. Throughout the grades interested
Pupils should have opportunities to
extend and enrich their knowledge
of computer programming. In the
elementary schols these
opportunities might take the form
of after-school workshops or
special individual or group
projects; in the junior high
school. of projects( mini-courses,
or club activities; in the high
school, of specialized elective
courses, projects, or club
activities.
(Sow can we make this item more
specific?)
4. Interested high school pupils should
also have the opportunity to study
computer science, computer
technology, and data processing.
1 '70
Appendix C
Curriculum: Pour-Year Plea
1978-79
Elementary: None.
Junior High: Computer club plus?
High School: Computer Center open.Computers used inscience courses.
1. Three week mini-course forintermediate studentsrotates among fiveelementary schools.
2. Consultant and 2-4 primaryteachers work with K-3students.
Junior Highs
a. Each 8th grader has anequivalent of one weekmini-course oncomputers (hands-on).
b. Each 8th grader has two weekcomputer/literacycourse in spring aspart of social studiescurriculum.
High School:
Introductory programming (2semesters).
Advanced programming coursesdeveloped.
Integrating Computing into K-12 Curriculum 165
Computers used in science courses.
Computer applications in mathdeveloped.
Independent study projectsavailable.
Computer Club active.1980-81
Elementary:
Mini - courses for 5th /6th graders;most have hands-on timewith desk-tops duringboth years.Increased/experimentalactivity at primarygrades. Units forprimary grades includedin social studies,science curriculumguides.
Junior High:
Units on computers in socialstudies technologyunits.
Computer units in math courses bothyears.
Active Computer Club.
High School:
All before continues.
Increased computer use in socialstudies, business, andscience courses.
1981-82
Elementary:
Nini-courses for all 5 -6 graders,all have opportunityOr hands-on computerwork.
Activities for primary gradesincluded in socialstudies, science, andMath guides.
Junior High:
Continuation of 1980-81.
High School:
7G
166 NECC 1980
Continuation of 1980-81. Expandeduse of computers innon-math departments.
tzpi)*hrerehitirL PrpnitA
First Semester.
a. Consultant meets with staff invarious formats.
b. Guest speakers meet with staff.
c. Committee reads, discusses,meets, and argues.Engages others not oncommittee in similarconversations.
Second Semester.
a. Consultant continues to meetwith staff.
b. Consultant teaches literacycourse in Scarsdale.
Faculty visits other schools toexamine use ofcomputers.
Issues: What do we want kids toknows
What learnings are appropriate foreach age?
How do we best teach aboutcomputers?
How do we go about organizingourselves?
How do we involve the entire staff?1979-80
a. Consultant advises computercommittee and others.
b. Advanced seminar for those whohad first course.
c. Four session course on BASICfor novices andliteracy graduates(Sept./Oct.).
d. Follow-up mini-courses oneducationalapplications &programming led byin-house staff (Nov. -
February).
e. Course on graphics and music.
f. Guest speakers - Computerliteracy.
g.
h.
1.
j.
a.
b.
c.
d.
e.
f.
g.
Visits to other schools on anexpanded basis.
Social studies, science, andbusiness faculty gainexperience withsimulation materialsthrough BOCES course.
Send one or two to computerrepair school.
Course on programming for CAI(2nd semester).
1980-81
Consultant continues supportgroup on computers.
One or two faculty takeadvanced collegecourses.
Course in FORTRAN, etc.,offered to all withprevious experience.
Introductory BASIC courserepeated for those with-no computer experience.
Release-time workshops forthose with no computerexperience.
LOGO on the APPLE course forelementary schoolteachers.
CAI programming course.
LI19=78-79
1. Prepare statement ofaspirations.
2. Educate ourselves.
3. Design structure to involvesentire district.
4. Plan and implement system-widein-service program.
17
S. Plan 1979-80 computer budget(16,000)
6. Purchase 1979-80 computers andsoftware.
7. Design courses at high schoollevel.
8. Plan 3-4 yr. purchasingprogram for computers.
9. Design a mini-course forelementary students.
197v-80"
1. Continue system-wide generaleducation.
2. Provide follow-throughin-service forteachers.
3. Plan 1900-8i budget (21,000 forhardware, software).
4. Make 1980-81 purchases.
S. Design courses at high schoollevel.
6. Pilot units at juor highlevel during 2ndsemester.
. -
7. Pilot mini-courses atelementary school uppergrade level.experiment at primarylevel; expand primarylevel computer use,teach technology unit.
8. Develop maintenance procedures.
9. Integrate computers intoexpanded Sth gradetechnology unit.
10. Integrate calculators,computers into mathcurriculum.
11. Search for and write grant forcomputers.
12. examine software.1980-81
1. Focus in-service on advancedcourses.
Integrating Computing into K-12 Curriculum 167
2. Continued expansion ofcurriculum off4rings.
3. Plan 1981-82 computer budget.
4. Make 1981-82 purchases.1981-82
1. High school curriculum inplace.
2. Junior high curriculum inplace.
3. Elementary curriculum in place.
4. Write computer scope andsequence.
1 *;*fcI tj
168 NECC 1980
INTEGRATING COMPUTING INTO K-12 CURRICULUM
Beverly HunterHuman Resources Research Organization300 North Washington StreetAlexandria, Virginia 22314(703) 549-3611
ABSTRACTaliFFically, schools and school dis-
tricts have employed a wide variety ofstrategies to integrate computer-relatedactivities into the curriculum. Several,such strategies are reviewed, based upona national sample. The need for a compre-hensive computer literacy curriculum,K-12, is described. A plan to infusecomputer literacy objectives and activ-ities into the traditional curriculum isset forth. Implications of such a planon teacher training and curricular mate-rials me discussed.
Catherine E. MorganDept. of Instructional Planning andDevelopmentMontgomery County Public Schools850 Hungerford DriveRockville, Maryland 20850(301) 279-3321
ABSTRACTUgger-based instruction in the Mont-
gomery County Public Schools, a largepublic school system, includes mmputer-assisted and computer-managed instruc-tion, computer literacy, computer math-ematics, and problem solving. The inter-active use of computers for childrenbegins as early as third grade while themanagement system in mathematics monitorsindividual student progress from kinder-garter.
Invited Session
FUNDING ACADEMIC COMPUTING PROGRAMS
Sheldon P. GordonSuffolk County Community College
Belden, New York 11784
Lawrence OliverNational Science FoundationWashington, D.C. 20550
ABSTRACTAt most institutions, the single most
critical problem in the development, imple-mentation, maintenance, or expansion of anyacademic computing program is lack of moneyThis presentation will focus on a varietyof possible solutions to this problem. Anumber of alternate sources of funding,including grants from private foundations,grants from government agencies, andcontributions from local business andindustry, will be discussed. Primaryattention will be devoted to the mostlikely source of outside funding, namelygovernment grants, especially thoseavailabl through the National ScienceFoundation. In particular, details willbe provided about the respective objectivesand limitations on those NSF programs thatwould most likely sponsor various types of
computer oriented activities. These wouldinclude CAUSE (Comprehensive Assistance toUndergraduate Science Education), LOCI
. (Local Course 'mprovement), ISEP (Instruc-tional Scientific Equipment Program), andMISIF (Minority Institution ScienceImprovement Prograzi",.In addition, the presentation will
include a discussion of some of theimportant do's and don'ts connected withproposal writing, and will relate them tothe process of gre.nt review as conductedby the NSF,
169 1 S
Tutorial
TECHNIQUES FOR INSTRUCTIONAL SOFTWAREDEVELOPMENT USING MICROCOMUTERS
Kevin HausmannMinnesota Educational Computing Consortium
2520 Broadway DriveSt. Paul, Minnesota 55113
ABSTRACT7%saWocomputers become more and moreavailable and easy to use, more and morepeople will be designing and writing soft-ware for them. Before a project is begun,however, there are several obvious but oftengotten considerations:
1) Is this a reasonable application fora microcomputer?
2) What are the limitations of myequipment?
3) What modes of interaction shouldbe used?
4) What support materials may berequired?
Once the analysis phase is completed,the design and layout stage may begin.Design should center primarily on theorganizational content and division ofthe material. Also at the design stage,one should consider the mode of presentationof the material. So far, all the timespent on the project has not involved anytime programming the microcomputer.After the design is planned out, one
enters the ithplementation stage.Briefly, some items to consider hereincludes
170
1) Adequate spacing of material on thescreen.
2) Efficient movement between frames.3) Effective presentation techniques.4) Good ways to ask questi,ns.5) Effective feedback for answers.6) Adequate arS appropriate use of
graphics and other special featuresof the microcomputer.
After an application is developed andwritten, it is very important to adequatelytest the program. Many times this is bestdone by watching someone use the programwho is totally unfamiliar with it.
More detailed written material of thisnature is available from the MinnesotaEducational Computing Consortium, 2520Broadway Drive, St. Paul, Minnesota 55113.Ask for the Apple Authoring Guidelines.
Mathematics
A NETHOD FOR EXPERIMENTINGWITH CALCULUS USING.
COMPUTER-ASSISTED INSTRUCTIONFrank D. Anger L Rita V. Rodriguez
Department of MathematicsUniversity of Puerto Rico
Rio Piedras, Puerto Rico 00931(809) 764-0000
INTRODUCTIONThe project which is the object of
this report was carried out in the (a-cuity of Natural Science at the Univer-sity of Puerto Rico, Rio Piedras campus,during the years v.976 to 1979. The Na-tional Scienc. ration sponsored theproject through its Minority InstitationsScience Improvement Program (MISIP),making possible a wnole new aspect ofinstruction and learning for 1600 stu-dents. Although the ptssent paper willconcentrate on the computer- assistedinstruction modules developed for theintroductory calculus course that webelieve are innovative themselves, wemost begin by describing briefly a fewpoints which make tnis project, as av121e, unique.
First of all, as the compositionof the student body is that of theSpanish-speaking culture of Puerto Rico,the various materials of the projectwere produced in Spanish. In addition,experience has shown that although UPRmantains a selective admissions policy,entering students are frequently defi-cient in analytic skills and in expe-rience with methods of scientificinqciry. Inadequate or, frequently,incorrect preparation may lead a studen4..to frustration or failure when facedwith laboratory or problem-solving si-tuations.
171
Second, the project introduced oradvanced sweeping changes in the curricu-lum. Rather than simply paste on a littlecomputer-assisted instruction (CAI), wetook into account the full importance ofthe computer to modern science and inparticular to the working scientist. Thus,programming was made mandatory for allmajors in the faculty, CAI was introducedin all basic science courses, and inter-active computer utilities were madeavail-mble.for students of certain advancedlaboratories. In the calculus courses,greater emphasis was placed on the use ofthe calculator and on the numericalmethadeof calculus.
Finally, the project created the Na-tural Science Academic Computer Centerheaded by a member of the faculty andstaffed exclusively by students. Ratherthan implanting a ready-made system towhich students have recourse, we therebycreated an organic exterugion of the stu-dents' environment, providing her imme-diate incentives for probing deeper intothe uses and operation of computers. Thislast area has been one of the high pointsof Ale project. (4)
THE MATHEMATICS COMPONENTThe project developed in an orderly
series of stages, beginning with thetraining of staff, the search for appropriatehardware, and various necessary
172 NECC 1980
administrative measures. Although everypart of the project seemed to be fraughtwith unexpected difficulties, the facultyas a whole responded more positively thanexpected, and their cooperation along withthat of various student assistants keptthings moving forward at all times. Inthe Mathematics Department, the necessarychanges and investment of time were thegreatest, since to 'his department cellthe full weight of training professorsand implementing the new programming courserequirement for the whole faculty. We soonrealized that the applications in the dif-ferent discip'ines involved were betterhandled by U..* new programming coursesrather than by the old course (see Figures1.a and 1.b). The pre-medical and biolog-ical science students were better servedwith a programming course that would in-clude data gathering, statistics, andexamples in their own discipline. Thereare many packaged programs that, with someprogramming knowledge, the student mayeasily adapt to his needs. On the otherhand, the physics, chemistry,and mathe-matics majors were to be taught moresophisticated programming, also wit% exam-ples from these discipliner.
To train more of the faculty membersin PL/1, programming seminars were given.We chose the PL/I language in particularbecause we believe that it is extensiveenough that the student who knows it mayteach himself whatever langUage the ma-chine he has at hand uses. Parallel tethis, the calculus course was dividedinto two courses; the first, as in theprogramming course, designed to serve thepre-medical and biological science stu-dents and the second to serve physics,chemistry, and mathematics majors (seeFig. 1.a and 1.b). These divisions re-spond to long-standing curricular tensionsin a faculty in which about two thirds ofthe students will go no further inmathe..matics than the first semester of calcu-lus while the other third will need atleast two or three semesters of moreadvanced mathematics.
THE COMPUTER-ASSISTED INSTRUCTION FORCALCULUS
The major reason for the implementa-tion of CAI at the University of PuertoRico was to combat a lack of analyticalskills and to overcome, to whatever extentpossible in such a limited program, thetendency toward rote learning and memori-zation as the basis of learning. We de-sized, moreover, to avoid doing with thecomputer what could be done equally wellor better by more traditional methods,with the possible exception of the one-to-
one tutorial. Thus, we sought three basicthings from each module:
1. It should perform some graphicor numerical simulation of some funda-mental process, concept, or technique.
2. It should be highly interactiveand open ended, allowing the student somefreedom to experiment with the conceptunder investigation.
3. It should minimize the difficul-ties faced by the student using the com-puter.
These objectives have much more ob-vious implementations in the experimentalsciences than they have in mathematics,and we feel that there is still much roomfor exploration and development in thisdirection. Nonetheless it is herein thatwe feel much of the novelty and value ofour modules lies.
Math 103Precalculus I
4 credits
Math 104Precalculus II
4 credits
Figure 1(a) Previous Basic MatikaaticsSequence
/hdt 100Remedial Math 105
Math 1 Precalcuks3 5 credits
credits
Fiigu a 1(b) NewBasic Mathematics
Sequence
Math 200Calculus forBiology and
Life Science5 credits
fiaUt 205Calculus I for
PhysicalSciences andMath 5 credits
Math 211Calculus I3 credits
Math 107Programmine3 credits
Math 21:Programmingwith Statistics
3 credits
Math 218Programming
3 credits
Before the actual structure and con-tent of the individual modules are dis-cussed, it will be instructive to look atsome of the story of their development.At the start of the project we wrote tonumerous publishers and universities inorder to obtain Information on texts andalready existing CAI for calculus rouses.Although we then found some interestingsupplementary texts, it soon became ap-parent that very little had been done forcalculus like the interactive programsthat we envisioned. Algebra, statistics,and linear algebra seemed to be the areasthat had attracted the most effort for avariety of reasons running from the sizeof the student populace to the appropri-ateness of the material for programming.We found the nearest thing to our aspira-tions at Georgia Institute of Technologywhere Dr. J. C. Currie and a small grouphad developed some rather brief programsfor illustrating limits, derivatives, ap-proximate areas, and other topics alongwith a driver program which we totallyfailed to appreciate at the time. (6) Far
more impressive things had been done forlogic at Ohio State University, for sta-tistics at University of Akron, and fornon-mathematical subjects in many otherplaces; we wanted to see something similardone for calculus.
As we learned more about the diffi-culties of compatibility and transferabi-lity and thought more about the problemsof translating whatever we found intoSpanish for our students, we leaned to-wards developing our own programs out ofthe many materials and ideas which we hadcollected during the first year of theproject. We had finally settled on buyinga Hewlett-Packard 2000 System for its re-putation for fast response with a largenumber of users, to give us independencefrom the main competes center with itsmany large administrative jobs, and be-cause it supported Coursewriting Facility,an approximate implementation of IBM'sCoursewriter with available BASIC numericalfunctions. We had experimented with thelatter language on UPR's IBM 370/145 com-puter and believed that it was a ratherpowerful tool providing the necessarystructure for developing good CAI. Wewere eventually to come to grief withthe Coursewriting Facility, but we remainpleased with the rest of our system;after many months of work in this lan-guage we found ourselves unwillinglyforced to convert everything to BASIC,which is now the language in which allour programs are written.
We arrive, then, at the modulesthemselves. It was for the calculuscourse of the hard sciences and methane«tics (Math. 20S in Fig. 1) that the CAImodules were produced. The various pro-grams which make up the entire CAI systemin calculus are depicted in Fig. 2.
Within the system, the student hascomplete freedom to choose among thedifferent modules at any time, althoughhe is informed in class when it would bemost appropriate to study each module.The manager program, which interfacesbetween the student and the individuallessons, also allows the students toregister themselves, avoiding one of
Manager Pram
Registrationof Students
S sa-on anSecurit
Mathematics 173
the time consuming operations often asso-ciated with the beginning of the semester.The structure of this program is shown inFig. 3. (Of course the security with thiss;.tem is minimal, but in our environmentthis is not a problem.) This same programmaintains data on the total number of usesof each module by each student and thetotal number of minutes of terminal timeof each student. Three separate reportprograms have been written to produce dif-ferent usage reports for the teachers ofeach section and for global evaluation.No data are kept on right and wrong res-ponses: it will become clear further onthat such information is either not ap-
-plicable or irrelevant for the majorityof response situations presented.
The general form which most of themodules follow is given in Fig. 4. Itmust be remarked here that these modulesare nut in any way conceived of as re-placing the textbook or the lecture. :heyare strictly supplementary and each onepresupposes that the student has alreadybeen introduced to the concept or tech-nique in his regular class work. Thus,4mthe first part of the module there are'usually a few questions to find out if thestudent is at all familiar with the mate-rial, and if not, to recommend that helearn more about it before gotng on. Theexamples which the computer presents areas much for the sake of familiarizing thestudent with how the computer output looksand how to later enter his own data as theyare for illustrating the material beingstudied. That is to say, it is expectedthat the student will continue making uphis own examples and investigating thecomputer's responses and will do most ofhis learning during this latter exchange.Those students who stop after the preparedexamples are not really likely to ;lavegained much. Fig. S presents the eightmodules and their content essentially asthe manager program presents that menu tothe students, except of course, it appearsin Spanish on the screen.
4. SECANT S. GRAPH
7. NEWTON
6. CURVE
1ElETh Me I Modules
I. INTRO 2. NUMBER]
Figure 2. The Calculus CAI Programs
Report, Facilities
Rised Students by Section I
Amber of Successful Completions!of tea tbdula by Student
I Total. Usage Time for each tbduS1by Student
174 NECC 1980
/Present
Modulesof
1-----SelectModule
I
Figure 3. Manager Program Flowchart
sail format as in 3 . The computer provides
the necessary data and prompting. The optionto rerminate is offered before each example.
1. INTRODUCTION
2. 9pESTIONS. Does the student know enough to benefitfrom the module? If not, he may be askedto return to the text before continuing.
3. INTERACTIVE EXAMPLES. The student participates inworking through an cx..mple selected bythe program.
4. EXPERIMENTATION. The student is given the opportunityto make up and enter his own examples in the
ob.a.mee dm. el re.wwer...mrFi3ore 4. CFneal Scheme for Modeles
Mathematics 175
1. INTRO Introduces the terminal, thekeyboard, and its traits.
2. 111142814 Presents real numbers, round-off,and how to read mathematical tables.
3. LIMIT : Studies Limits of rational functionswith values that make the denominatorequal to zero (0) by the use of a table.
4. MAW: Illustrates how the slope of the secanttends toward the slope of the tangent.
5. GRAF : Calculates values of a given functionand its derivrtivesinorderto help plotthe graph.
6. CURVE : Evaluates the function,finds the inter-cepts, critical points, and points ofinflection of a polinoeisl.
7. NEWTON: Finds tne roots of a given polinomialusing Newton's method.
8. AREA : Calculates the approximate integral byvarious methods.
?IMRE 5. The Eight Calculus CAI Modules
Table 1. Calculus *Wale Usage
,1Pall
78-79
tamdwar ofSections
TotalStudents
Number ofStudents
participating
X ofStudents
participating
ModulesUsed perStudent
AverageTotalMinutes
5 150 37 25 % 3.3 65
Spring78-79 9 270 123 441 5.6 84
Pall
79-807 200 67 34 2 3.1 68
2
Sable 2. Average of Responses to Module Questionnaire', 1978-79
relevant relevantto to
long clear interesting helpful course studieshigh recommendlevel fast it
6 3.6
AA
3.5
tart confusing boring not irrel- irrele- loso slow don'thelpful elevent vane to level recommend
to course studies it
is 6
178 NECC 1980
The first two modules in fact have nothingto do with calculus. They are there tominimize the shock of dealing with thecomputer for the first time and increasethe probability that the student will beable to successfully use the modules andinterpret the results. Next comes a mo-dule on limits and then on the derivative,both centered around investigating tablesof values of the appropriate expressionsin the neighborhood of a chosen point.There are no epsilons or deltas here, andit is hyped that the student will get moreof a feel for how varying values approacha fixed number. The next two modules dealwith the use of the first and second de-rivatives to describe certain propertiesof the graphs of functions and to drawthose graphs. It is our continuing dis-appointment that these modules containvery few graphs. Despite the fact thatwe had proposed to buy some graphics ter-minals and even a plotter, we were neverable to do so. Several programs weredeveloped to produce graphs with alphabeticcharacters but were never actually incor-porated into the fiaal modules. It isstill not clear whether a graphing routinecannot make a straight line look like astraight line that will be veryconvincing for the students. The seventhmodule illustrates the Newton- RaphsonMethod for finding approximate roots topolynomials, while the last module doesapproximate integration by three of thestandard methods frequently treated inintroductory calculus courses.
To demonstrate the implementation ofthe general structure of the modules asshown in Fig. 4, we shall disease in moredetail the seventh module: Newton. Theintroduction reminds the student of theobjective and iterative nature of Newton'sMethod ant' requires him to recall the pre-cise form of the iterative formula, usingone of the very few multiple choice formatsin these modules. Once he gets thatstraight, the sample function, f(x) e
4 x3 -5, is presented. The studeat is di-rected to choose a starting value, x , andthe computer then calculates, one 0 ata time on request, x, until thestudent decides *.hat
,
hAt 4 he is ready tosay what is vot ,eing approximated.Upon entering his Guess, he is either toldit is not good enough and to look at morevalues, or that it is right, or that it isclose enough, but that a better valuewould be ,Once the .student has success-fully named the root, the computer offershim the ,..pportunity to enter a polynomialof his choice by entering, successively.the degree and the coefficients.The program then runs through the same
L
sequence of steps until the student namesthe root being approximated with suffi-cient accuracy or until it becomes clearthat the process is not converging. (Asyou may expect, the only method the pro-gram has for determining the root is thesenses Newton's (Method, so that in someuntidy cases the coMpeter gets it wrong,leading perhaps to some confusion. Thissort of thing is, however, a constantsource of difficulty in many of the mod-ules and in fact in all numerical methods.Some care has been taken in the program-ming to avoid some of these pitfalls andto give some warnings in the text portions.The student manual, which offers variousforms of encouragement and hints, alsopoints out same of the difficulties.) Thestudent then has the option of trying for.another root of the same polynomial, en-tering another polynomial, or terminatingthe module.
RESULTS OF THE CAI IN CALCULUSThroughout the duratioa of the proj-
ect, and continuing.now, certain evalu-ative procedures were used to assure somekind of objective information about par-ticipation, reaction, and effect of theCAI. The first of these, participation,was constantly monitored by the computeritself. Table 1 shows the percentage ofthe enrolled students who actually usedthe modules to some degree or other, theaverage number of modules run by eachstudent, and the average time spent byeach student at the terminal during thesemester. The relatively low use is dueirincipally to three factors. Foremostis the totally voluntary nature of thecomputer laboratory. Any such voluntaryactivity cannot hope to attract more than75% of a class unless it guarantees bettergrades. Our modules offer no such guar-antee. Secondly, there were considerabledifficulties during the first semester of1978-79 due to the delay in completingthe final home for the computer center and theresulting lack of terminals and comforta-ble working conditions. At that timethere also was no student manual, leavingthe students more or less on their own inattacking the modules. Finally,the pro-fessors teaching the various sections ofthe calculus course change from semesterto semester, and there is a clear corre-lation between the enthusiasm of the pro-fessor for the computer laboratory and thestudents' participation.
Student reaction was checked eachsemester by a standard questionnaire dis-tributed in class and in the computercenter. It was originally supposed thata student would fill out one of these
1
questionnaires for each module that heused, but we soon decideC that we shouldbe content if we could elicit one overallreaction from each participating student.The results of these questionnaires areshown in Table 2. The reaction is clearlyfavorable, being most favorable on theissues of "interesting" and "would recom-mend it." (Responses here were made on ascale from 1 to 5, 1 in full agreementwith the bottom description and 5 withthat of the top.) The form used also in-cluded more direct questions over problemsencountered in using the modules, and someof the students' comments led to correc-tions and minor improvements.
The actual effect of the CAI on stu-dent performance and general understandingis obviously the most difficult to measureof the three parameters, but perhaps themost important. We have made two formalattempts at measuring this. The first ofthese was in the second semester of 1977-78,when we set up non-participating controlsections for comparison of results. Dueto a plague of difficulties, many relatingto the problems which led to giving up theCoursewriting Facility, it became impos-sible to gather any reasonable data. Thesecond attempt was made at the end of thefirst semester of 1978-79. At that tira study was made comparing certain in-dices of calculus students who had actuallylama some of modules. These indices weretheir SAT scores, both achievement andaptitude, their final grade in calculus,and the number of modules used. In thissmall sample, no actual statistical cor-relation could be established, but cer-tain tendencies were apparent. Porexample(see Table 3), all the students who usedfive or more modules obtained a grade ofC or better in the course, while of thestudents who used five or more modules,only one had obtained better than 700 onthe SAT math achievement.
Table 3. Comparison of Module UsageCourse Grade, and Previous Aptitude
and Achievement Test ResultsPail Semester 1978-79
(Fig
part
0Ae
c
o u
Aetitebcores
ores in percent ofLcipating students.)
27 4 4
11 08 8
11 15 11
4 27 15 11 4
8 4 8 4 11 4
4 4 11---0...8----44 8 4 111 4
11 4 111 1
de 0400 601-700 701-400 1-2 3-4 5-6 7-8.
1Imammelmvir.
0
0ot-acio01-700t
0-600 X
44"
I
Number
doulfes
Mathematics 177
(One fourth of the participating studentsran five or more modules). One is temptedto draw the conclusion that the studentswho used the modules most are the hard-working students; not necessarily highlyintelligent, but nonetheless successful.
To conclude this discussion of theevaluation of the project's results, wewould like again to recall that the ob-jectives of these CAI modules and thebasic philosophy used in constructingthem are to overcome certain experimentaland analytical deficiencies in the stu-dents' preparation (as well as to offer achallenge to the more highly motivatedstudents), and hence they are not directedat the routine type of problem solvingthat is prevalent on examinations. Itwas, therefore, never expected that usingthe modules would be directly and imme-diately reflected in better test scores,but rather in a deeper, long-range appre-ciation of fundamental concepts and amore enquiring approach to attacking newmaterial. Any correlation obtained be-tween exam performance in the calculuscourse and module use is therefore muchmore likely due to pre-existing attitudesof the student than to the contributionthe calculus CAI made to his or her knowl-edge. The calculus is neither a bag oftechnical tricks, as it- is frequentlytaught to non-science majors, nor an ex-ercise to mathematical logic, as it oftenappears in honors courses. It is a co-herent system of ideas elaborated for themodeling and analysis of certain clavier:sofphenomena; experience with some of thesephenomena and with the way calculus pur-ports to capture these processes is neces-sary for anyone aspiring to scientificinvestigation or teaching. We hope thatour modules make a start in the directionof providing some of, this experience.
DIFFICULTIES AND WARNINGSThe creation and implementation of
successful computer-assisted Instructionis a major undertaking involving manyfactors, any one of which is capable ofnullifying many months of serious effort.Our project, which we consider to havebeen reasonably successful overall, didnot escape from its share of mistakes andfrustrations, and it still faces the dif-ficulties of maintenance and imbrovement.The initial challenge is the selection ofan adequate system of hardware and soft-ware, unless one is locked into a pre-existing system. The choices today arefar wider than they were three years ago,and serious consideration must be given tomicrocomputer systems as well as to minisand main frames. Although our mini
1 S
178 NECC 1980
computer, the Hewlett-Packard 2000 System,has functioned well in our environment,its upkeep may be too great for a smallerinstitution, or capacity too limited fora more ambitious project. The lack ofspecial symbols on our terminals (H-P2640B), such as integration signs or evenlower-case letters, puts undue strains onprogrammers and users alike. The mathe-matical symbols as they appear in textsand are used by the teacher on the boardare important for a quick understandingof the material and for adequate rein-forcement. For example, INT X2 DX needsmuch more attention thandoes the usual expression,jx2dx. Relatedto this point is the ability of the ter-minal to recall previous pages of material(terminal memory), or careful programmingto imitate this ability. Although all ourterminals now have memory, at the outsetof the project they did not, and we foundthis situation much less flexible and moredemanding.
Problems with software can be moreinsidious. It is not easy to find outfrom salesmen exactly what a computer ora software package is capable of doing,particularly in the area of educationalapplications. We began, as mentionedabove, with iBM's Coursewriter which isan especially designed system for CAIand its management. We soon discovered,however, that it is incapable of doingany kind of calculations other than in-teger arithmetic. This can be overcomeby begging, buying, or writing assembler-
- language functions to increase the powerof Coursewriter, but this solution wefound to be painful and restrictive. H-P,on the other hand, allows these functionsto be added to their Coursewriting Fa-cility via simple BASIC programs. To alarge extent this is what sold us on thesystem we bought. Several months laterwe gave up trying with the CoursewritingFacility and converted to BASIC. Com-plications in the compiling procedure,bugs in the management facility, andsystem crashes occasioned, at best guess,by the many needed disk accesses finallywore down the patience of even the moststalwart among us. Of course, BASiC isan entirely adequate language for mostCA/ ( consider its transfer.-ability, by far the majority of theavailable cAi packages are in BASIC),but it does not encourage the same kindof careful organization and answer pro-cessing as do Coursewriter and other CAIlanguages.
Some of the other components thatcan, and hence will, go wrong are: Inade-quate physical plant, too few terminals,
improper scheduling of student usage, poorresponse time when the computer is busy,and,inevitably, equipment failure. Whileit is idle to suppose that one can avoidall of the problems and delays that typ-ically arise, careful consideration canavoid many of the difficulties. Afterall, you can not change your computer sys-tem with the same frequency and ease withwhich you have been changing vourcalculustext over the years! We hope that as theadvice and experience of a large numberof generous and helpful people helpedguide us through the vagaries of thisproject, so some of our experience maycontribute to future endeavors of thisnature:-
REFERENCES1- M.J. Christensen, MATHDOC; A Control
System fox Computer-Assisted Calculus,Proceeding of Conference on Computersin Undergraduate Curricula, Denver,June 1978.
2- W.S. Dorn, G.G. Bitter, D.L. Hector,Computer Applications for Calculus,Boston; Prindle, Weber and Schmidt,Inc., 1972.
3- G. McCarty, Calculator Calculus, PaloAltos Page-Ficklin Publishing Co.,1975.
4- R.G. Selsby and M. G6mez Rodriquez,"On- the -Job Training of Students inComputer Science," Proceedings ofMinority institution Curriculum Ex-change Conference, Concord, NorthCarolina, January 1979.
6- K.D. Stroyan, Manual for a ComputationLaboratory in Infinitesimal Calculusand Linear Algebra, MathematicalSciences, University of Iowa, 1976.
COMPUTER APPLICATIONS IN A FINITE MATHEMATICS COURSE
by
R. Abernethy,G. Piegari,
andA.L. Thorsen
Mathematics DepartmentVirginia Military Institute
703-463-6335
INTRODUCTIONThe course described in this paper sup-
plements a standard course in finite math-ematics with computer applications. Itwas developed with the following goals inmind. First, since the course is offeredto students in engineering or mathematicscure^ula, it was the authors' hope thatthe computer applications would increasestudent interest and motivation as well ashead off the common complaints about theirrelevancy of abstract mathematicscourses. Second, it was felt that thecomputer assignments would lead to an in-crease in the student's proficiency inprogramming, which would prove benefi-cial in his later mathematics and engi-neering courses. Finally, it was hopedthat the student's grasp of certain top-ics, in particular probability, would beenhanced by the computer assignments.
Three sections of the course weretaught to a total of 51 students in thespring semester of their freshman year.Each of the students had received, in thefall semester, at least 4 hours (contacthours, not credit or semester hours) ofinstruction in BASIC. However, most werefar from being competent programmers, andthe authors are convinced that our coursecould be made self-contained by beginningit with 4 or 5 lectures on BASIC. Westarted our course with a diagnostic teston BASIC and followed it with a one-dayreview of BASIC. We resolved to discussprogramming only as specific problems a-rose in connection with computer assign-ments and spent no more than an averageof 15 minutes class time per week onprogramming. Eight computer assignments,timed to correspond to the material beingcovered in class, were given. Each as-signment was to be completed by its as-signed deadline, which was between oneand two weeks. These assignments, in to-tal, were counted as an hour test (15% of
Mathematics 179
the final grade in the course). No comput-er work was included on hour tests or thefinal examination.
The computer assignments were executedon an HP2000 computer through 8 time-shar-ing terminals in three locations on campus.However, a group this size could easilyhave been accommodated with as few as threeterminals (perhaps fewer with proper sched-uling). To provide help on the computerassignments, each of the three instructorswas available for a one-hour scheduledhelp-session each week. Each session wasopen to all students in the three sectionsand was conducted in a computer terminalroom where five terminals were available.
One restriction on the course, placedby our department, was that none of thecore mathematics content of the course bedeleted or diluted in order to accommodatethe computer work. We believe that we sat-isfied this restriction, even though cer-tain introductory topics from chapters 1and 2 of the course text (1) were omittedfrom the syllabus to provide the additionaltime required for the computer assignments.The time gained proved more than adequate,and in addition to the computer work wewere able to give a reasonable treatment oflinear programming, which was not in theoriginal course syllabus. The major topicsin the course were set theory, discreteprobability (including Markov Chains andstochastic processes), matrix algebra, sys-tems of linear equations (including aLeontief model in economics), and linearprogramming. The text material was chapters3,4,5, and 6 (1), but there are numeroustextbooks which offer similar coverage.
The course was evaluated in two ways.A student evaluation form was completed byeach student at the end of the course,which provided some insight into whetherthe goal of increased interest and motiva-tion had been achieved. More formally, wecompared the performance of the students in
180 NECC 1980
the computer-supplemented course to that ofstudents in the standard course. A lamepart of the final examination which wasgiven in our course had been given to sev-eral earlier sections and one concurrentsection of the unsupplemented finite math-ematics course. To test our hope thatcomprehension of some of the mathematicaltopics would be enhanced by the computerwork, we compared the scores of the stu-dents in the standard course with thescores of the students in the computer-supplemented course for the relevant sec-tions of the exam. Results of these evalu-ations are discussed in section four ofthis paper, following a discussion of someof the computer assignments in sections twoand three.
COMPUTER ASSIGNMENTS IN PROBABILITYThe first five of the eight computer as-
signments were related primarily to proba-bility theory. Pour of these five weresimulations of probability experiments.One assignment was a program to computecombinations and permutations and usethese results, in turn, to compute the prob-abilities of certain events.
The first program was simulation of onehundred tosses, one thousand tosses, andthen ten thousand tosses of a fair coin.It was in this assignment that the stu-dents were introduced to the random numbergenerator and its vital importance in sim-ulating experiments. Also, we indicatedthe need for a test condition to translatethe value of the random number generatedinto an outcome for the experiment. Forexample, since the random number generatedis greater than or equal to zero and lessthan one, declaring the outcome as headsif the number is less than 0.5 is a simpleand reasonable way to simulate the tossingof a fair coin.
The second assignment used the randomnumber generator introduced in assignmentone for a more complicated simulation.The experiment was rolling a pair of un-biased dice. The student was reminded thatthe outcome of one die is independent ofthe outcome of the other, and the outcomes2 through 12 must be generated in conformi-ty with their natural frequencies. The useof the integer function INT was introduced.Thc, outcome of one die can be simulated bygenerating a random number, multiplying itby six and adding one, and then taking thegreatest integer (INT) of this result.The outcome of each die is generated inde-pendently, and the two numbers are addedtogether to obtain the outcome of the ex-periment. Assignment two consisted ofrunning 100, 1000, and then 10,000 trialsof the experiment and approximating the
correct probabilities of the various out-comes. Varying the number of trials wasintended to convince the student that thelarger the number of trials, the closerthe approximating probabilities reflectthe true probabilities obtained by analyt-ical methods in the classroom.
The third computer program was a simu-lation of the classical Buffon's NeedleProblem. It is both easy enough to simu-late on the computer and difficult enoughto defy solution by the analytical methodsstudied in finite mathematics (although itcan be solved by integral calculus). Thestudent is thus introduced to the mainvalue of simulation--solving problems thatcannot be solved by other means. InBuffon's problem, a needle one unit inlength is dropped on an infinite grid ofvertical lines. The probability that theneedle touches a line is the desired quan-tity. For our experiment the lines weretwo units apart. In order to accuratelysimulate the experiment two independentrandom numbers must be generated, one toobtain the angle the needle makes with ahorizontal line and one to obtain the lo-cation of the center of the needle betweentwo lines. The test condition to determinea hit or miss amounted to comparing theprojection of the needle on the horizontalline to the distances of the center of theneedle from the adjacent vertical lines(5). The student was required to computethe approximate probability that the nee-dle hits a line for 100, then 10,000 simu-lations of the experiment. Incidentally,since the exact probability is 1/11, weasked the student to compute the recipro-cal of his approximate probability to seeif he would recognize this answer as anapproximation of t.
Assignment four involved programmingalgorithms to compute C(n,r) (the numberof combinations of n objects taken r at atime) and P(n,r) (the number of permuta-tions of n objects taken r at a time) andusing these algorithms to compute proba-bilities, for example the probability ofgetting four aces in a seven-card pokerhand. One by-product of this assignmentis the lesson that computers have practi-cal limitations. The largest factorialthat could be computed on our computer was331. Therefore, the student was forced todevelop a more sophisticated algorithm tocompute C(48,3)/C(52,7) (the probabilityof four aces in a seven-card hand) thansimply computing each of the factorialsfirst and then performing the division.
The fifth assignment had three parts:(1) the simulation of 3000 World Seriesbetween two evenly matched baseball teams;(2) the simulation of 3000 World Series
in which one team had a .6 probability ofwinning any single game; and (3) the simu-lation of 3000 World Series if it werechanged to a best of nine game series,rather than r best of seven. In the firstpart, where the teams are evenly matched,the length of each series was tabulated, toillustrate the non-intuitive fact that asix-game world series is as likely as aseven-game series. Since the probabilitiesof having a six-game series and a seven-games series are equal, about half theclass can be expected to conclude that asave: -game series is most probable and halfthat a six-game series is most prothle.It should be noted that a slight deviationaway from exactly evenly matched teamsmakes a six-game series the most likely.The second part of the fifth assignment wasintended to illustrate the fact that if oneof the teams has an advantage over t...) otherteam in each of the games, then itz advan-tage ie magnified when the trials aregrouped into a contest. The probabilitythat a team with a .6 probability of win-ning each game wins the series is about .7.The third part of the assignment illustrat-ed that as the number of trials grouped ina contest are increased, the initial advan-tage is further magnified, in this canefrom about .7 to abut .73. Although eachof these parts can be solved analytically,the solutions are complicated and usuallybeyond the objectives of a finite mathe-matics course; so, the student is againimpressed with the primary value of simu-lation. (Similar problems are ..iscussed inKemeny (2), pages 143-144 and section 1.5and Kemeny (3), pages 161-167.)
COMPUTER ASSIGNMENTS USING MATRICESAssignments six through eight were de-
signed to introduce the students to the ma-trix (MAT) commands in BASIC and to the sys-tem library. Assignment six was a set a!routine exercises that required the use ofthe commands for the matrix operations ofaddition, multiplication, scalar multipicce-tion, multiplicative inversion, and trans-position. Additionally, the students weregiven a 2x2 matrix B and asked to use a loopto determine B4.
Assignment seven paralleled our classdiscussion of Markov Chains, and its purposewas to introduce the students to subroutines,round-off error, and a double loop. For theproject, the students were given two 3x3transition matrices and asked to write a pro-gram to test each matrix for regularity. Atransition matrix A is regular if An has allpositive entries for some nonnegative integern. The students were asked to recall that inperforming calculations with a computer,round-off error is almost always present.
Mathematics 181
Hence a quantity, which may actually beequal to zero, may be calculated by thecomputer to be, for example 0.0000021.Thus, in the subroutine to test for zeroentries in BRAn, the students were forcedto structure their conditional statementsin such a way as to ignore any differencein compaAd quantities which are less thanthe suspected round-off error introducedby the computer. In addition to a loop tocalculate An, assignment seven also re-quired a double loop (one for the row in-dex and one for the column index) to deter-nine if An had any zero entries. The tran-sition matrices in the assignment were 3x3matrices. Thus the students knew thgttheir program need only test forzeroes because in class we gave the theoremstating that if an nxn transition matrixis regular, then at least one of the first(n-1)4+1 powers of A contains no zeroentries.
For the first part of the final computerassignment the students were required towrite a program to solve two problems basedOn our classroom discussion of Leontiefmodels. This prugram involved solving asystem of linear equations using the INVfunction to find the inverse of the coef-ficient matrix.
It has been noted that one of our ob-jectives was to increase the student'sproficiency in programming. Toward thisend, each of the above computer assignmentsrequired the students to actually write aprogram- -not merely supply data to a pro-grew called from the system library. How-eVer, the authors felt that students shouldbe given an opportunity to realize the im-portance of the system library; hence, theyintroduced the system library in the secondmitt of assignment eight. The problemsupplemented our class work on linearprogramming and the simplex method. In theassignment the students were asked to max-imize an objective function subject to fiveconstraints using a program (LINPRO fromour system library. The instructors feltthat the introduction to this powerful fac-et of computer use was an appropriate endto ;.he computer segment of the course.
EVALUATION OP THE COURSEThe results of the student evaluation
given at the end of the course were general-ly quite positive. Most students sewed notonly to enjoy the course, but to feel thatit was useful. For example, a question onthe overall rating of the course had anaverage of 8.3 on a 0-to-10 scale. Theaverage response to the question; "Wouldyou recommend this course to others?"(again on a 0-te-10 s ^ale , where 0 = notrecommend at all, and 10 = highly recommend)
182 NECC 1980
was 8.1. in addition, many responses inthe comment section of the evaluation formindicated that the course was well-received.
A statistical evaluation of the coursewas done by comparing scores of students inthe regular course (group R) on 21 problemscommon to both final exams. We proposedaverage scores on the mathematics portionof the SAT as a possible measure of howgroups C and R compared in ability. Imme-diately a problem arose because the twogroups were significantly different whenthis variable was measured. Two comparablegroups were acquired by deleting all stu-dents with Math SAT scores over 600 in eachgroup of students. We were left with 24students to group C with an average MathSAT score of 531, and 29 students in groupR with an average Math SAT score of 533.Students in both groups C and R were engi-neering and mathematics majors.
The results of the 21 examination ques-tions are given below.
Grou C GrouN er o st ents 29
Average Math SAT 531 533
Total number ofquestions
504 609
Number of correctresponses
330 365
Sample successprobability
PC r .655 .599
From the table above we get the differenceof the sample proportions to be Pc-Ae.056.
Let PC be the success probability for the
grow, C population and PR the success prob-
ability for the group R population. Test-ing the null hypothesis Pc 4 PR against the
alternate hypothesis Pc > PR, we found that
the probability that fc fR > .056 is less
than .028 (c.f. chap. 8,(4)). Therefore,at the 5% level of significance, we rejectedthe zAll hypothesis and accepted the alter.-nate hypothesis that the performance of.stu-dents in the computer-supplemented coursewould exceed that of students in the regularcourse on the 21 exam questions. We pointout that the one-sided test was suggested bythe restriction that the coverage of thecentral mathematics topics in the new coursebe at least equal to the ooverage given inthe regular course. Consequently, it wasexpected that the value of Pc would be atleast that of PR .
1Aere may be other faotors which con-tributed to the superior performance ofgroup C. For example, those in group C
were exposed to each of the three instruc-tors, while those in group R were taughtby only one of these instructors. in ad-dition, the class size for group C averaged17 students, while the average class sizeof group R was 24 students. However, theauthors believe that the introduction ofthe computer into the course was an impor-tant positive factor contributing to thesignificant difference in performance notedabove.
Finally, we remark that when the stu-dents with Math SAT scores of over 600were compared, the students from group Cagain performed better on the 21 examquestions than the students from group R.However, the difference was not as gag-nificant. This result is consistent withthe authors' feeling in the beginning ofthe project that the better students werelikely to perform well no matter how thecourse was designed and that the greatesteffect of the computer supplements wouldbe on the average students.
SUMMARY AND RECOMMEHDATiONSit was our intention to add a computer
supplement to a course in finite mathe-matics without compromising the traditionalapproach to the mathematical content in thecourse. In short, we did not want our stu-dents to become computer experts whilesacrificing their ability to solve problemswith paper and pencil. it is obvious, how-ever, that one cannot introduce a computersupplement into a finite mathematics coursewithout making some adjustments. Since wewere restricted to the same 45 contact hoursas In the standard course, we had to elim-inate certain introductory topics mentionedat the beginning of this paper; however, wedid not reduce the amount of time devotedto probability and matrices. We eliminatedenough material to give us an additionaleight contact hours, and we found this ad-ditional time sufficient for the discussionof computing in general and the computerassignments in particular. If, on theother hand, one wished to add a computersupplement without sacrificing any of thestandard course material, the best approachwould probably be to increase the contacthours per week from 3 to 4. The extra hourper week mould be more than adequate toaccommodate the computer supplement.
We also believe it is important for thestudents and the teacher to maintain asense of perspective about the goals ofthis kind of course. it is too easy todrift unintentionally into a computer pro-gramming seminar to the detriment of thetraditional theoretical and problem-solvingaspects of a finite mathematics course.The student must be kept mindful of the
193
main thrust of the course, even if, as isoften the case, he is more interested incomputers than in mathematics. Keeping thecourse properly balanced becomes a questionof the instructor's resolve, but he canhelp himself (as we did) by keeping thecomputer part of the course out of the hourtests and final examination.
Finally, we believe that when computerassignments are returned to the student,the student's program should be accompaniedby the instructor's solution. The solutioncan be typed in capitals to resemble theactual print-out obtained at a teletype-writer terminal. Besides showing the stu-dent the correct answer, an instructor'ssolution provides the student with a valu-able reference if he pursues computer pro -gramming after completion of the presentcourse. Such a policy, of course, requiresnew or altered computer assignments forsubsequent courses, but we consider theprice small in relation to the benefits.
We must admit, in summary, our satis-faction with the results of this course.It will likely be a permanent part of ourcurriculum, and we anticipate further re-finements and additions to the computerassignments in future offerings of thecourse. To this end, we invite others whohave had similar experiences with a com-puter-supplemented finite mathematicscourse to share their views with us.
3. Kemeny, J.G. and Snell, J.L.,Finite Markin, Chains, Princeton:Van Nostris,Ur
4. Mendenhall, W., Introduction toProbability and Stat fisFair,3raedition, Belmont,Ca1-7MrtiasWadsworth, 1971.
5. Mosteller, F., Fifty Challen inProblems in Probability, Read n ,FIREWEVarais Addison- Wesley,1965.
1 94
.Mathematics 183
184 NECC 1980
A COMPUTER-ASSISTED COURSEIN BIOMATHEMATICS
Pui-Kei WongMathematics Department
Michigan State universityEast Lansing, Michigan 46824
517-353-6880
Applications of mathematics to prob-lems in the physical sciences and engin-eering are well known and have long beenan integral part of the undergraduate cur-riculum of mathematics departments. touringthe past two decades courses in numericalmethods and elementary finite mathematicshave also been added by many colleges anduniversities, reflecting the growing im-portance of mathematics in economics,management. and social sciences as wellas in modern technology. We need onlycite the contributions of such individualsas Kenneth Arrow,' George Dantzig, NasallyLeaatiff. Paul Samuelson. and Johnvon Neumann in this connection.
On the other hand, applications ofmathematics to the biological and lifesciences, though no less important, havenot found their proper pla'ce in the under-graduate curriculum. It was J.B.S. Haldanewho observed in 1928 thatt "The permea-tion of biology by mathematics is onlybeginning, it will continue and (grow)into a new branch of applied mathematics."In the decades since, we have witnessed aphenomenal growth in the applications ofmathematics to the agricultural, biolo-gical, and life sciences. beginning withthe work of Haldane, Lotka, Volterra.and Wright and onto that of Hodgkin andHuxley, who shared the Nobel prize inphysiology or medicine in 1963. And mostrecently, Allan Cormack and GodfreyHounefield were awarded the 1979 Nobelprize in physiology or medicine for theirpioneering work in applying mathematicsand computer technology in the x-ray imagereconstruction technique called Computer-ized Axial Tomography (1).
Recognizing the growing impact ofmathematics in the life sciences, theMathematics Department at Michigan StateUniversity introduced five years ago anew course MTH-461 titled "SelectedMathematical Ideas in Biology." This is
a one-quarter, four-credit course withcalculus as prerequisite. At MichiganState we have a two-term, ten quarter-credit sequence in calculus for the bio-logical and social sciences. A studenthaving completed this abbreviated calculussequence is usually adequately preparedto enter MTH-481. No prior knowledge ofcomputer programming is aesumed.
The major objective of this course isto provide the student with a sound intro-duction to mathematical methods and deter-ministic models in biology, especially inthe use of computer simulations to analyzemodel behavior. To this end a number ofproblems from biology are introduced andstudied in some depth. We start with theunderlying biological description, hypo-thesize, and then derive a working mathe-matical model. Once constructed, the modelcan be manipulated and analyzed to revealits possible behavior. Modifications onthe basic model are made, and the systemis then studied again under various initialconditions, external inputs, and pertur-bations. All this can be done quickly andeconomically by computer simulations.Instead of numerical solutions. it is oftenmore desirable that the long-term trend orqualitative behavior and stability proper-ties of the model be known. In this casecomputer graphics provide a particularlyeffective tool in the analysis of modeldynamics.
In adlition to conventional class --room lectures and homework, students arerequired to spend time each week in thecomputer laboratory doing modeling andsimulation exercises. These exercises useinstructional modules written specificallyfor the course, and they run on Tektronix4051 graphics computing systems. Thesestand-alone microcomputers are capable ofhigh resolution graphics (780 lines by1324 pixels per line) and are especiallywell suited to our needs. A Tektronix
195
-4631 hardcopy unit provides a record of thegraphic output of any desired simulationrun.
We also use computer graphics in theclassroom. Demonstrations are done usinga Tektronix 4025 graphics terminal and anAdvent 1000A television projection system.The 4025 is a raster scan device with 480x 640 resolution and provides a compositevideo output to the Advent. It is drivenby a Tektronix 4051 computer. The Adventhas a seven-foot diagonal screen and issuitable for viewing by a class of up toforty students. Dynamics of a model canbe displayed on the screen before a classand discussed very effectively this way.A Tektronix 4662 digital plotter is alsoused to make high quality multicoloredtransparencies for use on ordinary over-hied projectors.
The mathematicg treated in MT8 -481include matrix algebra, difference, anddifferential equations, and applicationsare drawn from various areas of biology.Special emphasis is given to the qualita-tive behavior and stability of nonlinearequations. The course will typically coverthe first twenty-two items of the outlinebelow plus occasional substitution fromthe remaining ones.
MTH 481 Syllabus1. Algebra of vectors and matrices.2. Linear equations and determinants.3. Eigenvectors and eigenvalues.4. Methods of least squares.S. Discrete time single species population
models.6. Linear difference equations.7. Depletion of nonrenewable resources.8. Discrete time predator-prey models.9. Populations with age structure (Leslie
model).10. Harvesting and exploitation of renew-
able resources.11. Population genetics and difference
equations.12. Asymptotic behavior and stability
(period three implies chaos).13. Analysis of growth data.14. Logistic, Gompertz, and other continu-
ous time 9opulation models.15. Geometrical analysis of differential
equations.16. Models of photosynthesis.17. Linear differential equations and
systems.18. Compartmental analysis.19. Lotka - Volterra and other models of
competition and interaction.20. Volterra-Gauss principle.21. Theory of chemostat.22. Stability and limit cycles.23. Enzyme kinetics.
Mathematics 185
24. Simple epidemic models.25. Cell cycle analysis.26. Models in neurophysiology.27. Fishery dynamics and marine food chains.
The interactive computer-assistedinstructional modules used in the courseare designed for the student with noprior programming experience, and they arestored on data tapes. An orientationlecture on the use of the Tektronix graph-ics computing system is given at the startof each term, explaining how the studentcan access the system and use theseinstructional tapes. These modules rangefrom tutorials on vectors and matricesthrough population growth, fitting data tologistic growth curve, differential equa-tions, and modeling using compartmentalanalysis and the method of peeling.Students have found using interactive com-puter graphics particulty helpful andhave later adapted and modified some ofthe programs written for this course fartheir own research use.
To illustrate the use of computer sim-ulation in teaching mathematic modeling,let us consider the problem of growth ofhuman populations. First the problem isanalyzed by the very simple differenceequation
(1)xk+1 Ark k 0, 1, 2,
Here ick is the population size or density
at the kth census and A is the intrinsicgrowth rate, which is defined as the dif-ference between the crude birth rate Band crude death rate D. In this case thepopulation is treated as a homogeneouscollection of individuals so age structureand sexual differences are ignored. Equa-tion 1 admits the closed form solution
(2) xk = Akx0 k =
where x0 is the initial population if A
is constant. The future behavior of thepopulation is therefore completely deter-mined by the magnitude of A. Howeverreal populations can rarely be describedrealistically by such a simple model inwhich the growth rate is constant overlong periods of time.
The first modification we make thenis to allow the intrinsic growth rate Ato be time- or population-dependent.Closed form solutions to Equation 1 are ingeneral no longer possible, but the prob-lem is easily handled by the computer. Inthe Chi module for this course, actualcensus data is used to construct populationprojections. Figure 1 shows several dif-ferent functions used to fit the crudebirth rate data for Costa Rica. The
1 u
186 NECC 1980
student is asked to select one of these,and the computer will then calculate anddisplay the corresponding population trendfor a specified number of years into thefuture (see Figure 2). At the end of eachsimulation run the student has the choiceof changing the parameters or entering anentirely different birth rate function andrunning the problem again. In this casethe same death rate function is used forall the simulation runs.
At the next level of complexity, apopulation is studied by separating it in-to sex and age classes using the basicmodel of Leslie (2). For human populationswhere the ratio of males to females isessentially constant over long time periods,it is customary to consider only the fe-males and divide them into five-year ageclasses and take census every five years.The basic equation is still Equation 1,but xk is now a population vector with
n components, and A is the projectionmatrix.
A
f1
f2 f 3 fn
sl 0 0 . . 0
O s2 0 . 0
O 0 0 an_
Here fj is the average number of female
offspring born to a member of the jth classevery five years, and si is the prob-
ability that'a member of the jth classwill survive and advance the (j + 1)stclass at the end of five years. If werestrict our attention to females of agesO - 59, there will be twelve age classesand A will be a 12 x 12 matrix. if weassume constant fecundities and survivalprobabilities, the population for subse-quent years can be calculated usingEquation 2 as in the zimple scalar case.Figures 3 - 6 are the computer-generatedpopulation profiles for American femalesbased on the 1964 data for fecunditiesf1,...,f12 end survival probabilities
81"*"1112'In the final stage the effects of
population-dependent .A as well as har-vesting ere added to the'Laslie model andstudied. Figure 7 shows the menu selec-tion for this model; and Figures 8 and 9are the output of two different simulationruns for a hypothetical population havingthree age classes. Extensions of the model
to include insects, trees, and other popu-lations that are more naturally groupedinto developmental stages or size classesare also included at this stage.
All the CA/ material was written inTEK-BASIC specifically for interactiveuse and has been revised and tested overthe past five years. Although much of itwas designed with teaching mathematicalmodeling to biologists in mind, some ofit can be and has been adapted and modi-fied for use in other courses as well.The equipment was acquired in part withfunds from the National Science Foundationunder an Instructional Scientific Equip-ment Grant.
LESLIE'S MODEL: coNsar . FEC., POP. DEPENDENT WRY. PROD.FNDCW) m 1/C1.EXPCW/01-02)), DI on 180 02 m SF1 so 9 F2 m 12 $1 - 9.93333413$33 $2 m 0.5XCS) m 6 WO) so 4 ZOO - 1 WOO - 11PLOT OF TOTAL POFULATION WCK)
alm
1999
880
998
790
999
500
480
see -I-
MO
189
..,
I
\
\ t
i Iia a
19 28 30 40 SO 90 79 80
FIG. 9
204
194 NECC 1980
COMPUTER SYMBOLIC MATH
David R. StoutemyerElectrical Engineering Department
University of Hawaii at ManoaHonolulu, Hawaii 96822
(808) 948-8196
Most current calculators and mathemat-ical computer programs are oriented towardapproximate numerical computations usingan occult arithmetic called chopped non-decimal fixed-precision floating-point.If this revelation causes discomfort, con-sider what Anston Householder, dean ofnumerical analysts, said "I don't liketo fly in airplanes, knowing they are de-signed with floating-point arithmetic."
This arithmetic has undeniable advan-tages, but it also has decidedly bizarreimplications which are difficult toconvey even to graduate numerical analysisstudents. More important, it is not thearithmetic that any of us uses for manualcomputation.
THE NATURE et COMPUTER SYMBOLIC MATHThe good news is that computers also
can do exact rational arithmetic. More-over, even small personal microcomputerscan do the nonnumeric symbolic operationsof algebra, trigonometry, analytic geo-metry, and calculus. In fact, use of suchcomputers for these purposes does not evenrequire a knowledge of computer program-ming, other than the widespread custom ofusing "a" to denote multiplication and"+" to denote raising to a power, in orderto unambiguously specify expressions usinga one-dimensional format compatible withthe limitations of standard keyboards anddisplays. For example, here is a sample,_interactive dialogue using the muMATH-79 T'msymbolic math program on an inexpensiveRadio Shack ?RS-Seta computer, which isincreasingly popular in schools:
In order to expand the expression(5x- 7)(2y +3)4 the user types the line
(5*X-7)*(2*Y+3)+4;
and a few seconds later the interactiveresponse is
2 (23-42s) =x3 - a2x for x, the user typesSOLVE (2* (X+3-M2*X ) x+3x+2*X, X)and the interactive response is thesolution set
00.0,XBIA,
X = -A).
Next, to invert the matrix
l00 ql,
the user types
([1,19(0,Q]) -1;
and the response is
(11, PAN(0, 1/01).
Then, to simplify the trigonometric-logarithmic expression (tan a)cos a)+1/csc(a)+1D(e2y) 2 In a, the usertypes
TAN(A)*COS(A) + 1/CSC(A) + LN(X4.2*Y)-2*LN(X);
yielding the response
2*S1141(A) + LN(Y).
To sum the series
(a j2+ b
j)
J1the user types
SUM(A*.7+2+1344, J, 1, N)
to get the response
205-
(2*A*Nt3+3*A*N+2+A*N)/6(B1(N+1)-B)/(11-1).
(The arguments of the sum function arerespectively the summand, summation index,lower limit, and upper limit.) Finally, toevaluate the integral
jlax2 x sin x21dx ,
the user types
INT(A*Xt2 X*COS(Xt2), X)
yielding
A*Xt3/3 + SIN(142)/2.
In contrast, traditional programminglanguages such as APL, BASIC, FORTRAN orPASCAL provide built-in math facilitiesessentially only for limited-precisionarCtIonstia.
The above examples have illustratedsymbolic math capabilities relevant togrades 0 through 14. However, computersymbolic math also features exact rationalarithmetic, which is more suitable thanfloating -point for supporting math educa-tion at the kindergarten through sixthgrades. For example, continuing the abovedialogue, to simplify the expression3030/12/401 the user types
There are none of the roadoff, taufile1444or overflow problems that beset traditionalmathematical programming languages.
Although none of the above examplesrequires a knowledge of computer program-ming other than the convention of using*** for multiplication and "V* forraising to a power, curiosity of needsnot net by the built-in facilities causessomeusers to seek deeper involvementwith the mathematical and programmingtechniques used to implement these sym-bolic math systems. Accordingly, sym-bolic math systems generally provide aprogramming language for writing exten-sions. Examples of slch extensions are:
1. introducing new functions (such ashyperbolic functions) and theirsimplification properties;
2. extending the class of expressionswhich can be factored, differen-tiated, or otherwise operated upon;
3. extending the class of equationswhich can be solved;
4. automating the sequence of stepsnecessary to do an inductive proof,determine a limit, test a series
Mathematics 195
for convergence, or determinethe first few terms of a Taylorseries.
EDUCATIONAL USESComputer symbolic math has long been
available to applied mathematicians whohave generous computing allowances on thelargest computers, but this tool has onlyrecently become available on the inexpen-sive small computers most often availableto kindergarten through sophomore cal-culus classes. How can this increasing]available tool support math education?
1. A symbolic math system can be usedbx a computer-aided instructionprbgram to provide far more flex-ible, intelligent, aild responsiveautomatic drill or examinationthan is otherwise possible. Forexample, symbolic math systems areable to recognize the equivalenceof a wide variety of mathemat-ically equivalent expressions, anda computer-aided math instructionprogram can adaptively take alter-nate courses of action dependingupon the users' performance. Suchprograms free teachers to do whatthey do best--provide warmth,understanding, individual high-level guidance, and assistance forunanticipated difficulties.
2. The ability to do numerous largeexamples encourages creative ex-plorations which can reveal pat-terns and thus suggest generaltheorems. Students* inductivegeneralization powers can be exer-cised and developed to a fargreater degree than is otherwisepossible.
3. Built-in two facilities can allowstudents to see each step of acomputation, rather than merelythe final result.
4. A demonstration that an operationcan be done automatically by com-puter can encourage average andpoor students that the flashes ofinspiration given only to brilliantstudents are unnecessary for thatoperation. Hope for plodders isrevealed.
5. Students who are more enthusiasticabout computers than about mathare provided with a new avenue forappreciation of math. Computersymbolic math vastly enhancesthe opportunity for beneficialmutual reinforcement and cross-motivation between math andcomputers.
196 NECC 1980
6. Inspection of the underlying com-puter symbolic math implementationprograms can help students learnthe methods for accomplishing theoperations manually.
7. Programming extensions to thebuilt-in operations can reinforceunderstanding of both the built-inand new operations. In fact, aninstructor can withhold portionsof the algebra system implementa-tion and challenge the studentsto implement them.
AVAILABILITYFor educational use, a symbolic math
system should be interactive, general pur-pose, and available for a modest fee oncomputers typically available to students.In order of increasing computer memoryrequirements, here are four symbolicmath packages which meet theserequirements:
PICOMATH-80th is a set of three smallsymbolic math demonstration programswritten in BASIC that should run on vir-tually any computer. For information,write The Soft Warehouse at Box 11174,Honolulu, Hawaii 96828.
muMATH -79 is currently available formanY57Mvarious brands of personalmicrocomputers based on the Intel 8080,Intel 8085, and Zilog Z80 microprocessorchips. Computer memory can be measuredin units called bytes, and muMATH requires
from 32,000 to 64,000 bytes of memory,depending on how many of the various math-ecatical facilities are loaded simulta-r =sly. For information, write Micro-...ft, at 10800 N.E. 8th, Suite 819,Bellvue, Washington 98004.
FORMAC is available for medium tolarge IBM 360 and 370 computers which haveat least 140,000 bytes of memory. Forinformation, write Knut Bahr at GMD/IFV,D-6100, Darmstadt, Germany.
REDUCE is currently available for IBM360war110, DEC POP 10 and 20, Univac1100 series, CDC Cyber series, andBurroughs 6700 computers. REDUCE requiresa minimum of from 300,000 to 400,000 bytesdepending on how many of the various math-ematical facilities are loaded simulta-neously. For information, write ProfessorAnthony C. Hearn at the Computer ScienceDepartment, University of Utah, Salt LakeCity, Utah 84112.
LITERATUREMost of the relevant literature is
oriented toward research rather than edu-cation, but a good way to get started inthis area is to join the Association forComputing Machinery Special Interest Groupon Symbolic and Algebraic Manipulation.Their A6WSIGSAMBulletin is a timely sourceof information on such topics asmeetings, abstracts, and new systems.For information about joining, writethe ACM at 1133 Avenue of the Americas,New York, NY 10036.
Invited Sessions
COMPUTER-BASED RESOURCE SHARING:DIVERSITY AND OPPORTUNITY
Chaired by Donna Davis Mebane and Rodney MebaneEDUCOM
P.O. Box 364Princeton, NJ 08540
(609) 921-7575
ABSTRACTComputer-based resource sharing occurs
in many places, by many people, in manyforms, for many reasons. A professor whoasks a colleague to evaluate the pedagogicquality of a CAI tutorial, a student userof SPSS, a demographer working with censustapes, a librarian with access to OCLC --all of these people are engaged in someform of computer-based resource sharing.This special session will report on amajor EDUCOM research project, funded byNSF, to examine the diversity of sharingactivity that takes place within thehigher education and research communityand the opportunity that exists forincreased cooperation.
In describing, explaining, and evaluat-ing the sharing phenomenon, two scenariosare apparent. One is of relatively small-scale activity, very informal and withinformation exchanged primarily by word ofmouth. Many are neither aware of whatothers are doing with computers nor of thepossibilities for shared use. The otherscenario is of an active attempt byvarious resource-sharing organizations(RS08) to promote the orderly exchange ofcomputing materials and experiences.Focusing on the latter scenario, this
session will begin with the benefits ofsharing and the various pathways tosharing. It will then specificallyaddress interinstitutional resourceexchange and nine representative organ-izations that provide a link betweenresource developers and a resource usersand effect the shared use of hardware,software, machine-readable data bases,and other computing systems. The RSOsparticipating in the EDUCOM study includethe following nonprofit organizations:CONDUITZOOMGeorgia Information DisseminationCenter (GIDC)Health Information Network (HEN)
Inter-university Consortium forPolitical and Social Research (ICPSR)MeritNew England Regional Computing Program(NERComP)
North Carolina Educational ComputingService (NCECS)Research Libraries Information Network(REIN)
Profiles of these organizations will bepresented, and the nature of servicesoffered will be explored in depth.Specific RSO activities will be cited toillustrate the diversity of organiza-tional response to such questions asthese:What kinds of resources are availableand where do they originate?What options do developers have?How ate resources packaged anddistributed?Who are the current users of theseorganizations' and what services dothey find most valuable?
What are the technical, economic, andorganizational conditions of usingshared resources?How do the sharing organizations them-selves adapt to a rapidly changingenvironment?
The primary objective is to make praoti-tioners and leaders in the field moreaware of organizational options availableto meet their own diverse computing needs.Materials will be distributed describ-
ing the study and the participatingorganizations, and representatives of thenine organizations have been invited toapply for space in the exhibit area.
197 208
198 NECC 1980
COMPUTERS AND INSTRUCTION:DEVELOPMENT, DIRECTIONS, AND ALTERNATIVES
Chaired by William GrusnerAddison-Wesley Publishing Co., Inc.
Reading, MA 01867(617) 944-3700
ABSTRACTTlifrsession will examine the world of
computers in the instructional environment.Our three panelists have used micros, minis,and mainframes in various educationalsituations and will present papers as abasis for discussion and examination ofcomputers and instruction.
During the session directions and alter-natives for new uses of computers will beexplored and the existing systems examined.Participants and attendees are encouragedto exchange information and ideas. It ishoped everyone involved will leave withfesh insight into the creative use c!..omputers for educational purposes.
CREATIVE LEARNING WITH COMPUTERS - -THEORIZEOR PERISH
Margot CritchfieldPittsburgh, PA
ABSTRACT*Thperience of the dyed-in-the-wool
computer hobbyist provides educators onepoint of reference for our warm enthusiasmfor the computer. Our observation ofstudents who learn to program or who playcomplex games on interactive microcomputersprovides another point of reference. Whatdo these vastly different kinds of learnershave in common? A theory of learning isneeded to tie these disparate computerusers together. Such a theory must beapplied to the design of educationalexperiences that include computers in orderto ensure computersl.success and continuedgrowth.
This paper will attempt to clarify thedefinition of creative learning and tostate some of the explicit values of formaleducatita as they relate to computer use.The possible contribution of some currentlearning theories to a set of principlesfor creative learning with computers will bediscussed.
PLATO: COMPUTERS AND INSTRUCTION, ALARGE-SCALE SYSTEM
Robert HartUniversity of Illinois
ABSTRACTThe Language Learning,Laboratory at the
University of Illinois makes available an80 terminal PLATO site for humanities usage.During the past seven years, the LanguageLearning Laboratory has supported PLATOmaterials development for French, German,Spanish, Russian, Swedish, Swahili, Hindi,Hebrew, Chinese, Japanese, Latin, ClassicalCivilizations, and E.S.L. The relativeindependence of development projects hasled to a number of models for incorporatingPLATO in classroom activity, ranging fromtotally PLATO-centered to voluntary andsupplementary usage. Research now inprogress seems to reveal a relativelystable pattern across a wide variety ofsituations:
(1) There are two overlapping but distinctpopulations of users, one PLATO-receptive,the other non-receptive.
(2) Both instructor and student dissatis-faction with current materials centers onresponse analysis and feedbadk character-istics, which are perceived to be relativelyinaccurate, inflexible, and unable to dealwith meaning as opposed to fdrm.
MICROCOMPUTERS IN ELEMENTARY AND SECONDARYEDUCATION: WHERE WE'VE BEEN, WHERE WE'REGOING.
Dan IsaacsonUniversity of Oregon
ABSTRACT4*=Bducational research says about the
use of computers in teaching, what'shappening now, what the future holds, andwhat's holding us back will be discussed.
C'9
Tutorial
VIDEODISC TUTORIAL
Bobby R. Brown and Joan SustikWeeg Computing CenterUniversity of IowaIowa City, Iowa 52242
(319) 353-3170
ABSTRACT'mss tutorial is for individuals who havelittle or no experience using a videodiscAn instruction. The tutorial will consistof a presentation of the performance char-acteristics of videodisc as they relate to.tnstruction. A variety of applicationsc...-vnt and potential will be discussed.Various approaches to the development ofmaterials and mastering of discs forvideodisc applications will be presented.The tutorial will also include a demon -stration of an operational intelligentvideodisc system. Handouts will be pro-vided to the _participants.
199 210
Testing/Placement
MICROCOMPUTER-ASSISTED STUDY AND TESTING SYSTEM(MASTS)
Hugh GarrawayThe University of Texas at Austin
EDB 43GBAlPtin, Texas 78712
(512) 471-4014
INTRODUCTIONComputer-managed instruction (CMI) has
proven to be an effective teaching/learn-ing strategy. Studies using large com-puter-based instruction (CBI) systems suchas PLATO (Nievergelt, Jurg and others,1978) and TICCIT (Reigeluth, 1978) haveshown that in addition to facilitatinglearning, students enjoy CMI. The mili-tary is one of the largest implementersof CMI. The Navy is using several TICCITsystems in a CMI application that is ef-fective in terms of both cost and in-struction. The Air Force is using a de-dicated CMI system, the Advanced In-structional System (AIS), developed withthe McDonnell Douglas Corporation.
The Teaching Information ProcessingSystem MIPS) and the Program for Learn-ins; in Accordance with Needs (PLAN) aretwo CMI systems available for use inall levels of education. Both of thesesystems are batch-oriented and are in-tended to be used on mainframe compu-ters. TIPS and PLAN are designed to beused mainly with large numbers of stu-dents.
Successful medium- and small-scaleapplications of CMI (Bork, 1977; Brock-osier, 1977) have helped define somecommon problems in using CMI including'
1) Time and expense involved in customdesigning CMI systems for individual ap-plications.
2) Difficulty with adapting existingCMI programs to run on computer systemsother than those for which they were de-veloped.
3) Knowledge of programming necessaryfor an instructor to create or modify aprogram.
4) High initial cost for setting uphardware or adding to existing systems.
5) Communication expenses and problems,and down time associated with time-sharingsystems:
Recommendations for improving the costeffectiveness and efficiency of futureCMI systems have included the use ofstand-alone microcomputer systems and ad-vanced authoring languages that allow aninstructor to specify or choose instruct-ional and testing strategies and enterrelated content without having to learna complicated programming language.
PROTECT OBJECTIVEThe objective of this project was to de-
sign, produce, implement, and evaluate aCMI system with the following specifi-cations:
1) The system will be based on inex-
200211
AUTHOR DISK
AUTHOR PROGRAM
Creates files for assignment/test mod-ules, edits fi:es, maintains index.
.00 401 mks mk.
Collects data from completed studentdisks.
Displays results from completed nodule.
YES I
Testing/Placement 201
STUDENTDISX
STUDENT PROGRAM
Executes essignment/tests, updates"STUDEX" file.
INDEX FILE
Holds execution sequence.
STUD= TILE
Holds student data, past scores, se-quence pointer, and re-entry status.
ASSIGNMENT/TEST TILES
Holds couplets text and data for 1 sod-uls in sock file.
?JOUR! 1
MASTS INSTRUCTIONAL ALGORITHM
REPEAT THROUGH LAST REMEDIAL NODULE X, N
NEXT REME-DIAL ORSLOCX 4 SIXINSTRUCTOR
YES
GIVE NEXTMAIN ASSIG%KENT (X+1)0
.111=111M OR. Vila .00
YEE
SANE AS ABOVE THROUGH ALL MAIN AND RUMEDIAT. RDE14ENENT/TDst NODULES I
?IMRE 2
212
202 NECC 1980
pensive microcomputers with disk storageability.
2) The system will be interactive.3) The CMI program(s) will be con-
tained on individual student disks so thatthey may be used on any compatible micro-computer.'
4) An interactive, intelligent auth-oring program will allow an author tocreate CMI modules without having tomaster a programming language.
5) The system will allow an authorto gather student data from individualdisks for storage on a single datadisk.
6) The system will perform simplestatistical computations and displaygroup and individual results as rawscore, mean, standard deviation, ands-score.
7) The collected data will bestored in a format that can be easilytransferred to other statistical anal-ysis packages.
MASTSThe result of this project is the Micro-
computer-Assisted Study and Testing System(MASTS). MASTS is contained on two disks(Figure 1). The author disk contains theauthoring program which creates the dataand index files to be stored on the stu-dent disk. The student disk containsthe student program which executes ac-cording to the data stored by the author-ing program. Copies of the student diskare distributed to each student in a CMIcourse. The progress for collecting andmanipulating student data from the indi-vidual student disks are included on theauthor disk. The heart of MASTS is atwo-dimensional variable instructionalalgorithm (Figure 21 with parameters andbranching or looping conditions set bythe authoring program. This algorithmis assembled by the author program throughinteractions with the author. To assem-ble an algorithm, the author need onlybe familiar with the options availablein the algorithm model (Figure 2), as theauthoring program is menu-driven and alloptions at any point in an authoringsession are displayed on the monitor.
AIOTHORIN0.SYSTEMAn instructor wishing to use MASTS
receives an author disk, a student disk,and a short booklet outlining the optionsavailable in the instructional algorithmfor creating a CMI module. After read-ing the booklet, the instructor plans theassignments and tests and determines thecriterion levels to allow the student toprogress or receive remedial assignments.The instructor or clerical assistant may
now interact with the authoring programto create the CMI unit. A complete CMIunit may consist of numerous assignmentsand tests, but MASTS does not requirethat the complete unit be entered in asingle authoring session. The smallestamount of information that may be enter-ed during a session is one module con-taining an assignment, the test based onthe assignment, and the data such as com-petency level for advancing, number oftest questions to give, whether questionsare to be given in random or fixed order,and what remedial strategy is to be takenupon the student's failure to reach theestablished competency level.
Each assignment/test module is given atwo-number label X, Y, which determinesits position within the instructional al-gorithm. X represents the main assign-ment/test. Y represents a remedial as-signment/test for the main assignment-test X. For example, the first assign-ment in a MASTS unit would have the label1,0. I: the author-specified competencylevel were met for test 1,0, then assign-ment 2,0 would be made. If the compe-tency level were not met, the MASTS wouldexecute the author-specified remedialstrategy. The student could then repeatassignment 1,0, be given a remedial as-signment Test 1,1, or be blocked from fur-ther assignments pending a visit to theinstructor, at which time the block canbe removed from the student's disk. Whena remedial test is passed, the student isadvanced to the next main assignment.,When a remedial test is failed, the op-tions are the same as outlined for 1,0.Thus, a remedial assignment/test pre-scribed for failing 1,1, would be 1,2.Passing 1,2 would advance the student to2,0. If the last remedial test fora mainassignment is failed MN, the studentis automatically blocked and told to seethe instructor. When a student has suc-cessfully completed all of the main as-signments on a disk, he is instructed toturn in the disk to the instructor. Anynumber of main assignments (X) and anynumber of remedial assignments tY of X)may be created within the file boundariesof the microcomputer disk-operating sys-tem.
To begin a MASTS authoring session', theauthor places the authoedisk in themicrocomputer and turns on the power. Theauthor program automatically loads, runs,and presents the author with a mastermenu. The author may select one of sev-eral options. If he selects to createnew assignment/test modules, he must puta student disk into a second disk drive,and is so informed by the author program.The author will then be asked to enter
21 3
the 2C,Y number of the new ;nodule. Afterthis information is entered, the programwill prompt the author to enter the assign-ment, the text of which can be any length.At the end of the assignment, the authorenters the letter "S* on a new line to sig-nal the end of the assignment. At thispoint, the author may instruct the pro-gram to execute the test immediately af-ter the assignment or allow the student toleave the computer to complete the assign-ment. It is also possible to run a pro-grain specified by the author at thispoint, which allows the integration ofother CBI programs with MASTS.
The program will now prompt the authorto enter the test. The test is construct-ed through interaction with a menu-drivesubroutine which allows any number of(mixed) multiple choice, true-false; orkey word questions to be entered. Themenu allows the author to select a ques-tion type or end the test. The processfor entering the text for all types ofquestions is similar to that for enter-ing the assignments. A multiple choicequestion may have any number of choicessince they will be presented one at atime to the student. (After all choiceshave been shown, they will be repeateduntil the student picks one.) The keyword question allows the author to specs-fy a number of keywords or characterstrings to be considered a correct answerto the question. In the student program,a key word question asks the student totype in a response. This response willbe searched by the program to determineif it contains any of the key words. Ifit does, it will be counted as correct.After entering a test question, the pro-gram returns to the menu which allowsother questions or the end of the test tobe specified. At the end of the test, theauthor is prompted to enter the competencylevel in percent for passing to the nextmain assignment. He is then prompted toenter his choice for a remedial strategyfor students who do not pass. When thisinformation has been entered, the programwill return to the master menu. From. themaster menu, the author may add other mod-ules as outlined above or print a masterstudent disk when all modules for theMASTS unit have been entered. When thelatter option is selected from the mastermenu, the program will prompt the authorto enter the information fields he wishesto gather on each student's disk. Theauthor simply types a one-line instruction,such as "Please enter your last name," foreach information field he wants. Any num-ber of fields may be entered, and they willbe presented before the student's firstassignment. Tim responses will be stored
Testing/Placement 203
on each student's disk.When the author finishes entering in-
formation fields, the author program willplace this information on the studentdisk, along with other index informationmaking the student disk complete. In-structions are then given to make copiesof the master student disk for distribu-tion, which necessitates the program re-turning to the microcomputer's disk-oper-ating system.
STUDENT DISKSEach student is given a copy of the
master student disk. To use his disk, astudent places it in any compatible mi-crocomputer and turns the microcomputeron. The program automatically loads andstarts at the proper point in the instruc-tional algorithm. On the first session,the program gathers the informationfields entered by the author and the firstassignment is given. Then according tothe instructions entered by the author:1) the program stops and will not contin-ue until the computer is turned on again,thus allowing the student to remove hisdisk and carry out the assignment beforereturning for a test; 2) the program pro-ceeds directly to the ti it on the_ assign -sent, in which case, the assignment textcould be used as an instructional framerather than as an assignment as such; or3) the program chains directly to anotherprogram which could be placed on the stu-dent disk by the author.
When a student finishes a test, he isgiven instructions which may be the nextmain assignment, a remedial assignment,the same assignment again, directions tosee the instructor (a block is placed onre-entry until the instructor unlocks thedisk), or instructions to turn the diskin, if the student has finished the lastassignment. All scores for each test arestored in the studex file which also holdsthe information on a student's locationin the algorithm and his re-entry status.
MASTS DATA COLLECTION AND DISPLAYWhen students have completed their
MASTS disks, the disks are turned in tothe instructor for data retrieval. Oneof the options on the authoring programmaster menu is collecting data from disks.Each student disk is placed in the seconddisk drive and is read by the author pro-gram aollect subroutine. The informationfields and scores for each test are storedwith those for all other students. Whenall student disks have been read, the in-structor may use the result display optionfrom the master menu. This option causesthe computer to chain to a data displayprogram that computes and displays
2 `r4
204 NECC 1980
statistical information for each test andeach student. If a particular test hasbeen repeated, the statistics are computedfor each attempt. this information may bedisplayed on the monitor, printed on aline printer, or both. The student datais stored in standard ASCII text files sothat simple programs may be written totransfer student data to other data baseformats used by more sophisticated statis-tical analysis packages.
APPLICATION AND EVALUATIONMASTS has been used for two semesters
in an upper-level media course for educa-tion majors at the University of Texas atAustin. Most of the class time for thiscourse is spent teaching production pro-cesses while MASTS is used to assign out-side readings describing media applica-tions. In the past, assignments were madewith a reading list, but many students didnot finish the readings. Another approachused was to require students to write re-action papers or short reports based onthe reading assignments, but this approachreduced the number of readings that couldbe assigned during a course. Using MASTSfor these assignments allows the instruc-tor to set due dates for completion thatcoincide with related in-class activities.
These first applications of MASTS haveserved as formative field evaluations help-ing to define needed changes and to catchbugs as only trial by fire can do. Anadded benefit from using MASTS has beenthat the computer anxieties of students,most of whom have never worked with a com-puter, have been lowered (according to pre-and post-MASTS questionnaires).
MASTS HARDWAREA Radio Shack TRS -80 with 48K memory
and two disk drive units is used for auth-oring and data collection from studentdisks. The student programs require onlyone disk drive and 32K memory.
CONCLUSIONSThe definite student enthusiasm and
lack of major problems encountered in im-plementing MASTS indicate an open door foxexploring the use of small scale CMI.MASTS or a similar system could form theheart of a programmed text or workbook.Another application might be in individuallearning centers, using all forms of in-structional materials as possible assign-ment media.
REFERENCES
Baker, Frank B. Computer Managed Instruc-tior. Englewood Cliffs, N.J.: Educetrail Technology Publications, Inc.,1978.
Bork, Alfred. "Course Management Systemfor Physics III." Proceedings of theConference on Computers in the Under-graduate Curricula,1977, 213-222.
Brockmeier, Richard. "Report on a HighlyUsed Computer Aid for the Professor inHis Grade and Record Keeping Tasks."Proceedings of the Conference on Com-puters in the Undergraduate Curricula,1977, 93-100.
Hilgendorf, Allen. A Study of the Trans-portability and Effectiveness of theUn-Stout CMIS and Individualized In-structional System Based Upon LearningStyles. Menomonie, Wisc.: Univ-Stout,1976. (ERIC Document Reproduction Ser-vice No. ED 131 305).
Nievergelt, Jurg, and others. Alterna-tive Delivery Systems for the Computer-Assisted Instruction Study ManagementSystem (CAISMS). Urbana, Ill.: Illi-nois University, February 197R. (ERICDocument Reproduction Service No. ED149 785).
Reigeluth, Charles M. TICCIT to the Fu-ture: Advances in Instructional Theo-ry for CAI. Salt Lake City, Utah:Brigham Young University, 1978. (ERICDocument Reproduction Service No. ED153 611).
Roll and Pasen. "CMI Produces BetterLearning in an Introductory PsychologyCourse." Proceedings on the Confer-ence on Computers in the UndergraduateCurricula, 1977, 229-238.
Van Metre, Nicholas H. "Problems in Re-searching Computer-Managed Instruct-ion." Paper presented at the AmericanEducational Research Association, To-ronto, Canada, March 1978.
21 5
RIBYT -- A DATA BASE SYSTEM FOR FORMAL TESTINGAND SELF-ASSESSMENT
F. Paul Fuhs
School of BussnessVirginia Commonwealth University
1015 Floyd Ave.Richmond. Va. 23284
804-257-1737
ABSTRACTThis paper describes the function and
structure of a data base system calledRIBYT, which simultaneously controls andassociates questions in question pools formany courses of instruction. The database stores questions created by both fac-ulty and students and is used for kormaltesting and student self-assessment. Thestructure of the data base allows rapidretrieval of multiple types of sets basedon predetermined logical associations.The data base also provides for the col-lection and association of feedbackinformation not only on student perform-ance, but also on the quality of the ques-tions in the data base. This quality isassessed through students' subjective com-ments about the questions they receive andthrough statistical item analysis. Feed-back is also provided to students on howthey are able to improve the quality oftheir question input, The structure ofthe data base is easily transferable tomany data base systems which currentlyexist.
INTRODUCTIONMany educators administer one or more
pools of objective test questions as thebasis for examinations. These pools usu-ally are in the form of multiple-choice,fill-in-the-blank, or column-matchingquestions. Educators have a need to gen-erate, organize, store, .modify, andretrieve such test questions'.
1Publiehing companies like Harcourt,Brace, Jovanovich (7) and SRA (10) haverecognized this need and are willing tosupply such pools in certain courses tofaculty who adopt their texts. Theirpools, however, lack associational struc-ture and are merely sequential filesorganized by text page number.
Testing/Placement 205
The logical associations among thequestions within a pool present a diffi-cult administrative problem. Questionsmust simultaneously be ordered in a numberof ways, they may be grouped into sets bycourse topics, by textbooks, by page num-bers within textbooks, by tests in whichthe questions were used. and by many othercategories. Question pools are also dif-ficult to manage because they are dynamicsquestions have different life spans, sub-ject to such factors as over-exposure inusage, changes in textbooks, and obsoles-cence due to technological changes. Newquestions must continually be added tothese pools and sach new question formsmultiple associations with previouslyexisting questions. In addition, periodicrewording and pruning of existing ques-tions are required and, unless the ques-tions in a pool are already well organ-ized. it is difficult to detect synonymousquestions.
Many traditional file systems haveattempted to solve these associationalproblems. Yet, frequently these systemsare expensive, overly complicated, anddifficult to use. They lack the capabil-ity to make multiple associations amongthe data items.
DATA BASE TECHNOLOGYData base technology over the past ten
years has been successfully applied tomany storage and retrieval problems espe-cially where multiple associations existamong data elements. Such associationsare logically represented as relations(2), hierarchical structures (8), and sim-ple or complex networks (9). They arephysically implemented through suchmethods as hashi0g, tables, and lists.The data base approach has penetratedbusiness and government and is in theprocess of replacing many traditional file
21
206 NECC 1980
systems. DBMS are currently availableeven on minicomputers like IBM's Sys-tem/38, Hewlett Packard and DEC machines.However, educational institutions havebeen slow to use data base systems exceptfor some purely administrative functions.
FUNCTIONS OF RIBYTIn this paper, we describe the func-
tions and structure of the data base sys-tem called RIBYT at Virginia CommonwealthUniversity. This system applies data basetechnology to the associational problemsof using and maintaining pools of ques-tions for formal testing and studentself- assessment2. The single data basecontains multiple pools of questions, onepool for each course using RIBYT. Thequestions are of various types, includingmultiple-choice questions with single ormultiple correct answers; fill-in-the-blank questions with multiple and alterna-tive correct answers; and column-matchingquestions. Each pool is under the directcontrol of the faculty members responsiblefor the course. In addition to faculty,students in some courses are allowed tocreate and add questions into the database. Student questions, however, are con-sidered unedited until reviewed by a fac-ulty member. Such questions, while physi-cally intermingled with edited questions,are logically distinct. During the reviewprocess, student questions are accepted,modified, or deleted3. The question poolsgrow at a much more rapid rate when stu-dents are allowed to enter questions.With a large pool of questions and thesecurity provided by a data base system,faculty allow student self-testing in anonthreatening spirit similar to theself-assessment procedures afforded mem-bers of ACM (M. Students view eitherfaculty selected questions or random setsof questions, and receive feedback ontheir performance. The faculty controlwhich questions the students are allowedto view, but students control (within lim-its) the number of questions, the coursequestion topics, and text page ranges overwhich questions may be selected for self-assessment. Students are generally enthu-siastic about participating in question
2The self-assessment aspect of RIBYT isone of the data base's unique character-istics, from which it derives its name,(Review It Before YOU Test).'The review process is also used to modifyor delete questions which have failedstatistical item analysis.
generation because they understand themerits of the self-testing.
To improve the quality of studentquestions, faculty, while reviewing thequestions, optionally enter into the database comments directed to the studentauthors of the questions. This feedbackis a learning experience in test construc-tion, but even more so in course content.There are two other feedback mechanisms toimprove the quality of the questions inthe data base. First, from an inspectionof patterns of student responses facultymay detect acceptable alternative respon-ses that were not previously considered.Secondly, students can enter into the database complaints concerning any of 'theself-assessment questions they havereceived. As a result faculty and stu-dents critique the self-assessment ques-tions. Questions for formal examinationsare selected from questions which havealready passed through these filters.
Student responses to test questionsare used not only for student performanceevaluation, but also for statistical itemanalysis. Each response entered into thedata base is associated with a uniquequestion number within the data base, acourse test number, the question numberwithin the test, the student making theresponse, and the student's score for theresponse.
The RIBYT data base system associatestwo course topics with each question inthe data base. Topics assigned to ques-tions by students are considered tempo-rary, until reviewed for editing b7 fac-ulty. Each course can organize itspermanent topics hierarchically andindependently of any other course. Thereis no limit to the depth of any of thehierarchies. The topics for all coursesare physically in the same data sets ofthe data base, but logically kept dis-tinct. The hierarchical organization oftopics provides flexible retrieval possi-bilities. Questions may be retrieved basedon specified topics. Alternatively, theymay also be retrieved by topics that existas children nodes in the hierarchicaltopic structure. Retrieval, therefor, ismade as generic or specific as desired.Faculty and student users can view at anytime the hierarchical structure of thetopics which, similar to the questionswithin any question pool, changes overtime.
IMPLEMENTATION OF RIBYTRIBYT was designed under the assump-
tion that faculty and students using thesystem would know little or nothing about
21 7
data base processing. Application pro-grams written in COBOL and BASIC act asthe communication between the DBMS and theusers, who access the system through batchand on-line devices. RIBYT is implementedon a Hewlett Packard 3000 Series II com-puter using the DBMS package, IMAGE, whichhas been widely acclaimed (5). The designof RIBYT is not restricted to IMAGE, how-ever. With minor modifications RIBYT canbe implemented on many network-orientedDBMS. CINCOM's TOTAL DBMS is but oneexample (3).
ASSOCIATIONAL STRUCTURE OF RIBYTFigures 1 and 2 illustrate the 22 data
sets (files) of RIBYT. There are eightmaster4 and 14 detail data sets. Masterdata sets contain records stored by hash-ing. Each master record, composed of oneor more data items, can be linked to oneor more detail data sets. Master datasets can be viewed as either independentfiles or as sets of records where eachrecord acts as a chain head to one or morerecords in a detail data set. Within adetail data set records are linked by for-ward and backward pointers to form sets. Adetail record may belong to one or moremaster data sets. Master data sets are oftwo types, manual and automatic. Unlikerecords in manual data sets, automaticdata set records are stored or deleted bythe DBMS itself. These records act asChain heads for records in detail datasets. With this data base architecture,networks of associations among records anddata sets can be created and then manipu-lated by the DBMS. Figures 1 and 2 showfor each data set its name, its type5, andits associations with other data sets.Within a detail data set a chain ofrecords can be further logically orderedby specifying one of the data items withinthe record as a sort item6.
4Aside from the terms "master" !_and"detail" which are peculiar to HewlettPackard, we will adhere to CODASYL(1)terminology as closely.as possible.
5In figures 1 and 2, B- Binary andX=Alphanumeric. The numeric value fol-lowing the alphanumeric specification "X"represents the number of bytes assignedto that data item.6In figures 1 and 2 these sort items aredesignated as "S" under the appropriatedata items.
Testing/Placement 207
Before faculty or students store orretrieve questions, information concerningthe users is stored in the data set calledROSTER; course information in the dataset TEXTBOOK-MASTER, and system initiali-zation information in the data set LAST-MASTER. Each record in ROSTER contains auser's name, course and section number, astatus (faculty or student), and a uniqueuser identification number. TEXTBOOK-MAS-TER contains one record for each block oftwenty pages of each textbook. Each recordcontains the textbook title, and a dataitem BOOK-BLOCK composed of course and anassociated textbook number and the pageblock. Such structuring gives more rapidretrieval for those questions based ontextbook page %webers. The data setLAST-MASTER contains one record. Storedhere are the sequence numbers of the lastquestion entered into the data bare(LAST-QUES), the last temporary topic(LAST-TEMP-TOPIC), the last permanenttopic (LAST-PERM-TOPIC), and a facultypassword (FAC-CODE). The latter allowsfaculty to perform such special functionsas dumping an entire question pool for acourse. In addition to the security pro-vided by the operating system and theDBMS, FAC-CODE and ROSTER are used to con-trol access to the data base. SYSTEM-STACK is the other system data set, imple-mented as a stack that containsinformation on the last 50 interactive orbatch jobs run. Each record contains atime stamp, plus a user and program iden-tification number. This information aidsin backing out of system malfunctions anddetecting unauthorized access.
A user, wishing to enter questionsinto the data base, begins a session at aCRT by choosing from a menu of textbooksassociated with the user's courses. Thismenu uses information from the data setsROSTER, ALTERNATIVE-COURSE, and TEXTBOOK-MASTER. Then, for each question, usersare prompted for background information onthe question before entering the text ofthe question. This information will beentered into the data set QUESTION-INDEX.The user enters two course topic numbers,a textbook page number, and the type ofquestion. The application program assignsa permanent or a temporary number to eachnew topic. The names and numbers of newtopics are entered into the data set TOP-IC-MASTER. If the user wishes to index a
'Mr readability the data.base, data sets,and data items are shown in capital let-ters.
208 NECC 1980
question on only a single topic, the sec-ond topic becomes a dummy. The programthen adds to the record in QUESTION-INDEXthe DATE, the BOOK-BLOCK, a unique ques-tion number (QUES -NUM), the version numberof the question (QUES-VERS), and theuser's identification number. The statusof the question (QUES-STATUS) is set to'edited" if the user is a faculty member,
'otherwise it is -set- to "unedited." Theother data items in QUESTION-INDEX areinitialized to zero.
After all data item values have beencoastructed for a record in QUESTION-IN-DEX, as the record is physically stored,the DBMS creates multiple associationsusing hashing and linked lists. The ques-tion's background information in QUESTION-INDEX is linked to its author in ROSTER,the textbook referenced in TEXTBOOK-MAS-TER, the question's topics in TOPIC-HEAD,and its unique question number in QUES-TION-HEAD.
The text for a question and itsanswers are next entered into the datasets QUESTIONS and ANSWER-CHOICES. Forfill-in-the-blank questions the text isstored in QUESTIONS and the answers arestored in ANSWER-CHOICES. Alternativelyacceptable answers for each blank arestored in ANSWER-CHOICES. For multiple -choice questions, the stem of the questionis stored in QUESTION and the choices inANSWER-CHOICES. Each choice is designatedas correct or wrong (CORRECT-WRONG). Onecan not afford to assign to each question,choice, and answer a fixed length equal tothe largest possible number of bytes thetext might need, rather, data base tech-nology is used to solve this problem.Text is considered as 40 or 72 Characterblocks linked together. Records in QU2S-VERS-HEAD are chain heads for sets ofcharacter blocks in the data sets QUES-TIONS and ANSWER-CHOICES. Within the dataset QUESTIONS each textual block of char-acters is assigned a sequential line num-ber (LINE-NUM), which is used as a sortitem to keep the text in correct logicalorder on a chaine. Since any choice of amultiple - choice question may also havemore than one line of text, the sort itemcalled SEQUENCE is composed of the alpha-betic choice designator (A,BiCiDtor E)concatenated to a text line number withinthe claim. This system automatically
ftecords in any of the data sets are notnecessarily stored in the same order in_which they ;sere .input.
keeps the choices and the lines of textwithin each choice in logical order. Thetotal number of times each choice isselected as an answer in either formaltesting or self-testing is later stored inthe data item RESP-TOTAL.
QUESTION RETRIEVALThere are many selection options for
question retrieval from RIBYT, aside fromretrieval for editing purposes. Questionretrieval is first divided into formaltesting and self-testing. Faculty canrestrict questions so that these are usedonly in formal testing. Each type oftesting is further divided into threeoptions. Under the first option facultyspecify by question number the actualquestions to be included in a test. Underthe second option faculty specify that asingle set of questions are to be randomlyselected for an entire class. under thethird option faculty indicate that a dif-ferent set of questions are to be randomlyselected for each student. Faculty mayspecify that questions are to be selectedbased on one or more topics, page ranges,or any combination of topics and pageranges. In self-testing, faculty mayleave topic and page selection to eachstudent's choice.
DATA SETS USED IN QUESTION RETRIEVALThree data seta are used to control
question retrieval. These data sets con-tain specifications that are used by aretrieval program to select questions fromthe data base for formal testing andself-testing. The first data set, TESTS,has a primary key COURSE-TEST-NUN, whichis a combination of the course designationand a test number within the course. Eachrecord in TESTS represents a single test.Each record has a SECURITY-CODE to controlaccess to the test. The percentage thateach test represents towards the totalcourse grade is stored in TEST-WEIGHT.The number of questions per test, the num-ber of times a test may be taken, theaverage difficulty factor per test ques-tion, the types of questions, the selec-tion options, and the mix between vali-dated and unvalidated questions are storedin the data set TESTS. When faculty wishto designate the actual test questions,the set of question numbers is stored inthe data set SPECIFIED-QUES by questionnumber (VERB -NUM). Each number acts as asearch,argument to retrieve the text ofone questl m, This retrieval is performedas follows. First the system uses thevalue VERS -NUM in SPECIFIED-QUES to hashinto the data set QUES -VERS -HEAD. Theretrieved record than points to sets of
22
textual blocks, representing a question.its choices, and answers in the data setsQUESTIONS and ANSWER-CHOICES.
The data set PAGES-TOPICS is a multi-purpose data set. Each record in PAGES-TOPICS stores either a textbook page rangeor one or two course topic identificationnumbers. Each record is linked to anindividual test in the data set TESTS.The value of the data item PAGE -TOPIC -SWITCH indicates-whether a record containspage or topic information. For pages, thedata items START and END store page rangesand the data item BOOK-HIER stores acourse textbook identification number.For topics, two topic identification num-bers can be stored per record; one inSTART and one in END. BOOK-HIER is thenused as a binary switch to indicatewhether only designated topics or theirhierarchical children are to be includedin question retrieval. Question retrievalbased on textbook pages is accomplished byhashing into TEXTBOOK-MASTER and followingappropriate chains into the data set QUES-TION-INDEX and from there using the com-bined data items, QUES-NUM and QUES-VERS,as a key to hash into QUES-VERS-HEAD.Then the program follows chains into QUES-TIONS and ANSWER-CHOICES as describedabove. Question retrisVal by topics issimilar after entry into QUESTION-INDEXfrom TOPIC-HEAD. Since all hierarchicaltwins of a given parent topic are organ-ized as a set in the data set TOPIC-INDEX,whose members are linked to a topic inTOPIC-HEAD, sub-topics are found by firsthashing into TOPIC-HEAD and following achain into TOPIC-INDEX.
STORING AND RETRIEVING TEST RESPONSES ANDTEST SCORES
Information concerning student respon-ses is stored in six data sets/ TEST-SCORES, TESTS, RESPONSES, STUDENT- GRIPES.ANSWER-CHOICES, and RESP-FEEDBACK. Theset of all test scores for a given test islinked together within the data set TEST-SCORES, and this set is itself linked tothe data set TESTS. Within the data setTEST-SCORES is a chain associating alltests for each student. Since RIBYT han-dles the possibility that the same studenttakes the same test more than once, eachtest score is made unique by being associ-ated with a time stamp DATE-TIME in TEST-SCORES.
The student responses to each testquestion are stored in the data setRESPONSES. Each record is composed of aquestion identification number (VERS-NUM),the question number within the test, thetext '.ine number within the response, acode indicating whether the response was
Testing/Plaoment 209
correct or incorrect, the student'sresponse, and a combined data item (USER-DATE-TIME) formed from the student identi-fication number and the time stamp. Eachresponse in the data set RESPONSES is ontwo link paths, one originating from thedata set QUES-VERS-HEAD, the other fromthe data set TEST-HEAD. In addition,since it is desirable to have the respon-ses continually sorted by test questionnumber within a test and since there maybe more than one text block per studentresponse, the two sort items, TEST-QUESNUM and LINE-NUM, are used as major andminor sort fields for the records on aUSER-DATE-TIME chain.
Besides receiving the traditional per-formance feedback to their responses (alist of correct answers, questions, andtest scores), students may receive textualcomments concerning particular responses.Faculty store these in the data set RESPFEEDBACK. After students receive feedbackon their tests, they may enter into thedata set STUDENT-GRIPES their reactions toany individual test questions, anony-mously, or by student identification num-ber. Course grades can easily be deter-mined based on the data sets TESTS andTEST-SCORES.
STATISTICAL ITEM ANALYSISTwo measures of a question's quality
under statistical item analysis are itemdifficulty and item discrimination (4). Adifficulty index is calculated concur-rently for every question in a test. Theset of student responses is retrieved byhashing into the data set TESTS with thekey COURSE-TESTNUM and obtaining allchained records in the data set TEST-SCORES. For each record retrieved thedata item values of USER-NUM and DATE -TIMEare used as a concatenated key to hashinto the data set TEST-HEAD. All testresponses for a single student areretrieved from the data item CORRECT-WRONGof the data set RESPONSES through a chainoriginating in the retrieved TEST-HEADrecord, and this procedure is thenrepeated for each student. The retrievalis rapid and passes over all other testsin the data base. The calculated diffi-culty indices are then storcd in the dataset QUESTION-INDEX. The calculations forstatistical item discrimination requirethe same data sets as item difficultydetermination.
CONCLUSIONThe associational power inherent in
the data base system RIBYT allows manyrelationships to be made and maintainedamong courses, textbooks, coursa topics,
220
210 NECC 1980
students, faculty, tests, test questions,test construction options, and studentresponses to questions. RIBYT stores sub-jective and statistical feedback on thequality of the questions in the data baseand aseociates this feedback with the testquestions. This system supports formaltesting and self-assessment procedures andsalads the storage of many types of ques-tions, while imposing no limit on the sizeof any question or answer,---The centrali-zation of the logical associations intoone data base provides better security andeasier maintenance than traditional filesystems.
ACKNOWLEDGEMENTSI am grateful for the assistance
of Dr. Kathleen cokrigan Fuhs and Mr.Paul Thompson for helping to specifyand clarify some of the user require-ments of RIBYT.
REFERENCES
(1) CODASYL, CODASYL Data Base TaskGrp* Report, ACM, New York,
(2) Cwdd, E. F., "A Relational MOdelof Data for Large Shared DataBanks," CALM, 13,6, 1979.
(7) Hilgard, E. R., R.L. Atkinson,and R. C. Atkinson, Introductionto Psychology, 7th ed., New York:Harcourt, Brace, and Jovanovich,1979.
(8) IBM, IIMB/VS, Application Pr ram-aka eference Manual, or er no.SR-20-1026, whit7-17017ins, NewYork, 1978.
(9) Kroenke, David, Data Base Processing,Chicago: SRA, 19777-
(10) SRA, "Mid -Term Blues? ComputerizedTest Service Available," Data Proc-essing News (Spring 1979)71
(11) Wong, J. W. and G. Scott Graham,"Self.-Assessment Procedure VI," CACM,22,8 (August 1979),449-454.
(Manual)LAST-MASTER
STACK LAST-
QUESLAST-
TEMP-TOPIC
LAST-
PERM-TOPIC
FAC-
CODE
X4 B B B X8
key
---- ..
(detail)
SYSTEM-STACK
:.;ACK DATE-TIME
LAST-USER
LAST-PROGRAM
X4 X12 X10 X4
I S
FIGURE 1: SYSTEM CONTROL DATA SETS OF RIBYT
(Auttratic) (Manual) (Automatic)
TOPIC-HEAO TEXTBOOK- QUESTION-
MASTER HEAD
TOPICURN TITLBOOKE
X30
KBOOBLOCK
X12
key
(Manual)ROSTER
LAST-NAME
FIRST-NAME COURSE \SECTION STATUS
USER -
NUN
X20 XI6 X6 ' X2 X2 XIO-
key
1 _
(Detail)QUESTION-INDEX
TOPIC-
NMITOPIC-NUN2
DATE BOOKBLOCK
PAGE-NUN
QUES-NUR
QUES-'MRS
DIFFI.CULTY
QUES -
TYPE.QUES-DISCRINI.,°"STATUS NATION
;"NUM
X4 X4 X6 X12 X4 X4 X2 X2 X2 X2 B -X10
,S S S S. . .
(Detail)TOPIC-INDEX
TOPIC- SUB- SUB -
NUM TOPIC Bic.
X4 X30 X4
(Detail)
QUES-FEEOBACK
(Detail)ALTERNATE-COURSE
VERS-NUM
LINE-NUM
FEE°maala X2 MU X10
MINIM
USER -
NUN COURSE SECTION
X10 X6 X2
I
FIGURE 2: DATA SETS OF RIBYT
(Manual)
TOPIC-MASTER
TOPIC
ICOURSE \TOPIC NUN
X6 X30 X4
241 223
(Automatic)QtES-YERS4AD
VERS-_. NUN
X6
(Detail)
QUESTIONS
VERS- LINE-NUN NUN TEXT
X6 X2 X72
S-
(Detail)RESP-FEEDBACK
(Detail)
ANSWER-CHOICES
VERS- SEQUENCE RESP- CORRECT- TEXTNUNTOTAL WRONG
X6 X4 B. I---..-X2 X40-
S
VERS-NUN
LINE-NUN FEEDBACK
X6 X211. X72
S
VERS-- HUH
X6
(Detail)'
RESPONSES
TEST-S.
NUN
LNUINE-
NCORRECT-WRONG
X6
4X6
-s
X2 X2
(Detail)STUDENT- GRIPES
STUD-NI/4
LINE-NUN GRIPE
X10 X2 1 X72
S
(sqealTEGTu-HEAD
)
USER-DATE-TINE
X22
STUDENT-RESPONSE
USER-DATE-T
X40 X22
Figure 2 (cent): DATA SETS OF RIBYT
2?
(manual)TESTS
,0,,,,I0U5E;
TEST:SECURITY-CODE WEIGHT"'u"
HUM-OF-
QUES
TEST-TIMES
AVE-OIFF
QUES-TYPES
SELECT-OPTIONS
VALID -
MIX
X12 X8 8
-
8 8_ 8
-
-..
X2. X2
.
8
key
...._ ,
. ,
I ,
-I, -
(dtai)SPECIFIC/lIVES
COURSE- %MRS-
TEST-HUM NUM
X22 X6
2 5
(detail)
PAGES-TOPICS
COURSE-TEST-NUM
START ENO BOOK-
HIERPAGE-TOPIC-SWITCH
X12 X4 X4 X2 X2. , .
Figure 2 (cont): DATA SETS OF RIBYT
(detail)TEST-SCORES
(fromRoster)
COURSE-
TEST-HUM
---
CATE-TINE
SCORE
_
USER-NUM
X12 X22 8
.
X10
S
214 NECC 1980
COMPUTER MANAGED PLACEMENT
IN
MATHEMATICS INSTRUCTION
FOR
HEALTH OCCUPATIONS STUDENTS
Thomas A. BoylePurdue UniversityWest Lafayette, Indiana 47907(317) 749-2256
.INTRODUCTION
The Indiana Vocational TechnicalCollege at Indianapolis (IVTCIN) main-tains an instructional program to servethe mathematics needs of students be-ginning study in health occupationstechnologies. The program is based onStreeter and Alexander's Fundamentalsof Arithmetic (1), beginnlTifh-ngtwhole number operations, continuingwith fractions, decimals, and percents,and extending to ratios and signednumbers. The materials are organisedin 15 modules, each with a pre -test,a practice test, and a mastery testused to guide individual students tothe study materials appropriate totheir evident needs.
Although the instructionalmaterials in the program provedappropriate and were reasonably wellreceived by students using them,certain problems soon developed in'guiding students to the modules speci-fic to their needs. in particular, thefollowing were noted:
1. The handling of pre-tests onan individual basis was time consumingand otherwise a burden on testingpersonnel.
2. Because of the limited numberof pre-test 1orms and the level ofsupervision for actual testing, it waspossible for students to use the pre-test materials in ways which would foolthemselves regarding their mathematicsskills.
3. Capable students were botheredby the need for working through severaltests in order either to find a' suitablestarting module or to demonstratecompetency and by-pass the programentirely.
Peter T. MagnantIndiana Vocational Technical College1815 East Washington StreetIndianapolis, Indiana 46202
4. Neither the student nor thestaff could readily get an impression ofhow the student stood with respect tothe whole program.COMPUTER APPLICATION
in an effort to mitigate theseproblems, work was begun early in 1978to adapt test materials and procedureswhich had been in use at another XVTCinstitution (2). These materials con-sist of a programmed test format, whichrequire. students to attempt equal num-bers of test items in each of severalcategories, together with computerscoring and data processing. This useof a computer enables first the scoringof each student in each item category,with subsequent determination of scorestatistics for each student group. Aswill be shown in subsequent examples,this use of a computer yields informa-tion in detail which would be practi-cally impossible to get from hand scor-ing. Previous work with the combina-tion of test format and computer dataprocessing had demonstrated efficiencyin the use of test administration timeand the possibility of rendering in-formation from each of several relative-ly independent item categories (3).
Figure 1 shows the computer scor-ing output for a select group ofstudents. The category of subtestscores appear in six columns followingthe student names. A key at the bottomserves to identify six subtests. Forthis test some grouping of modules wasnecessary, e.g., instructional modules1-3 dealing with whole numbers are allrepresented in subtest one. The keyfurther identifies the SCR column as
** 6 = SIGNED NUMBERS **** SCR = SUM OP SMUT SCORES ** NF = NUMBER OF ITEMS ATTEMPTED ** RATIO SCORE = SCORE
DIVIDED BY RELATED NP **Figure 1. Facaimile*GlftthemfticeScaingPrcammOiftlat
holding the sum of six subtest and the NPcolumn as showing the total number ofitems attempted by each student. Numbersat the far right identify the first ldmistakes made by each student.TEST SCORING
Subtext scores for each student arepresented in two numbers. The first, theraw score for each subtext, is based onone point for each correct response. Apenalty is applied for making errors insequence, this scoring yields an expectedscore of zero for a student strategy ofpure guessing. If performance is worsethan that from pure guessing, i.e., if astudent consistently makes errors in asubtest, then that subtest score becomesnegative. The second number, the ratioscore appearing below each raw score, is
toe result obtained from dividing the rawscore by the number of related items at-tempted. As should be evident, the com-putation of seven raw scores and sevenratio 'corm would present a formidabletask for anyone attempting hand scoring ofthis type of test.APPROPRIATE PLACEMENT
The students represented in Figure 1were selected to present a range of stu-dent performance. Bob, the first student,obtained a row of scores which could wellresult from guessing. It appears likelythat he made consistent errors in itemsdealing with fractions and ratios, but heearned only three points in seven attemptswith whole-numbers, so we may be reason-ably confident Bob should start at thebeginning of the instructional program.
218 NECC 1980
INDIANA VOCATIONAL TECHNICAL COLLEGEINDIANAPOLIS
M - 31 MATHEMATICS SKILLS ASSESSMENT PROGRAMFOR
Pat W. 4 679 80981 44
THIS PROGRAM IS AIMED AT HELPING YOU DEVELOP THE BASIC MATHEMATICAL SKILLS NEEDEDTO DO AN EFFECTIVE am IN THE IVTC SPECIALTY YOU INTEND TO ENTER. THE PROGRAM HELPS YOUAND YOUR INSTRUCTOR FIRST BY NOTING THE PARTS OF THE IVTC MATHEMATICS SEQUENCE IN WHICHYOU MAY ALREADY HAVE ADEQUATE SKILL. NOTE IS ALSO MADE OF THE SECTIONS, OR MODULES INWHICH YOU APPEAR TO NEED ADDITIONAL HELP AND PRACTICE. AS YOU KNOW, MOST PERSONS EN-TERING TECHNICAL EDUCATION NEED TO IMPROVE THEIR SKILLS IN SOME PARTS-OF-MATHEMATICS.
THE PRIMARY PURPOSE OF THIS ASSESSMENT PROGRAM IS TO HELP YOU GET STARTED AT THEPOINT WHERE YOU CAN DO YOURSELF THE MOST GOOD., ONCE YOU KNOW THE MATHEMATICAL MODULESYOU NEED TO WORK ON, YOU WILL FIND YOUR INSTRUCTOR AND THE LEARNING LABORATORY STAFFREADY TO HELP. ,AND AFTER A PERIOD OF INSTRUCTION AND PRACTICE, YOU CAN CALL ON THESKILLS ASSESSMENT PROGRAM TO CONFIRM THE PROGRESS YOU HAVE MADE.
THE FOLLOWING MESSAGES SHOULD HELP -
Pat W.YOU WORK SOMEWHAT FASTER THAN THE AVERAGE WHEN TAKING THE TEST. THIS IS A REASON-
ABLE STRATEGY, OUT YOU MAY BE GUESSING ANSWERS TO ITEMS WHICH YOU COULD SOLVE IF YOUSPENT A LITTLE MORE TIME ON THEM.
IT APPEARS YOU DO YOUR BEST WORK WITH THE MATHEMATICS IN MODULES 1 - 3. YOUR SCOREOF 100 INDICATES ADEQUATE SKILL WITH WHOLE NUMBERS.
JUDGING FROM YOUR RESPONSES, IT APPEARS LIKELY THAT YOU HAVE ADECIJATE SKILLS IN THEFOLLOWING TOPICS.
TOPIC SCORE
PERCENTS 83
IVTC MATH MODULES
TEN
ALTHOUGH YOUR SCORES DO INDICATE SUFFICIENT SKILL, THERE MAY BE SOME ROOM FOR IM-PROVEMENT, AND YOU MAY BENEFIT FROM A QUICK REVIEW OF ONE OR MORE OP THE MODULES LISTED.
EVIDENTLY YOU HAVE MADE SOME PROGRESS IN LEARNING ABOUT FRACTIONS, HOWEVER YOU DOAPPEAR TO NEED MORE PRACTICE WITH THE MATERIALS IN MODULES 4 - 7.
PROBABLY YOU CAN DO YOURSELF A LOT OF GOOD.BY GETTING TO WORK ON THE MATERIALS INMODULES 8 - 9, EVIDENTLY YOU DO NOT HANDLE DECIMALS VERY WELL.
NO DOUBT, YOU REALIZE THAT YOU KNOW SOMETHINGS ABOUT PATIOS, BUT YOU DO MISS SOME POINTSHERE AND PROBABLY WILL BENEFIT FROM ATTENTION TO IVTC MATH MODULI ELEVEN.
EVIDENTLY YOU MADE MISTAKES ON THE FOLLOWING TEST ITEMS
The next three students receivedvery low scores inmost 'attests, yetscores which are evidently adequate inthe whole number subtest. Bev had thelowest total score (-7) observed to date,but did five out of six whole- numberitems correctly. Evidently these studentswill benefit from starting with the in-structional materials on fractions. patwill also benefit from starting with theinstructional materials on fractions,however, she obtained acceptable scores
in other subtests.Pan, Dee. And soy, the next students
listed, had scores characteristic of ade-quate performance in two or three sub -tests. Recommendations for these studentswould lead to intermediate modules in theinstructional program. The scores forPAM and Dee indicate some study is neededbeginning with module eight. Joy maywell be advised to begin with module ten.
The two students remaining, Lis andKay, received practically perfect scores
229
in all but one subtext. Evidently thesegirls made all but one of their mistakeson items in the percent subtext. May'sscores are especially noteworthy. Sheworked rapidly on the test, attempting 48items in the 30 minutes allowed. Shegot all items correct in five subtestsyet made errors in all eight of the per-cent items she tried: the negative sevenscore results because the first error isnot penalized.INDIVIDUAL OUTPUT
The computer program which producesthe test scores has been extended toyield a one-page message for each studenton which the student's ratio scores areranked, and the subtests on which he didbest are identified. Appropriate mes-sages are selected, including recommenda-tions for notion. Duplicate copies areproduced for the student's file and forinstructional supervisors. A facsimileof this output is presented as Figure 2.
INKRUCIFICING.
FFOGIOM
1 48 mem =ARMMO 71
2 28 MIID LAS TECHMO 61
3 46 RADEECGIC 'MICENs 57
4 42 SURIOCALIECHMO 57
5 44 LICPRACEICRSEN- 455
6 37 minawAserMO 39
7 93 MUMMIES:REMOMO 31
TestingiPlacement 217
SCORE STATISTICSDuring the period to July 1979, the
M-31 test was administered to approxi-mately 1000 IVTCIN students. Prom thetotal group, some 763 were identified byinstructional program. The response datafrom these students were grouped and testscore statistics obtained for studentsentering each of seven health occupationsprograms at IVTCIA. The numbers of stu-dents identified by_ instructional programranged from 31 for health career prepa-ration to 455 entering the practicalnursing program. Numbers of students ingroups may appear to change and the sumsof groups may be different from the totaltabulated in Figure 3. This is becauseadditional data were received and pro-cessed during the time the report was inpreparation.
The score statistics were arrangedby ranking the mean total ratio scoresfor the instructional groups. This
1NDIMMIKMIXAMLIECENVFICILLEGE1101/MOLys
14-31 19111114=8 VMS= SUMMRRY
AUGUST, 1979
029421 DE m= maw maw exam Ian rissNOS. PEACTICAS NOS. SOIREMOrtED
MEAN .93 75 .74 .49 .76 .71 .73 40.56STD DEN .09 .35 .33 .44 .39 .34 .22 8.62
MEAN .93 .72 .75 .45 .72 .70 .71 40.079P1) MY .15 .41 .29 .52 .37 .37 .25 9.06
wan .92 .62 .66 .32 .61 .55 .61 38.52ETD DEV .15 .42 .32 .50 .47 .44 .26 8.69
115t9111 3. 14-31 =SE E5312791,101 DIMSWEID BY INVIRICYICti% GROUP
23
218 NECC 1960
ranking placed the respiratory theraphy studentshighest, and the health career preparationstudents latest. The 0:Inplete statistics for allseven Student grape are presented in Figure 3.Mean scores are presented for all six of theN-31 subtests, as well as for total score andnumber of item attempted. With the exceptionof the items-attepted mean, the scores repre-sented are ratio mores (i.e., the raw scoredivided by the number of related items atteipted)for each student. As can be seen the subtestmeans range fran 0.95, correeronding to a mean ofninety-five percent, down to a negative 0.06.This latter score indicates average test per-formance just slightly worse than would be mc-rented for a etudett strategy of pure guessing.
Figure 3 Indicates that, an the average, the14-31 test is of reasonable difficulty forstudents entering health service instructionprograms at IPDZIN. However, the difficulty isnot even for different groups. The respiratorytherapy and medical laboratory techniciantrainees and the radiologic aid surgical techni-cians find the test quite easy. The statisticsghat that, especially in the first grouPs, manystudents; handle the designated range ofarithmetic skill adequately. Unquestionably,there are many of these students vim have littleto gain from further study of the skills =pre-sented. Among the students who do seen to needfurther related study, the need appears to begreatest for percent drills.
Data for all students, including sanetested after mid-July, We that there is sanecorrespondence between the arltleneticthat have been =Allred by the entering studentsard the sequence of interactional reddest. Onaverage, all groups score highest onthe whole amber subtest. With the eocceptionof the percent scores, there is a trend towardlaser scores in the subtests related to laterinstructional modules. The break in the se-quence of scores, at the percent subtlest,
may indicate either need for special effort ifstudents are to attain 80 percent performanceor disPlY that the test items meresonby difficult. t1te heck dose call at-tention to an advantage of the test format:evidently saw students oho have difficulty withtest item related to module 10 readily surpassthe criterion BO- percent perfonerms in itemrelated to module 11 and to signed ambers.Here the test faint and amputee data proces-sing enable identification of the difficultmaterials and may obviate the need for subse-quent modules.SIGNIFIGINZE
TIO3 statistical teats have been performedon 14-31 data obtained to date. One-wayanalyses of variance have been clone at eachooltent4Figure 3) of group subtest scores tocheck for significant differences between
grout's, for example to see *tether the respiratorytherapy students and the LPN's could be regardedas caning from the same PoPoleriro- The analyseswere me twice, once with and once without thehealth career preparation group. The results arepresented in Table x la teats of P. values.
WRTHOUf WITH;MEM SCP MCP
1 %tole Numbers 0.545 5.271*
2 Fractions 2.099 4.473*
3 Decimals 1.581 7.425*
4 Percents 5.246* 7.455*
5 Ratio 4.172* 7.413*
6 Signed mamba= 7.747* 8.135*
*Significant at 0.01 ac less
Table x. F-values for analysis ofvariance in subtext scares,with and without healthcare preparation students.
These values show that, excluding the OCPstudents, only marginal differences appearbetween groups wititxecex4 to ichole =bar,teactica, and decimal'ambtest scores. Signifi-cant differences appear in the other subtastscores. When the RCP students are included,sg=icant differences appear in all subtlest
APPLICATICN OF MUMSZang the ptvoedures discussed, students are
able to review their am strengths and weekneesesrelated to mathematics, both as individualapplicants and in relation to the other Maim**seeking admission into a program area. TheIndiana Vocational Technical College at Indianap-olis helps students in their areas of deficiencysib they can develop the mathematical skillsnecessary for admission aid eventual success. Ifscores are acceptable for program adedssicn, theRelated Education Department can use the date asa basis to increase mathematic skill in light ofthe health programs needs and objectives.
aben imbalance develops belmeen the ambersof students seeking to enter different healthspecialities, an 1ndividual's scores may wag*encouragement for consideration of otherspecialities also within the student's preterit orfuture range of mathematical developeent. Forexample, sane students seeking to enter thepractical nursing program may beneficially
231
ansider other instructimal program.OICELBICHS
A ccmputer-tesed testing prooedure has beenadapted to menage the mathematics placement ofMadames pursuing inetzuctia, in health mama-time. The anaiattion enables more efficientuse of scixol persamel and expedites pleasant
Testing/Placement 219
in, or by-pas3 of, a sequence of asthmaticsinstruction SWUM. Scam normative data havebeen obtained and ease significant differencesseen to exist between scores of studentsseeking inatructian insdifferent healthspecialities.
REFERIEWES
1. Streeter, Jame and Alexander, Gerald. Fundamentals of Aritbatic. New York: Raper and AcesPublishers, 1975.
2. Boyle, Timms A. and Shaver, Carroll G. Ccmwtar Amassment of Mathematics Skills for Studentsbeginniza Postnday Technical Evalw. Seventh 0:inference on Omenzbers in the Maks-gradeate OnTicula. KrOaftn: state university of New York, 1.976.
3. Boyle, Thaw A. and Wright, Guy L. Itcarater-aanisted Evaluatica of Student Acheivement."Engineering klucatice, vol. 68, No. 3, Decenber 1977, m. 241-245.
232
Invited Sessions
IMPROVING UTILIZATION OP TWO-YEARCOLLEGE COMPUTER CENTERS
Robert L. BurrowsTriton College
2000 Fifth AvenueRiver Grove, Illinois 60171
(312) 456-0300
ABSTRACTth=difiC11118i011 will begin with an over-
view of the areas of responsibility of acomputer-assisted learning agency and thenelaborate on what can be done in theseareas once one actually becomes an agency.Specific topic areas include the following:--Marketing approaches to attract teachersto the computer via such means asseminars, newsletters, and committees.
--Training instructors in computers,using as specific examples the threecourses now being offered to Tritonfaculty each semester: Introductionto Computers and Programming, Intro-duction to Statistical Analysis UsingSPSS, and Introduction to ComputerGraphics.
--Working with faculty to obtaineducational software includingcatalogs, magazines, educationalsuppliers, and user groups.
--Supporting software written or plannedfor academic users, including changesto system software, program filterswritten to convert foreign software,an electronic mailing system, aprogram directory system, and a carsystem.
--Working with faculty in obtaininghardware, using as example theterminals and microcomputers purchasedat Triton and its plan for a computerlab.
--Problems with computer educationspecific to a community college.
2')')4, to.
TEACHING .COMPUTER ETHICS
Walter ManerPhilosophy.and Computer Science
Old Dominion UniversityNorfolk, VA 23508
ABSTRACT--Wriinagemont, staff, and users ofinformation systems of all kinds willbenefit greatly from training in appliedprofessional ethics. These professionalswork in an environment where they mustdeal responsibly and knowledgeably withcritical moral problems (such as breau4of privacy, computer crime, and dehuman-ization) which are aggravated, trans-formed, or created by the advance ofcomputer technology.A general rational for a course in
computer ethics will, therefore, bedeveloped along with course designcriteria, a full course description, aset of proposed course objectives,associated bibliographies, and a taxonomyof subject matter areas within the fieldof information ethics.Attendees will break into small discus-
sion groups to study cases illustratingmoral dilemmas posed by the use ofcomputers in education.
23 .f
Invited Sessions 221
Tutorial
PASCAL TUTORIAL
Harry P. HaidukAmarillo College
P.O. Box 447Amarillo, Texas 79178
ABSTRACT--YETrEutorial is concerned with
(1) a brief historical view of PASCAL interms of its philosphy and stated designgoals
(2) a brief review of its current rele-vance in a diverse set of applications,particularly as it may relate to the newDepartment of Defense Common High OrderLanguage, ADA
(3) a comparison of PASCAL's logic anddata structures with those of BASIC, COBOL,FORTRAN, and PL/I
(4) actual running program examples con-trasting PASCAL with BASIC and FORTRAN.
222
2 3 5
IMM'y A11111111===1.11
Pre-College Instructional Materials
COMPUTER -BASED INSTRUCTION FOR THE PUBLIC SCHOOLS:A SUITABLE TASK FOR MICROPROCESSORS?
Dr. Timothy D. TaylorComputer Based Education Center
308 Carroll HallThe University of Akron
Akron , Ohio 44325(216) 375 7848
OVERVIEWThe microprocessor is proving itself
as an attractive, low-cost device whichhas the capability to serve its owner as arecord-keeper, accountant, entertainer,and tutor. The attractiveness of the de-vice has led teachers and other educatorsto explore its potential for providingcomputer-based instruction. While no de-vice, whether it be a microprocessor, alarge computer, or something in between,should be viewed as simply good or badfor education, all devices have charac-teristics which must be considered beforean investment is made. One might expectthat a computer terminal connected to alarge computer (resident or non-resident)might have capabilities that microcompu-ters lack, and this is certainly the case.Whether or not a school should purchase amicro, however, depends entirely upon thefunction which the machine will have.Following is a discussion of the educa-tional services that can and cannot beprovided by off-the-shelf, low-cost,microprocessors such as the Apple IX orRadio Shack's TRS-80 (Model I).
DISPLAY CHARACTERISTICSThe variety of characters, graphics,
and colors that microprocessors are ca-pable of displaying has been a major fac-tor in the success of its sales duringrecent years. The Apple II and OhioScientific micros, for exawle, can pro-duce graphics, colors, and sounds, inaddition to displaying standard keyboardcharacters. Graphics are often helpful,particularly for the ruching of such sub-jects as physics and hisher level mathema-tics, but the most important characteristic
223
required for much of the current instruc-tion in the public schools is the displayof upper- and lower-case alphabetic char-acters. None of the low-cost micropro-cessors is sold with upper- and lower-casedisplays, although most of them can bemodified, at extra cost, to add this capa-city. The use of color and sound can addsignificantly to the appeal of micropro-cessor-based courseware, but the lack ofdual-case alphabetic characters has to beconsidered a serious limitation for theteaching of such subjects as English andreading.
SIMPLICITY OF OPERATIONA public school student using a termi-
nal tied to a large computer typically be-gins his session by being seated and typ-ing one short message. At that point, hehas entered the computer-based educationcourse and is working where he left offduring his last session or where his pre-vious performance records indicate that heshould be. One CBE course may containenough material to teach the student andtrack his progress for a year or more. Astudent using a microprocessor (presumablywith at least one disk drive) typicallyhas a more complicated procedure for begirt-ming his session. He will obtain his disk,be seated, insert his disk into the diskdrive, power-on the machine so that theinitial_ program will be loaded into memory,and theft he is ready to begin, He may ormay not have to type a message telling thecomputer to start the lesson. He probablycannot start where he left off last time;however, a carefully programmed lesson mayenable him to choose his own startingpoint. One disk may contain enough meter!?
236
224 NECC 1980
al to tutor the student for a number ofhours.
In the public schools, students usingterminals connected to a large host compu-ter can be given the freedom to use theterminals without teacher supervision be-cause the sign-on procedure is simple andbecause no peripheral materials, such asfloppy disks, are required. Use of amicroprocessor may require the monitoringof a disk storage area and a check-out,check-in systeu for students to borrowdisks. Since a standard 5-inch floppydisk contains only a limited amount of ma-terial, the student will frequently needto exchange his disk in order to switch toanother topic.
In terms of simplicity of operation,the advantage appears to rest with theterminal connected to a large host compu-ter.
ADDRESSING A VARIETY OF LEARNING NEEDSA typical, well-planned CRE course con-
tains a sufficient quantity of instructionto meet the needs of a variety of differ-ent students: the slow learner, the rapidlearner, the student who knows nothingabout the topic being presented, the stu-dent who already has familiarity with thesubject matter. As the student worksthrough the instruction, his performancewill cause him to be branched forward foradvanced material, backward for additionalpractice, or laterally for s discussion oftopics tangentially related to the primarytopic being presented. The number ofpaths that a student could take throughthe material is infinite because thestructure provides almost limitless bran-ching opportunities. Since no two stu-dents possess identical needs, it is pro-bable that no two students will progressalong the same path of instruction. Acomputer-based course should be extremelyflexible, broad in scope, and capable ofhandling students' individual performancestyles. Failure to attain these charac-teristics constitutes a failure to placeCIE into a category eeparate from othermedia such as lectures, textbooks, andaudio-visual tapes, all of which possessa predominantly linear flow. :-
The above characteristics make presentcomputer-based instruction impractical, ifnot impossible, on today's most popularmicroprocessors. Microprocessor tapes,disks, and memory have a size limitationwhich prohibits them from delivering a CREcourse with the breadth of scope describedabove. At best, a student using a microfor this type of instruction would findhimself swapping disks constantly in orderto branch ahead, backward, and laterally.Rut microprocessor courseware with this
V. 4'.
degree of individualization is not avail-able today. The educational software cur-rently available for microprocessors con-sists of short lessons, each addressing aspecific topic. Seldom does any continu-ity exist among these short lessons. Toemploy these microprocessor lessons forgenuine computer-based education wouldmean that decisions as to which studentsneed which lessons at what time would bemade by teachers, aides, or by the stu-dents themselves. Whether these personswould make the correct decisions, andwhether continuity could be provided amongthe lessons collected, is questionable.Assuming that continuity could be achievedstudents would have to be given paper-and-pencil tests constantly in order to deter-mine which lessons were needed by eachstudent and in what sequence. Such tests ,
are unnecessary on a large computer sys-tem since branching occurs automaticallyas a result of present and past studentperformance. In a well-designed CRE les-son, students are generally unaware thatsuch intricate branching is taking place.They know only that the computer is movingin s logical direction from the beginningto the end of topics that they need tolearn.
TRACKING STUDENT PROGRESSMicroprocessors and large CIE systems
both have the capacity to provide on-lineassessments, to give the student some in-formation about his progrese, and tobranch him to various locations based onhis performance. (As mentioned before,the physical size limitations of themicroprocessor's disk would frequentlynecessitate disk-swapping in order toswitch topics.) Out in order to build agenuine performance history on each stu-dent, a large system is required. Large-system CRE can relieve the teacher of theburden of administering periodic tests andplotting each student's growth. By typingappropriate keywords at the computer ter-minal, the instructor can view immediatelyany student's current and past performancerecords. Typical information includes thefollowing:
1) the student's first and last dateon-line.
2) the total number of hours andminutes spent.
3) topics where mastery has or hasnot been demonstrated.
4) areas of particular difficulty.5) areas not tutored due to excellent
pretest performance.6) average response time on drill-and-
practice items.7) length of individual sessions.8) students' comments on parts or all
23 7
of the instruction.This information can be displayed for onestudent, a group of students, or all stu-dents in a teacher's class. Hardcopy ter-minals are often used so that the reportmay be kept for future reference. Teacherscan quickly identify any students havingdifficulty, and students can become veryhighly motivated when they see that theyhave achieved progress from one week (ormonth, or year) to the next.
The quantity of student performancedata afforded by the microprocessor istrivial when compared to the data routinelykept on the large system. Many (perhapsmost) of today's tutorial microprocessorlessons give no performance statisticswhatsoever. Those lessons that do giveperformance information do not store itpermanently for the teacher's later inspec-tion. Performance data are collected andreported to the student as he is working,but when his session is over and he re-moves the disk from the disk drive, allcollected information is instantaneouslylost. Although virtually no microproces-sor courseware exists that builds a his-tory of student performance, it is possi-ble to do so. The most practical way toaccomplish such a task is to use a multi-ple-disk system. In a two-disk system,for example, one disk can be devoted tocollection and storage of performancedata while the other contains lessons be-ing presented. In a public school, eachstudent would be assigned one or moredisk for storage of his own performancedata, and these disks would be checked inand out along with the disks containingthe instruction. Instruction would haveto be programmed carefully so that eachlesson would store performance data on adifferent part of the data-collectingdisks. An instructor wishing to view theperformance histories of his 30 studentswould sit at the microprocessor with hiscollection of 30 or more disks and slipthem in, one at a time. As each student'sprogress is displayed, the informationcould be duplicated in hardcopy form ifa printer is attached to the micropro-cessor. Obviously, this inspection ofstudent progress would be a time-consum-ing task, and the performance data pro-bably would not be as complete as thedata collected on a large system.
It was mentioned above that typicalperformance data provided on large CBEsystems includes the student's first andlast dates on-line, total time spent oninstruction, and average response timeon drill-and-practice exercises. Unfor-tunately, many of today's popular micro-processors cannot record any of this infor-mation. It is possible to ask the student
Pre-College Instructional Materials 225
what the date is or how long it took himto answer a question, and then the stu-dent's response can be recorded, but thereliability of such data would have to bequestioned. The absence of an internalclock precludes the recording of dates orelapsed time by the microprocessor. Othertypes of data, such as topics mastered andscores earned during on-line assessment,can be recorded by the microprocessor ifthe lessons are programmed to do so and ifa system of data-collecting disks, likethe one described above, is established,
In short, the collection and reportingof student performance data are essentialin order to track progress, provide studentmotivation, and isolate learning problems.Large CBE.systems address this need verywell. Microprocessors could provide some(but not all) of the same student perfor-mance records only after a fairly complica-ted system of specially written lessons,student data disks, and multiple disk-drivehardware is created. Retrieval of informa-tion from this system would be significant-ly more cumbersome than it is presently onlarger CBE systems.
languages are numerous and varied. Manycomputer languages perform similar tasksequally well, but some languages arespecialized to perform certain tasks betterthan the .others.
Computer languages that have been de-signed to provide instruction are sometimesreferred to as instructional languages.Such languages typically record some basicinformation automatically for every studenton every CBE course, such as his first andlast usage dates, his total time on eachlesson, and hia current location withineach lesson. In addition, storageties are provided to enable the courseauthor (programmer) to store an infinitevariety of performance data permanently ortemporarily. A student progress report ismerely a systematic display of these stor-age facilities. An instructional languagealso permits the course author to accept awide range of responses from the student.For example, if the author expects "false"as the answer to a question but realizesthat students may misspell the word, he canuse a atat-ment which essentially says, "Ifthe student types a five-letter word begin-ning with 'f' and ending with 'e', I recog-nize this as the correct answer." Anothercharacteristic some instructional languageshave is the automatic re-starting of thestudent at the point where he left off dur-ing his list session. Characteristics suchas these are not found in computer lan-guages that have been designed for purposes
233
226 NECC 1980
other than instruction.BASIC is an example of a non-instruc-
tional language, not because it is defi-cient in some way, but because instructionis not the purpose for which it was crea-ted. It is a general-purpose computinglanguage which is standard on virtuallyall microprocessors. The language has nopre-defined storage area for recordingstudent performance data, no provisionsfor accepting a wide range of responses,and no automatic restarting of the studentwhere he left off last tine. Naturally,the BASIC language has capabilities thatsome instructional languages lick, notablyin the area of mathematical computation.Wu if BASIC is to be used on microproces-sors for CBE, numerous sophisticated sub-routines must be created in order to giveBASIC the characteristics of an instruc-tional language. These subroutines willoccupy space on the disk and permit lessspace for instruction and collection ofperformance data. The subroutines must beduplicated and stored on every disk usedfor instruction. And, in spite of theirsophistication, the subroutines will notpermit the recording of certain types ofdata such as dates and student responsetime unless the microprocessor chosen con-tains an internal clock.
As mentioned above, most languages canaccomplish most tasks. But if a choice ismade available, it is only logical to pickthe language which is least cumt.ersome.On large systems, the choices are numerousand include many instructional languages.With microprocessors, the choices are fewor nonexistent and do not include instruc-tional languages.. On the basis of lan-guages, it appears that present-day micro-processors are not yet prepared to handlethe level of aophistication found in large-system CBE courseware.
LARGE - SYSTEM AND SHALL-SYSTEMCHARACTERISTICS
Regardless of the language employed orthe performance data collected, there arecertain. inherent characteristics of largesystems which must be evaluated prior tomaking a decision to employ or abandon sucha system. On the negative side, when thelarge computer malfunctions, all terminalsconnected to it become useless. But thatall terminals are using one computer alsohas advantages. A major advantage is theability to avoid duplication of effort.For example, when a courseware revisionbecomes necessary on a large system, onecorrection constitutes a system-widechange. But if 50 microprocessors wereused in place of a 50-terminal network,then 50 individual corrections would needto be madeone for each copy of the defec-
tive course. Such duplication of effortis a general characteristic of micropro-cessors. Each user has his own computerand his own copy of each program he uses.Fifty users wishing to use a mathematicslesson would need to purchase 50 copies ofthat lesson. On a large system, only onecopy of the lesson is needed, and it canbe used on any number of terminals simul-taneously.
Another important feature of largesystems is that all users are accessingthe same disk storage area. It is thischaracteristic which permits teachers toobtain class reports on groups of students.Such reports often point out the best,worst, and average oerformance record ofthe group (sometimes including statisticssuch as mean, range, and standard devia-tion), thus giving the teacher an immedi-ate picture of the group's homo or hetero-geneity. With micro-based CBE, perfor-mance data are likely to be scatteredacross many disks, thus making such groupreports impossible or impractical.
Large systems often possess a message-handling capacity which is not possible onmicroprocessors. Messages can be sentbetween course authors and students, forexample. A student who is puzzled or whowishes to communicate with the author ofhis lesson for any reason can usually loga comment which the author will later readand answer. Such communication assiststhe author in locating any areas of,hislesson which may cause confusion amongstudents and enables teachers or othersupervising persons to post administrativeannouncements to users of the system.Some large systems also permit on-line di-rect communication between terminals.
Microprocessors necessitate a duplica-tion of effort and lack message capabili-ties, but they do provide a greater degreeof independence and self-sufficiency. Asmentioned above, a computer malfunction ona large system renders all terminals use-less simultaneously, but one microproces-sor's failure has no influence on othermicroprocessors. A micro is also moreportable than a terminal connected to alarge computer since the latter requireseither a hardwired communications line ora telephone line. And, of course, thereare no monthly rental charges after amicroprocessor and peripheral equipmentare purchased.
WHY BUY A MICRO?A microprocessor can serve as a valu-
able educational tool, particularly inteaching students about computer hardware,and can help them develop an understandingof what happens when a program is writtenand executed. Mathematics students can
239
also profit from a study of the BASIC lan-guage, which can be used effectively tosolve math problems using algorithms.Limited amounts of instruction, of course,can be provided in tutorial or drill-and-practice form as well But nothing withthe scope, complexity, and record-keepingabilities of large-system computer-basededucation can be accomplished ih a reason-able way on today's microprocessors.
LAUNCHING A CBE PROJECT WITHMICROPROCESSORS
Assuming that apublic school has con-sidered the pros and cons of the low-costmicroprocessor as well as those of the CBEterminal connected to a large system, andassuming that the micro was chosen, theseare some of the steps that would probablyhave to be taken:
1) purchase a microprocessor.2) purchase at least two disk-drives
in order to provide the capabilityfor storing and retrieving studentperformance data.
3) purchase floppy disks for datastorage and for storage of theinstruction itself.
4) purchase a printer in order toprovide student progress reportsin hardcopy form.
5) modify the microprocessor so thatit will display upper- and lower-case alphabetic characters neededfor the teaching of spelling,English, reading, and relatedsubjects.
6) design machine-language or BASICsubroutines to handle tasks al-ready designed in instructionallanguages.
7) copy these subroutines onto alldisks to be used for, lesson con-tent and/or all disks to be usedfor data collection.
8) initialize each disk to be used forlesson content or data collection.
9) build courseware which branchesforward, backward, and laterallyto accommodate individual learningstyles and which keeps detailedstudent performance records on thedata-collecting disks.
10) establish and supervise a systemfor storing lesson and data-collecting disks, keeping track ofwhich disks are borrowed (and re-turned) by whom, and assuring thateach student has access only to hisown data disk.
11) make copies of lesson disks if theproject involves more than onemicroprocessor.
12) when errors are detected, make thecorrections on all disks containing
Pre-College instructional Materials 227
the defective lesson or data-recording rou-tine.
The above steps vary in complexityand would probably require years to accom-plish. The characteristics of such a sys-tem are contrasted with the characteristicsof present day large systems in the SummaryChart which appears at the end of thisarticle.
NOTHING IS IMPOSSIBLEIt is not the author's intention to
state that microprocessors are useless forcomputer-based education. In fact, most ofthe undesirable characteristics mentionedcan be rectified through hardware modifica-tion, addition of peripheral devices, or bysimply purchasing a larger microprocessor.But it is the small, comparatively unsophis-ticated microprocessor that has"been at-tracting the most attention in the market-place. This author becomes concerned whenteachers and other consumers become en-tranced by those very attractive devicesand when salesmen with little or no experi-ence in computer-based education make exag-gerated claims. Quality microprocessorcourseware which has been tested objective-ly and shown to be of educational value isNOT widely available. Microprocessorcourseware which provides the individuali-zation and record-keeping of large systemsis not available at all. Although it ISpossible to buy a microprocessor with upper-and lower-case alphabetic characters, andalthough larger disk drives CAN be used soas to reduce the amount of disk-swapping,and although an internal clock CAN be added,such modifications result in a home brewsystem which is costly and which makes thein-house development of courseware manda-tory. It is this message that the sales-men are failing to convey.
THE FUTUREMicroprocessors will probably play an
important role in future CBE systems for atleast two reasons:
1) Their capabilities will increase.It is hoped that a greater choice of compu-ting languages will be available along withgreater storage capabilities and, conse-quently, less required disk-swapping. Aninternal clock will enable the micro torecord dates and keep track of elapsed time.Upper- and lower-case characters will becommonplace.
2) Downloading will become morepractical. Downloading essentially meansthat the micro can communicate with a lar-ger machine for the purpose of copying aprogram into its own memory or onto its owndisk. Then the program can be used locally,independent of the large machine. Afterthe session is completed, the updated
240
228 NECC 1980
records can be transmitted back to thelarger computer.
It is rather impractical to attempt toemulate large-system CBE with today's smal-ler microprocessors. A tremendous amountof time and effort would be required, andthe final product would contain built-indisadvantages. By the time such a systemis built, better micro-based systems willprobably be available. At that time a re-evaluation of the enhancements made onmicroprocessors as well as on large sys-tems will be necessary.
DESIRABLE CHARACTERISTICS
Typical Low-CostLarge PopularSystem Micro
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Involves a one-time cost without monthly rental charges.
Displays graphics and colors, produces sounds.
Displays both upper- and lower-case alphabetic characters.
Has complete portability, not dependent on a telephone line.
Requires one step (one short message) to start instruction.
Affects only one user when the computer malfunctions.
Provides performance feedback while the student is on-line.
Provides detailed performance history for teacher's laterinspection.
Enables students to continue where they left off during theirlast session.
Records dates and students' response_ times.
Records session time and permits instructor-defined sessionlength for individuals or for all students.
Generates class reports showing histories of groups or students.
Has well-constructed courseware immediately available.
Is a system well-adapted for extensive branching to accommodateindividual learning styles.
Supports instructional computer languages.
Permits nessages among students, authors, teachers, and others.
2 II
UNDESIRABLE
TypicalLargeSystem
CHARACTERISTICS
Low-CostPopularMicro
X
X
X
Pre-College Instructional Materials 229
Involves a monthly rental charge.
Lacks graphics, colors, and sound.
---Requires a dedicated line or telephone line.
Has upper-case display only, unless modified.
Requires executing several tasks in order to start instruction.
Requires constant disk-swapping if individual learning stylesare to be accommodated.
Requires a disk storage and check-in facility which most besupervised.
Necessitates adding more disks proportional to the quantity ofperformance data desired.
Requires creating new courseware since virtually none existspresently.
Providing an appropriate education fot allchildren in the least possible restrictive envi-ronment will require drastic changes in our educa-tional processes, ftom assessment and prescriptionthrough instruction and monitoring. RecentlyBenjasin Bloom has stated, "If we are convincedthat a good education is necessary fot all wholive in modern society, then we 151st search fotthe alterable variables that can make s differencein the learning of children and adults in or outof school" (Bloom, 1980). The alterable variablesSloom emphasizes are time-on-task, cognitiveeuttY, and formative t4Stin. These variableshelp explain the learning interactions betweenteacher and student. Using these variables impliesindividualizing the entire educational process,ftom assessment to the monitoring of learning.
Mainstreaming is currently viewed as anecessity in providing the least restrictiveenvironment. Mainstreaming handicapped studentscrestes a special burden for the classroom teacher,however, introducing handicapped students intothe regulat classroom greatly inctsases the rangeof intellectual capacity and experience of thestudents. This broadened range makes the individ-ualization of instruction a necessity, althoughindividualization will be extremely difficultconsidering the limited resources available to theregular classroom teacher,
Educators have long felt that the computerholds a special promise in providing individual-ized instruction, but this promise has not beenfulfilled because of the limitations of audiovisual hardware ..id the high cost of computers.Recent changes in hardware should alleviate manyof these limitations. The microcomputer hasgreatly reduced the cost of computing and thevideodisc has the capacity to provide the audioand visual components necessary for effectiveindividualized instruction via computer-assistedinstruction (CAI).
A CAI system utilizing a microcomputer andvideodisc is currently being developed as pertof a research project conducted by Utah StateUniveruity's Exceptional Child Center, Thesystem is designed to be used by nonreaders and
specifically by mentally retarded children andadults. Even though the system is designed to beused by e handicapped, it will have all thecomponents necessary for a general purpose CAI
system. With the appropriate courseware thesystem would be relevant to learners at any levelof intellectual ability and expetience.
The system, teferred to as the MCVD (Micro-Computet/VideoDisc) System, was otiginally devel-oped ftom funds of a small grant received from theUnivetsity Research Office. Subsequently a stoatftom the Media and Captioned Films Division ofthe Bureau of Education for the Handicapped ofDREW was awarded. The grant is for two years andbegan October 1, 1979.
The major goal of the project is to develop,evaluate, and demonstrate a CAI system for usewith mentally handicapped learnets. The system is
unique since it can commmajcate with nonreaders,a capability nonexistent or very limited in pastCAI applications. Computer-assisted instructionsystems have been devised and developed to commun-icate with nonreaders, but in most cases they haveused serial devices such as audio tope recordersand slide projectors. The difficulty with thesesystems is a relatively long access time whenbranching to different segments of an instruction-al program. A short access time (less than 3seconds) is ctitical in most CAI applications,but especially critical when wotking with mentallyhandicapped learners since, at best, their atten-tion is difficult to maintain.
The hardware fot the MCVD system consists ofthe MCA videodisc player, an APPLE II microcom-puter with a digital disk, a Sony 12-inch tele-vision monitor, and a Carroll Manufacturing light -
lututruPt touch panel. The MCA 7802 videodiscplayer was selected becauae of its random accesscapabilities, and the APPLE II microcomputersystem was selected because of its pottability,reliability, and color graphics capabilities.A color monitor was chosen because of the addi-tional flexibility provided by color. The tout.%
screen is a light -intetrupt system which allowsthe learner to interact with the system by touch-ing the scream. The hardware components are
2.1:3
enclosed in a cabinet that disassembles for trans.-partation purposes. The monitor will be mountedon an adjustable pedestal that swivels, tilts,and can be adjusted for elevation. This flexi-bility and adjustment is necessary to accommodatethe wide range of ages and varying degrees ofmotor ability. The cabinet is mounted on retract-able wheels to facilitate mobility.
The computer programs that control the video-disc and receive input from the touch panel arewritten in Applesoft BASIC. These programswere originally written in PILOT, which we aban-doned because of problems with the PILOT trans-lator. PILOT was originally selected because ofits ease of use. lie are currently exploringvarious PILOT translators and compilers. We havedetermined that a PILOT interpreter does not pro-vide for sufficient execution speed to interactwith the three devices. We are also currentlyinvestigating the potential of PASCAL as a lang-uage for the system.
Communication between the three hardwaredeilces is accomplished through an interface boardthat was designed and built by USU's microcomputerlab. The interface board also can programmatic-ally switch the video source to the monitor fromeither the APPLE /I or the videodisc. This capa-bility allows us to transmit APPLE II graphics andaudio to the monitor while the videodisc is search-ing. It also allows us to present video informa-tion from the APPLE II and audio from the video-disc simultaneoua3y, which provides for a greatdeal of flexibility in altering the video imagesfrom the videodisc.
At present the system interacts with thelearner by presenting an audio instruction andthe associated visual image on the monitor viathe videodisc. The learner responds by point-ing to an object on the monitor screen. Whenthe learner touches the screen, two light beamstransmitted from each axis of the touch panelare interrupted, and the point of interruptionis detected by the touch panel. The X and Tcoordinates from the point of interrupt aretransmitted to the microcomputer. The computerprogram in the microcomputer contains the correctcoordinates for each segment of instruction, andthe coordinates transmitted by the touch panelare compared to these correct coordinates. If
the response is correct, the microcomputerresponds by referencing the segment on the video-disc which contains audio and visual positivefeedback. Other possible response conditionsare a wrong response, a close response, and anon-response. (A non-response is detected whenthe learner does not respond in a specifiedperiod of time. Recorded segments are containedon the videodisc for these response conditions aswell as a variety of feedbatk, including animationand motion picture sequences. Each segment ofinstruction has associated parameters that specifythe number of times a learner must respond cor-rectly to advance to the next segment and thenumber of trials allowed before the teacher is
Pre-College Instructional Materials 231
signaled for help. While the learner interactswith the system, data are collected by the micro-computer and stored on the APPLE II magnetic disk.
Various data are maintained which track aleatner's progress through a program of instruc-tion. The data are then available for analyticpurposes and are also used by the system to allowthe learner to start at a point where the previoussession was terminated. This automatic restartoption can be overridden by the teacher, who canspecify the point at which the learner is tocontinue. The data is readily available in hard -copy form via a Centronics Printer.
Pour instructional programs are currently invarious stages of development. These instruction-al programs are developed by the Outreach andDevelopment Division of the Exceptional ChildCenter and have been field tested and validated.They were chosen primarily because the instruc-tional sequences had been validated and becausethey are adaptable to the MCVD format. Thefollowing is a brief description of the programs:
1. Matching Program (Hofmeister, 77) - Thisprogram teaches the child.to match objectsthat are alike in_size, colpr, or shape suchas idencifying squares with other squares.This skill is necessary before the child canlearn the more advanced skills such as namesof colors and shapes and reading. Thisprogram was not designed to teach a child toto name colors, sizes, or shapes.2. Timetellina Program (Rofmeister, 75) -This program is to be used with any childor adult who cannot look at a clock andrecognize what time it is to the minute. It
has been especially useful in tutoringprograms for the slow learner, although it isnot restricted to such use. Upon completionof this program, the learner should be ableto look at any clock, showing any time, andrecognize the correct time to the minute.3. Recognition of Functional Words(Rofmeister, 76) - This program providesinstructions for teaching the learner torecognize functional words. This programteaches the recognition of twenty differentwords important to everyday living (e.g. stop,go, pull, push, etc.).4. Identification of Coins (Rofmeleter, 77)This program teaches the learner to recognizecoins upon sight.Preliminary development of the Matching
Program was completed in May 1979. A preliminaryfield test was conducted and changes were made tothe system as a result of this field test. Asecond preliminary field test was conducted andchanges were made to the system as a result ofthis field test. A second preliminary field testof the Matching Program is currently being con-ducte4. It was anticipated that this field testwould be tompleted by the submission time of thispaper, however, the field test will be continuedthrough February 8, 1980, in order to collectadditional data. This second field test was
24.4t
232 NECC 1980
necessary tom (1) evaluate the changes in thesystem that were precipitated by the first fieldtest; and (2) further define the population.
The following problem areas were determinedas a result of the first field test:
1. Even though the search times were rela-tively short (less than 2.5 seconds), theattention of the child was lost when themonitor went blank during a search time.2. The system was giving negative feedbackto a correct response when the child wasrequired to touch multiple objects on thescreen.
3. The child seemed to lose interest inspecific positive feedback segments if theywere repeated several times.la an attempt to rectify these problems a
number of changes were made to the system:L. A programmable switch was added to the;interface board that allowed the computerprogram to switch the source of video fromeither the microcomputer or the videodisc.This switch enabled the system to present acomputer-generated graphic while the video-disc was searching.2. The algorithm in the computer programused to detect Nultiple responses on thescreen was changed. We determined that thecomputer program was not responding rapidlyenough to detect rapid responses from thelearner. The algorithm was changed, and arapid increase in response time was realised.3. The algorithm used to select positivefeedback segments was changed so that thesegments could be presented randomly.The six children involved in the second
field test are from the special education class -roomest the Exceptional Child Center. Thesechildren are functioning at a lower mental devel-opment level than those involved in the firstfield test. The purpose of this change was todetermine if children at the lower level ofdevelopment could effectively interact with thesystem. The only criteria established forselection was that the children have sufficientmotor skills to hold and point .ith a smallpointer and sufficient receptive language tounderstand simple commands such as "touch theone like this," "that was good," "that was notcorrect," and "try again."
The field test has been in operation forthree weeks at the time of writing. Duringthis time numerous problems in working with thetest population have been identified. Sincethe field test is continuing, the data c.onotat this point be analyzed and specific con»elusions cannot be made.
The first half of the Timetelling Programis planned for production in March 1980. Thefollowing entry-level skills have been estab-lished for learners using timetelling courseware(these entry level skills will define the popula-tion):
1. Learners must be able to count to sixty
and recognize :gushers 1-12.
2. Learners must be able to identify basicgeometric shapes (e.g. circles, squares,etc.).
3. Learners must have sufficient motorskills to hold and use the pointer.At the completion of the first half of the
program, learners will be able tom1. Sequentially count and place the numbers1-12 around the clockface.2. Recognize the digital format (e.g. 2: )
for the little hand.3. Determine the value of the little handwith digital cues on the clock face and thebig hand as a distractor.As a result of information gained during the
second field test of the Matching Program, it wasdetermined that physical prompting and practicein_using the. pointer are required before inter-acting with the abstract objects on the tele-vision screen. The Timetelling Program is beingdeveloped to involve the teacher (or anotherstudent) in the beginning phase of the program.When appropriate interaction with the CAI systemis achieved, the system takes over, allowing thelearner to work independently with the CA1
system.Based on the unique capabilities of the
MicroComputer /VideoDisc System, the followingcourseware design considerations were formulatedand incorporated into the development of theTimetelling Programs
1. Although the availability of a secondaudio track affords many instructionaloptions, such as instruction with differentvoices or in a second language, as course-ware development proceeded both positiveand negative, it was determined the secondaudio track should be used to store specificfeedback, to the learner on his performance(see Figure 1). This feedback is specificto a particular segment of instruction asopposed to the standard feedback blockswhich are used by numerous instructionalsegments. Generally, it is the instructionalsequences at the beginning of the programthat require specific feedback. An examplewould be "good touching the one."2. Strategic placement of the standardfeedback blocks can substantially reducethe access time required by the videodiscplayer to search for a feedback segment.Because of the nonlinear nature of the MCVDsystem, standard feedback is used oftenthroughout the program. A regressionequation developed by Woolley and DeBlooisallows the prediction of access time basedon the number of frames (Woolley, 1980). Acriterion-of 2.5 seconds was established forthe maximum search time for providing eitherpositive or negative feedback. Based on the
regression equation, the feedback segmentswould have to be located within 6900 frames.of any instructional segments that access it.
241.-
Meeting our 2.5 second search criteria re-quired multiple placement of the standardfeedback blocks.3. Periodic testing is facilitated by thenonlinear capabilities of the MCVD system.The timetelling courseware and microcomputersoftware are designed to branch to appro-priate practice frames for criterion test-ing. This branching insures that the learnerwill be tested only on the material the learnerhas already covered. Criteria have beenestablished at 00X accuracy. Testing,consisting of II frames, begins and continuesonly as long as the learner meets the SOXcriterion. In other words, as soon as twoframes have inaccurate responses, the softwarebranches the learner back to the initiallesson for continued practice. Conversely,should the learner meet the BOX criterion onhis or her seventh frame, the last test framewill be omitted.4. The amount of repetition can be varied forindividual learners based upon their previousperformance. The learner who achieves apredetermined level of accuracy (determinedduring field testing) may be branched aheadto the criterion test and given the opportun-ity to demonstrate mastery of that particularconcept. Learners having difficulty with Coesame concepts can be given additional repeti-tion and subsequent prectia:-z,rThe two preliminary field tegiti and produc-
tion experience have yielded valuable informationon the development of videodisc-based CAI. Pre-sently one of the most limiting disadvantages ofCAI involving the videodisc is the high cost of discpressing. This presents instructional coursewaredevelopers with an entirely different set ofinstructional development problems. After a discis pressed, you have to live with it. Therefore,any experience gained from this project or otherprojects should provide valuable information indeveloping videodisc programs. Essentially,a developer needs all of the up -front knowledgeavailable before a disc is pressed.
REFERENCESBloom, B.S. "The New Direction in Educational
Research." Phi Delta Keegan. 1980, Vol.16, No. 6.
Nofmeister, A.M. 6 Gallery, M. A Program forTeaching the Identification of Coins. Niles,Illinois: Developmental Learning Materials,1977.
Nofmeister, A.M., Gallery, M. 6 Landon, J.J."Matching Sires, Shapes and Colors." AnInstructional program prepared by the Excep-tional Child Center, Utah State University,Logan, Utah for the Utah Division of FamilyServices, 1977.
Parent Teaching Package - Word Recognition."Available at the Outreach 6 DevelopmentDivision, Exceptional Child Center, Utah StateUniversity, Logan, Utah, 1976.
Woolley, R.D. and DeBloois, M.L. "PreliminaryBenchmark Data for the PR 17820 DiscovisionAssociates Videodisc Player." Center forInstructional Product Development TechnicalReport 11. Utah State University, Logan,Utah, 1980.
234 NECC 1980
Audio Track 1Audio Track 2
Video
Control TrackSync TrackAudio Mad( 3
One Segment of Instruction-4 Seconds on Vk isotope120 Frames on Videodisc
Directio of Tape Travel
One Inch Type C Helical VideotapeUsed For Initial Production
Figure 1
2 A.
Linear Framework of Timetelling Package
1 .instruction
AY.
_1 Instruction;i1A4.
....f.:::: Instruction, 4,
1,417`c
v.v.42Instruction
\
Standard Feedback Block ISFBI
Figure 2
2-1s
236 NECC 1080
THE NON-TECHNICAL FACTORS IN Tat DEVELOPMENT OF CAI
Michael MocciolaComputing Center/Academic Computing
Pace UniversityNew York, New York 10038
ABSTRACTThe most challenging opportunities and
the most serious problems that inhibit thedevelopment of computer-aided learning ineducation are non-technical. In mosteducational institutions decision making,communications, and bargaining about thefunding of computing, the purchase ofcomputer equipment, or the application ofcomputer methods is divided among themajor populations of the schools; schoolboard members, administrators, teachers,department heads, and students. Tounderstand the socio-political factorsrelated to computer-aided learning, thejob-related responsibilities of each ofthese populations must be carefullyexamined. Computer educators then needto use the socio-political forces to thebest advantage of students and teachersto provide enhanced classroom instruction.Educators must recognize and foresee theneeds of the schu'l community, have earlyaccess to pertinent information, andcommunicate effectively.
21o;
Invited Sessions
DATA SETS AVAILABLE FROM THE FEDERAL GOVERNMENT
chaired by Thomas E. BrownGeneral Services Administration
National Archives and Records ServiceWashington, DC 20108
(202) 724-10,d
ABSTRACTMany agencies of the O. S. Government
routinely make computerized data availableto educators and researchers. To do this,some agencies provide reference service forthe data which the agency created, otheragencies distribute data originally createdby other agencies, and still ether agenciesserve as a clearinghouse for a given sub-ject and refer researchers to agencieshaving information on that subject. This
""session will have representatives fromthree agencies performing these threefunctions. At the end of the session, oneshould have a clear indication of the pro-cedures involved in locating and obtainingdata sets from the Federal government.
SPEAKERS:EFrationcn, Coordinator, College Cur-
riculum Support Project, Bureau of theCensus: a discussion of the serviceswhich the Bureau of the Census providesto enable researchers to gain access todata collected by the Census Bureau.
Ross J. Cameron, Archivist, NationalArchives and Records Service: an outlineof the functions of the Machine-ReadableArchives Division in its mission to inven7__toffy, obtain, and provide reference ser-vice :or data created by other Federalagencies.
Edward D. Mooney, Program Specialist,National Center for Educational Statistics:a presentation on the Federal InteragencyConsortium of Users of EducationalStatistics whose purpose is to facilitateaccess to education data in the Federalgovernment.
237250
Minority InstitutionsECMI
ACADEMIC COMPUTING:A SAMPLER OP APPROACHESIN MINORITY INSTITUTIONS
Sister Patricia MarshallXavier University of Louisiana
7325 Palmetto StreetNew Orleans, LA 70125
504/486-7411
A variety of post-secondary minorityinstitutions using a variety of approachesto academic computing were interviewedcampus-wide and in-depth late in Fall 1979.These interviews were part of a larger as-sessment of needs in educational computingat 239 minority institutions.(1)
Faculty, students, and administrators..were_interviewei_on computing development,usage, problems, and successes. Diverseapproaches were discovered, correspondingto philosophic, demographic, geographic,Historical, political, and cultural fac-tors.
The institutions were selected in orderto obtain as broad a cross section as pos-sible within time and financial limits andparameters such as ethnic composition, typeof control, date of establishment, highestlevel of offering, academic orientation,enrollment, type of access to hardware, andexperience. While not all of the most suc-cessful institutions were chosen, we didtry to include schools which would not feelthreatened by the interviews and which hadhad enough experience to identify factorsinhibiting and promoting progress. Con-sciously excluded were such institutionsas the Universities of Hawaii and PuertoRico; partly for financial reasons andpartly because they are so much larger thanminority institutions in general. Thetable on the next page lists the institu»tions interviewed and some key parametersused Li their selection. For confidential-ity we identified them only by number. Wewill report here on four of them; these arenamed below with permission.
Institution #3
The Community College of Baltimore isau Eastern seaboard, public, urban, two-
238
year, three - fourths black institution es-tablished after World War II. It enrollsover 9,000 students in day and eveningdictator Two campuses, one emphasisingtechnical studies and one evolving towarda focus on business studies, are unitedunder one administrative structure. Largenumbers of minority students began to at-tend this school early in the 7Os as aresult of outreach by the institution.The stable faculty consists of only 28percent minorities. Although this insti-tution may appear small by national stan-dards, it is one of the largest of theminority schools and one of the few two-year colleges with any large degree ofexperience and planning in academic com-puting.
Unencumbered by the weight of tradi-tion and spurred by local employmentneeds, the Community College of Baltimorehad established a separate academic de-partment for data processing by the earlydate of 1965. Thia department, now calledComputer and Information Systems (CIS),attracts 300 majors and graduates about 25each year with an associate degree in com-puter studies. Four full-time and numer-ous part-time faculty staff the department.
The school has weat'ered tnree hardwarephases and is entering its fourth. Ini-tially an IBM 1620, acquired for the dataprocessing department in 1965, was theonly computer. A committee from electron-ics, data processing, and administrationhired a manager of computer services anddeveloped specifications for a UNIVAC 9300,which arrived in 1968. The college allo-cated its funds to lease it and pay supportstaff for administrative services. Depart-ments and individual users, however, havenever been charged.
1972 brought a UNIVAC 9480 (131K) with
Zvi
Minority institutionsECMI 239
TABLE 1
MINORITY POST-SECONDARY INSTITUTIONS
INTERVIEWED ON ACADEMIC COMPUTING, FALL 1979
/NWT-TUT/ON
DATEESTAB.
ETHNICTYPEB S I
TYPECONTROLPUB
OF
PRI
HIGHESTLEVELOFRERING'2'4
OF-
4M
ORIENTATIONFALL1978ENROLL-HENT
TYPE OF HARDWAREACCESSIBLE BY STUD -ENTS AND FACULTY
DEC 10 (remote)PDP 11/70 (remote)PDP 11/34IBM 360 (remote)IMSAI 8080TRS-80
DEC 10HP 1000Micros
1891 x x x x x 5,395
5 1884 BIA x x 1,013 DEC 11/34IBM 1401
6 1867 x x x x 1,526 DEC PDP 11/34IBM 1130
7 1968 x TRIBAL
x x 839 DEC PDP 11/45
8 1881 .. x x x x 3,296 HPM2000/158370/158 (remote)
SOL (3, micros)SWTP (micro)
9 1966 x x x
I
x 4,315 IBM 3031 (remote)IBM 3033 (remote)AMDAHL 470 (remote)
NOTE: Campus-wide interviews were conducted at theinstitutions listed above as part of a needsassessment of educational computing in minor-ity post-secondary institutions.
ETHNIC TYPES: B - Black1S - Spanish speaking
I - American Indian1
HIGHEST LEVEL OF OFFERING: 2 - Two years
.- 4 - Four years4M - Master degree
252
7
240 NECC 1980
four terminals for administrative data en-try and inquiry. COBOL, RPG, ernd FORTRAPwere supported for students, as well asASSEMBLER (on the 9480 and the 1620).This UN/VAC is still used for studentbatch runs. In addition, ten ports wererented in 1975 on a companion communitycollege's HP 2000, an arrangement that wasterminated in 1979 in favor of access to aDEC 10 at a nearby private university.Students use ten DecWriter terminals whichare dedicated during labs for a course inBASIC and used for other courses duringopen times.
The CIS faculty purchased five micro-computers in Fall 1979 through a Title VIgrant. These have replaced the 1620 inteaching assembler languages and computingprinciples. A fourth system is under con-sideration now because of the two campuses,growth of institutional and faculty re-search, and growing concern for academiccomputing. The institution had done itsown local needs assessment before consid-ering such a step.
Thus, the Community College of Balti-more hoe steadily built its computing cap-abilities, if at a slower pace than somemajority institutions certainly also at afaster pace than most minority and/or two-year institutions. This stable growth hasresulted from support by the college ofincreases in computing capabilities throughits own funds, supplemented by aggressivelysought federal grants. Because of itscareful planning and budgetary practices,the school will be able to assume ancillarycosts when grants expire.
Factors influencing the expansion ofcomputing capabilities at this institutionhave been the local employment market, sizeof the student body (necessitating automa-tion in record handling and giving earlyexposure to administrators and faculty),support by administration, key faculty mem-bers with interest and dedication to sup-plement the budget for computing resourcea,and explicit planning mechanisms within thecollege charged with studying computingneeds.
Administrative support is evidenced byhiring practices, release-time practices(for proposal writing and planning), andactivities involved in acquiring computingequipment. A few key faculty nembershavevisited other successful sites, attendedconferences (ECMI(2) and others) and con-ventions, served on committees in profes-sional associations, sought external funds,and lent one another intellectual and
moral support. The evaluator who conductedthe interviews at this community collegecommented:
These individuals represent ascarce resource at any institutionand seem to represent a necessaryif not sufficient condition forprogress in academic computing.(A broad base of faculty awarenessis probably not a necessary condi-tion for groiEg in an institution'scomputing capabilities.)
Periodic ad hoc committees on computerutilization WEISii exemplify the planningmechanisms for computing established withinthe college. Plans also exist to establisha line position for a director of informa-tion systems reporting to the presidentthrough a dean of planning, development,and communications. This new positionwould assume responsibility for academic,as well as administrative computi:3.
Tensions have existed between adminis-trative and academic computing, as at manyOther institutions. The manager of comput-er services reports to the vice-presidentfor administrative services. The CIS De-partment, on the other hand, is within theDivision of Business, Secretarial, and Com-puter Sciences under the vice-president foracademic and student affairs. Supportstaff have grown through several stages tothe current manager of computer services,three programmers, two keypunchers, and twocomputer operators. No students are em-ployed since the manager claims "it doesn'twork."
Students submit batch jobs through aslot in a. door. Output pickups are sche-duled twice daily for fifteen minutes each.However, during the two lab hours assignedeach course in a batch-mode language,turnaround time is closer to immediate.Students, who probably know of 40 otheralternatives, accept the pickup arrange-ment, complaining only about the number ofkeypunches. Terminals are available forlearning BASIC, and the APPLE microcomput-ers were made available in Spring 1980 forlearning ASSEMBLER. The commercial optionin CIS will be offered at the campus near-est the business community and the scien-tific option at the other campus. Academ-ic applications exist also in the businessdepartment (managing data on patient careand student performance). CMI(3) is ex-pected to be developed in remedial readingand CAI(4) in science courses after the newconfiguration is installed. Faculty with agood track record in attracting federal
grants will be responsible for developingCAI at the science learning center. Theproject director for the science learningcenter had attended an ECMI conference andattributes his current grant, in part, tothat experience.
Institution #2
Bennett College is a private, women'sfour-year, historically black, liberalarts college with a "ILy, though stable,enrollment of about by°. Located inGreensboro, a mid-Atlantic city, the in-stitution houses its computing hardware inone attractive location on campus. Thishardware includes an INK IBM 1130 batch sys-tem used for administrative purposes andstudent programming coursea and an inter-actsve HP 2000 with fifteen terminals usedextensively for CAI tutorial and drill andpractice in mathematics, English, and biol-ogy. Two of four keypunches are availableto students.
Computer center hours are 8 a.m. to 10p.m. (Saturdays on request), and consulta-tion is available during these times.Freshmen receive computer ID numbers at thebeginning of the school year, but upper-class student* have open-shop, direct ac-cess to both computers. They run their owndecks and retrieve their own output immedi-ately; however, lines do develop when thebusiness office is running at the same timeon the 1130. Tables are provided studentsin the computer center for such use asexamining output and correcting programs.
Reports on the success of the CAI pro -j ect, in which faculty prepare courseware(or modify existing courseware) using IDF,have been given at CCUC(5) and ECMI(4)conferences; a great deal of consulting andsharing with other colleges has also takenplace. In addition, this college has beencited as an academic computing exemplar byHumRRO(6).
Computing began at Bennett College ele-ven years ago. A rented teletype connectedthe college to TUCC(7) and provided forsome instructional work, as well as for im-proving computer literacy of the faculty.Influential in establishing the initialcapability were the president of the col-lege (who later attended an ECMI confer-ence) and the mathematics department chair-person (who later became the computer cen-ter director). A year later, the collegepurchased the 1130 through a federal grantin connection with another university.This computer was used, from the beginning,
Minority InstitutionsECM1 241
for both administrative and academic pur-poses. It was also used cooperatively bytwo other small, church-related, primarilywhite institutions in the same city. Allthree schools used it for registrationprocessing. Five years ago a Title IIIgrant made possible the purchase of theHP 2000 with fifteen terminals for CAI de-velopment. Basic skills disciplines ac-counted for early CAI use; CAI is builtinto course requirements in these disci-plines.
The success here with CAI has givenrise to new problems. Students need twiceas many terminals, and the 1130 is toosmall and is no longer supported by IBM.Additional personnel trained in computingare needed. Release time for faculty isneeded to develop additional CAI course-ware. Despite the success, the support ofthe president, end-the small size of theinstitution, some faculty are still una-ware of the potential in their disciplines.In such a small school, this is seen as aproblem by faculty who do use the computer.
Attendance at ECMI conferences helpedto develop a strong team of faculty, how-ever, who have used instructional computingheavily. Througn a federal grant the in-stitution began to share its expertise inSpring 1980 by conducting a regional con-ference similar to the ECMI conferences.Bennett collaborated with North CarolinaA & T University (Institution #4) in con-ducting that conference. Thirty smallcolleges participated.
One faculty member cited CAI as ex-tremely helpful to entering freshmen, somany of whom are in need of remedial workdue to inadequate preparation in secondaryschools. Students said they appreciatedthe CAI but not the downtime. The presi-dent of the institution would like to seeseparation of administrative and academiccomputing. Plans for upgrading the exis-ting hardware are on the drawing board.
Despite its small size and its depen-dence on hardware and software that areless than stets-of-the-art by today'srapidly changing standards, Bennett Col-lege has taken a position of leadershipin the development of transportable course-ware for use in basic skills courses. Suc-cess has brought with it new problems, butit has also nurtured confidence, plans, anddetermination to solve the problems.
242 NECC 1980
Institution #3
The Rio Grande Campus of Tens StateTechnical Institute is a technical-voca-tional, two-year, Hispanic institution lo-cated in Harlingen, close to the Mexicanborder in the southernmost tip of thestate. One of four such institutions com-prising a system in Texas, this school wasestablished on a World War II air forcebase in 1967. Its fast-growing studentbody numbered well over 1,100 at the timeof the interviews, and some of the growthis in the data processing program.
Curriculum design is a high-priorityand on-going activity at the school, asevidenced by its special stuffing anJ theexistence of school-industry cooperativecommittees and advisory committees. An-nual evaluation ensures that curricula areup to date and graduates are well preparedfor employment. (Some students commandentry-level salaries as high as $18,500without graduating.) Just minutes awayfrom this institution are vast farmlandson which thousands of Mexican Americansbarely survive. Thus the existence of anindustrial corridor and this institutionto serve it is pivotal to economic changein the area. (On the first day of theinterviews here the Wall Street Journalstated that this region was one of thefour fastest - 'rowing industrial areas inthe country.)
A new building houses the computingand electronics programs and computingfacilities, as well as some other mathe-matics and science or technical programs.Nearly one-third of ita 17,000 squarefeet is occupied by computing facilitiesand classrooms used in the industrialdata processing program (IDP). Prior tothe interviews (just two weeks after themove to the new building) less than afourth as much apace had been available.
IDP students dominate the facility,which, like the entire campus, is out-atandingly clean and well organized.Terminals are dedicated to students, orto entire classes, from 8 a.m. to 5 p.m.ften the facility is open till 6 be-use the IDP chairman stays late. Since
. t students are Mexican-American under-gr.Iuates who live with their families inthe area and are not accustomed to beingaway in the evenings, the facility is notkept to at night. A demand does exist,however, from working adults in the areafor an evening program. Finances andstaffing are the obstacles to be over-come.
Two instructors are employed in the IDPprogram, one the chairman. An additionalslot is open but not filled. The two in-structors bear heavy teaching and labloads, with most emphasis on tab work.They share the burden of supervising thefacility. Second-year students are trainednot only to program (and maintain programs)for some administrative and academic appli-cations, but also to assist as consultantsto first-year students when instructors areunavailable.
The /DP chairman has spent much of hisown time on outreach to other departments,providing demonstration projects, seminars,and classes. The degree of interest fromother departments likely to use the comput-er for instructional purposes ranges allthe way from "take it away" to "you can'tbegin to meet MY needs." Those least in-terested are traditional inatructors. Inprograms T.% individualized instruction(or in u instructors have had previousexperien-e), faculty are eager to attemptcomputer-managed instruction or computer-aaaisted test generation. Attitudes appearto stem from educational philosophiesrather than from familiarity with areas ofexpertise, such as nuclear technology ormathematics.
Two weeks after the move to the newbuilding, when the interviews were conduc-ted, a terminal room contained ten CRT's,two teletypes, and a line printer in fairlyconstant use. A Radio Shack TRS-80 and an/NSA' 8080 microcomputer were also avail-able. Six additional CRT's, a printingterminal, a digital plotter, and a tapedrive were also being readied for use.Three keypunches with acoustical shellswere available. Bulletin board displaysincluded charts, lists, schedules, comput-er-generated Mona ',isms, and a sign pro-claiming "The Dirty Dozen." ("The DirtyDozen," it turned out, were the second-year students who had survived out of amuch larger original field of beginnersin the IDP program.)
The terminals were on-line to a DECFDP 11/70 located in a department store45 miles away. COBOL& not the latest(1969), and RPG card (Mks are sent to aDEC 10 at a university 50 miles away, witha turnaround time of weeks. But studentsare obviously learning and being hired.The IDP department, having grown from fourstudents in 1974 to more than seventy atvarious levels of advancement at the timeof the interview, has gone through severalupgrades of remote connections. Currentplane call for an on-aite computer with 32
2 S6
ports for student use. However, instruc-tors who want to support individualized in-struction, record keeping, and test genera-tion in open-entry/open-exit courses feelthey may still not have sufficient capacitysince they would be in contention with IDPfor use of the resources. Meanwhile, ad-ministrators are working on additional ca-pacity for administrative work (remote ac-cess to a large state institution's network).
One instructor has done some work of hisown, some oz which was destroyed in one ofthe upgrades. Most recently he has beenusing a TRS -80 microcomputer on his owntime. He envisions a cluster of micros ina classroom, which he sees as a cost-effec-tive solution to his problems in a tradi-tion-oriented department. The electronicsprogram, which trains almost a hundredstudents for customer engineering on DECequipment dedicated to the program, canuse twice as much hardware. The chairmanof this program, a former ECMI(2) parti-cipant, keeps up with hardware period-icals and literature but is far too busyto move into academic computing generally.
Students who are beginners tend to seeno problems with existing hardware andsoftware, but the advanced "Dirty Dozen"talk like data processing managers. Com-pletely at ease in the jargon and what'sbehind it, they speak knowledgeably aboutthe shoreconings of the available soft-ware, the need for this version of a lan-guage and that many ports, and even theadditional justification needed in thecurrent proposal for new equipment.
Administrators give moral support, buttheir budget requests have to move throughseveral layers of state bureaucracy andcompete with other technical institutions.Even when funds exist, it is difficult tofind qualified staff in the area who srenot already working for burgeoning indus-tries at high salaries. Nevertheless, theindustrial data processing program grows,and other departments are beginning tovoice their needs. Among two-year tech-nical institutions, TSTI-Rio Grande may beon its way to becoming an academic comput-ing exemplar.
Institution #4
North Carolina A 61T University is astate-controlled, four-year, master degree-granting, black institution, with liberalarts and some engineering emphasis. Theenrollment is 5,400 students, Locatea 4CZ1188 town from Bennett College (Institu-
Minority InstitutionsECMI 243
tion #2), this school was the last (andonly minority) institution to procure amainframe computer through the NationalScience Foundation's original Office ofComputing Activities. That computer wasa CDC 3300, and it followed the first com-puter acquired in 1964, an IBM 1620. Itwas replaced by a DEC 10 in 1977. Federaland state funding was combined in each ofthe latter two cases, with the CDC beingsold to add to the funding for the DEC.
Although some impetus came from thecomputer center (especially recently), akey role was played by former dean of artsand sciences Arthur Jackson, after whomthe computer center is now named. Morethan 80 terminals are on-line to the DEC10, 26 of them in the computer center. Avariety of languages are available: BASIC,PASCAL, APL, COBOL, ALGOL, LISP, and oth-ers. The computer is available around theclock all week. The staff includes nineprogrammer/analysts, and experts areavailable to faculty and students asneeded. Students access the computer in-teractively through class accounts or in-dividually (both authorized by departments)as well as by batch jobs. A monthly reportprovides usage information but is not cur-rently used for charging. Most use is bystudents in coursework, but faculty do re-search, and administrative applicationsabound. Turnaround time is good except atpeak times. Jobs are limited to fifty ata time, which can cause a connection delayof ten-to-fifteen minutes. Work space inthe computer center for students is alsolimited.
During the past two years North Caro-lina A & T University has experienced whatthe computer center director refers to as"a quiet revolution" in computing. From asingle-job, batch-mode machine with 49Kmain memory and 25 megabytes of disk, theuniversity took a quantum leap to 256K,600 megabytes of high-speed disk and theability to process up to fifty jobs at atime in many languages interactively, inbatch mode, or in combination.
However, increased demand for computertime has lowered response from excellentto poor. Engineering, which six monthsbefore the interviews could run circuitanalysis, finite element, operations re-search, and other sophisticated jobs atany tine,'must now normally execute manyof these between 5 p.m. and midnight, andseveral between midnight and 8 a,m. Engi-neering school jobs also include CAI pac-kages needed for educationally disadvan-taged students, and it is difficult to
256
244 NECC 1980
find time during the day to run them.Plana are afoot for a memory upgrade at8100,000, half of which must be raisedfrom outside the university budget. Fur-ther upgrades will probably include a re-mote-site laboratory and a processor up-grade. Staff and space needs were citedin addition to hardware and stations forremote access.
An active computing advisory committee,chaired by a chemistry professor, func-tions as an agent of change and supple-ments positive pressurf.s from all types ofusers and administrators. Proposals havebeen written, and a computer science deg-ree program is to be launched in Fall 1980.An academic computing director and an ad-ministrative computing director will alsobe hired.
Some usage came about through partici-pation of faculty at ECMI conferences(2),where they learned about MINITAB, test as-sembly and scoring, test item banking, andCAI, all of which are now used. Althoughmost usage is by the engineering school,other departments (especially mathematics,chemistry, and business) are becoming ac-tive users. The list has expanded eachmonth in the last two years. Electricalengineering has also developed severalmicrocomputers for instructional use andan HP 1000 that is used for instructionalpurposes and research in upper-level cour-ses.
Instrumental in the growth of computingin engineering at this university has beenan engineering accrediting agency, accord-ing to some faculty. Employment of facultyby local industry was a factor in initiallycreating awareness of computer potentialamong engineering faculty. ECMI conferenceattendance is credited with increasingawareness among mathematicft, chemistry, andbusiness faculty.
Political problems associated with fed -eral. efforts at encouraging integrationhave caused difficulties in maintainingorderly development of academic computing,according to some, Also cited as a nega-tive factor was the pattern of federalfunding which historically favored main-stream institution and left out minorityinstitutions. This is one of the few min-ority institutions, however, that managedto begin to beat the system as early asthe '60s.
Conclusion
In the cases of the institutions des:-cribed above, though a variety of approach-es have been made toward academic comput-ing, certain factors surface as commoningredients of success. These includecampus-wide planning (or at least plan-ning beyond the walls of a single depart-ment or class), dedication on the part ofkey faculty or administrators, carefulbudgeting practices, the ability to puttogether funding from various sources, theability to learn by experience (as well asby capitalizing on the experience of oth-ers), and the will to get maximum mileagefrom the resources at hand. Interesting-ly, historical factors that could have de-feated some actually seem to have causedthese institutions to try harder.
References
(1) Needs/Strategy Evaluation of MinorityInstitutions in Educational Computing(NSF Grant No. SPI-7821515, in prog-ress)
(2) ECM (Educational Computing in Minor-ity Institutions): three workingconferences end one workshop forminority institution faculty in 1975-1977 to assist faculty with little orno knowledge of educational computingto learn about it in their own disci-plines. Courseware was developed bya small group at one summer workshop.
(3) CMI: computer-managed instruction
(4) CAI: computer-assisted instruction
(5) CCUC: Conference on Computing in theUndergraduate Curricula (annual,1970-1978)
(6) HumRRO: Human Resources Research Or-ganization
(7) TUCC: Triangle Universities Comput-ing Center
Minority InstitutionsECMI 245
COMPUTER USE IN CHEMISTRY AT A MINORITY INSTITUTION
James D. BeckDepartment of ChemistryVirginia State UniversityPetersburg, Virginia 23803
(804) 520-5481
INTRODUCTIONMany students in minority institutions
encounter severe difficulties in introduc-tory science courses. These difficultiesoften prevent minority students from ob-taining degrees in scientific fields andthus effectively shut them out of careersin the sciences, engineering, and thehealth sciences (1). These students areoften classified as "underprepared," aclassification which implies that theirbackgrounds, in science and mathematicsespecially, are not strong enough to en-able them to succeed in rigorous, quan-titatively oriented college courses.
A number of characteristics of theseunderprepared students have been identi-fied (1,2). These range from poor math-ematics background and lack of exposureto science in high school to poor studyhabits and a lackyof self-confidence.Poor reading ability and a general weak-ness in communication skills appear tobe significant factors which inhibitsuccess fcr these students. Many ofthese students have low personal stan-dards for academic achievement, expectfailure, and fear science courses. Al-though numerous approaches aave beentaken to improve the performance ofunderprepared students in Aciencecourses, limited use has bo4n made ofcomputer-based learning modes.
Virginia State University is afour-year stated supported institutionlocated in Petersburg, Virginia. About3400 full -time students are enrolled,with over 95 percent of the undergrad-uate population being black. Readingtesta given to incoming freshmen haveshown that most of them have seriousreading problems, with nearly 70 percentof them reading below the ninth gradelevel. Median scores on the ScholasticAptitude Test were ;bout 300 on theverbal part and about 350 on the quan-
titative part for incoming freshmen.Many incoming students have had limitedexposure to science in high school. Asan example, 16 percent of the students ina general chemistry class this past yearhad not had any chemistry in high school.
The problems of inadequate prepara-tion are compounded by the student diver-sity which is encountered in many courses.Incoming freshman chemistry majors, forexample, had reading scores ranging from9.0 to 15+, SAT verbal scores rangingfrom 240 to 480, and SAT mathematicsscores ranging from 280 to 580. Thisrange of abilities, backgrounds, and in-terests presents a tremendous challengeto an instructor.
INSTRUCTIONAL APPROACHFor several years we have been exper-
imenting with various approaches to theteaching-learning process in trying tomeet the varying needs of our studentswho enroll in chemistry courses. Most ofour efforts have been centered on ourgeneral chemistry course. This is thechemistry course with the largest enroll-ment and the one in which the problems ofdiversity and weak backgrounds are mostpronounced. The approach we are currentlyusing employe the computer in severalways (3), and we are exploring new and ex-panded types of computer use. In general,we try to make available to our studentsa wide spectrum of different ways tolearn chemistry in the hope that some ofthese will meet the needs of all of ourstudents. The variety of materials andmethods allows students to select learn-ing modes which suit them while permit-ting students who are not well-preparedto engage in extra activities to helpthem master the material.
Our approach has been to retain thetraditional components of our generalChemistry course--textbook, lectures,
25s
246 NECC 1980
recitation, and laboratory sessions. Tothese we have added a selection of supple-mentary materials and activities, includ-ing a study guide, sets of performance ob-jectives, lecture otes, slide/tape andfilmstrip/tape programs, voluntary practicesessions, copies of old examinations, andsuggested plans of study. The computerhas also played an important role in thissmorgasbord of learning activities. Aboutfifty interactive computer programs areavailable for student use; most of theseare drill-and-practice or simple tutorialprograms. A few are simulations. Thecomputer has also been used to generateindividualized problem sets for studentuse. These are used off-line and the an-swers may be submitted for grading by theinstructor.
Students are free to select which ofthese alternative learning strategies theywish to employ. No specific activitiesare required, but students are expectedto complete a certain number of activitiesduring the semester. Students are encour-aged to do more than the minimum number,especially if they exhibit weaknesses inidentifiable areas.
We have evaluated the effectivenessof these learning activities in two ways.One involved a simple frequency of usecompilation in which we counted the num-ber of times that students selected aparticular type of activity. The otherevaluation method used student surveys toascertain the students' perceptions of therelative usefulness of the activities.Both studies produced similar results.The interactive computer programs wereused heavily and were perceived by thestudents as being very useful to themin their efforts to learn chemistry.The computer-generated problem sets wereused by nearly all students and were alsoperceived as being very useful. Thevoluntary practice sessions rankedslightly below the computer programs andproblem sets in extent of use and inperceived usefulness. The slide/tapeand filmstrip/tape programs were ratedstill lower.
It is interesting that every activ-ity was used by some students and thatevery activity was rated by some to be"very useful." It is also interestingthat while all of the non-traditionalitems that were described above wereconsidered, on the average, to be"quite useful" to "very useful," thehighest student ratings were given tothe lecture notes and to the lectures.However, the textbook, study guide,recitation classes, suggested plans ofstudy, and sets of objectives were all
considered to be less useful than the com-puter programs, problem sets, voluntarypractice sessions, and media programs.The copies of old examinations were rankednearlyeven with the practice sessions.Students nearly all desired to continuehaving a variety of alternative methodsand materials available.
COMPUTER USEThe instructional approach has been
described in some detail to emphasize theintegral part played by the computer. Al-though the computer is not an essentialpart of the instructional process,it isa valuable component. The computer cando some things better than others, but itcannot do all things well. It should belooked on as one of many instructionalmodes available to the teacher, appropri-ate for some students studying some typesof material in some situations. When thecomputer is made a part of the instruc-tional package, its use should be care-fully planned with consideration given tothe learning objectives that are involved.
The use of the computer in this pro-ject is not new or unique. However, thesetypes of use can be duplicated at nearlyany institution, even a small college, aminority institution, or a high schoolwhere extensive computer facilities maynot be available. The methods also offerthe potential of expansion to more elab-orate and sophisticated computer applica-tions.
Most of the interactive BASIC programsthat we have used were written at VirginiaState University and tailored to the needsof our students. While simple in conceptand in construction, the programs haveproven to be quite useful in this setting.They have been designed to be short andsingle-concept in nature, each one re-quiring that the student spend only a fewminutes at the terminal. Most are open-_ended, allowing students to obtain addi-tional practice as they feel that theyneed it. In general, the programs coverbasic topics which must be mastered inorder to solve more complicated problemsand understand more complex relationships.A few programs are simulations of chemicalsituations.
These interactive programs do not usegraphics or special effects and generallyuse the most common BASIC instructions.Thus, they will run on nearly every com-puter system and can be used with eitherCRT or hard-copy terminal. Some of themhave been adapted for use on several ofthe popular microcomputers (4). Althoughthe programs may not be suitable for usein all environments, they are simple
227
enough that they can easily be modified.The generation of individualized
problem sets has also been done in a verysimple manner. Two different approacheshave been tried here. In one we have gen-erated the problem sets in batch-mode,using a FORTRAN program which merely se-lects questions at random from a questionbank. Any system could use this methodsince a similar program could be writtenin any language that has a random numberfunction. In the other method of genera-tion, individual BASIC programs have beenwritten for each problem set. In thesethe random number function is used to se-lect individual questions and to generateindividual data within questions. Withthis method, we can run the problem setsin batch mode or have each student ob-tain a copy from a terminal. Hard-copycapability is obviously required for thismode of generation.
EXTENSIONSWhile the two modes of computer use
that I have described are relatively sim-ple, they can be adapted to fit a varietyof instructional settings, they do notrequire special hardware, and they havebeen proven effective. There is, however,a great potential for developing more so-phisticated, and perhaps more useful modesof computer use, building upon these sim-ple applications. We have explored someinteresting extensions at Virginia StateUniversity and plan to try out more newthings in the future.
Prior to September 1979, our inter-active programs were limited somewhat byour terminals--slow and noisy IBM 2741hard-copy units. Installing IBM 3278 CPTunits and 3287 printers has increased ourcapabilities tremendously. By reducingthe time required to run each program, wehave been able to increase the number ofprograms available and allow students tospend more time in remedial or advancedwork. Although printers are available,students do not generally need to gethard-copy for their runs, since their re-sults are put into a file and can be re-trieved by the instructor for grading.
Although our new terminals do nothave true graphics capability, their rap-id print speed has enabled us to programmaterial which requires more extensivelayouts than we were able to use pre-viously. This equipment has expanded therealm of programming possibilities tosome areas that we could not consider be-fore the change in hardware. We wouldlike to have the capability to use realgraphics; this would undoubtedly enhancelearning even more (5,6).
Minority InstitutionsECM 247
Although we have been using computer-generated individualized problem sets forseveral years, we have just begun to ex-plore the possibilities involved withusing them. For example, the programthat is used to generate the problemsets can also be used to generate indi-vidualized repeatable exams or quizzes.Computer - grading, either on-line or off-line, is'a possibility.
Problem sets offer some significantadvantages over interactive programs.They do not require extensive connecttime, since the actual work is done bythe students off-line. They permit theinclusion of topics and problems whichare not appropriate for interactive pro-grams. For example, multi-step problemsthat require extensive calculations canbe included on the problem sets. Ques-tions that require the use of the text-book or of other sources of informationare suitable. In fact, there are almostno limitations on the types of questionswhich can be included on a computer-generated problem set. If grading isnot a problem, even essay questions arepossible.
One exciting potential use of com-puter-generated problem sets is for diag-nosis and prescription which we havetried on a limited basis. After the stu-dent has completed the problem set, aninteractive program is accessed on a ter-minal. This program checks some or allof the student responses. At this point,some of the advantages of an interactivemode can be realized. Immediate feedbackcan tell the student which answers arecorrect, and the correct answers or meth-ods of calculation can be presented. Acourse of action can be prescribed tohelp the student overcome the identifiedweaknesses. Prescribed activities mayinclude reading the textbook, doing ad-ditional problems, running interactivecomputer programs, viewing slide/tapeprograms, or seeing the instructor.These prescriptions are individualizedand are based on the student's perfor-mance on the problem set. In addition toproviding useful information to the stu-dent, the results can be made availableto the instructor, enabling areas ofweakness to be identified, for individualstudents and for the class as a whole.This capability is similar to the TIPSprogram (7).
The problem sets might also serve asa core for a system of computer-managedinstruction. We have been exploringsome different ways to use them and havebeen encouraged by the results. Our stu-dents have responded positively to our
26o
248 NECC 1980
primitive attempts at diagnosing problemsand prescribing remedies, and we hope toenlarge this effort in the near future.
Several years of working with com-puters in chemistry at a minority insti-tution have convinced me that many stu-dents, underprepared ones in particulur,can benefit from computer-based instruc-tion. I believe that this is possiblewithout elaborLte computer hardware, ex-tensive programming experience, or a hugeinvestment of time and m^ney. AlfredBork has stated that there is no oneright way to use - computers xn education,or even a most profitable way (8). Cer-tainly there is no one best way to usecomputers in science instruction or withunderprepared students or at a minorityinstitution. But there are many mean-ingful and wortl-while ways in which com-puters can be used to enhance learningfor all students (9,10,11). Instructorsat minority institutions should not con-sider computer-based instruction to bebeyond their c. Abilities. The compu-ter is an important instructional toolthat needs to be included in planningpresent and future educational deliverysystems.
REFERENCES
1. McDermott, Lillian C., Iiternick,Leonie K., and Roseng.tist, Mark L.,"Helping Minority Students Succeed inScience," Journal of College ScienceTeachin , 9, 135 (1980).
2. Ketn k, Louis J., "Teething Scienceto the Disadvantaged Student in anUrban Commbnity College," Journal ofChemical Education, 50, 46,71-973177.
3. Beck, James D., "Using i.L4 Computerin the Teaching of Science," Pro-ceedings of the Minority InstTEUrionsCurriculum Exchange Conference, Wash-ington, D.C., p. 37 (1979).
4. Several programs are available fromPrograms for Learning, Inc., P. 0.Box 954, New Milford, CT 06776.
5. Bork, Alfred, "Computer Graphics inLearning," Journal of College ScienceTeaching, 9, 141 (1980).
6. Soitzberg, Leonard J., "ComputerGraphics for Chemical Education,"Journal of Chemical Education, 56,644 (1979).
7. Shakhashiri, Bosom Z., "CHEM TIPS -Individualized Instruction in Undergraduate Chemistry Courses," Journalof Chemical Education, S2, 581------11371.477-
8. Bo; *4-ee "Computers and theFu AA .ucation," Computer -BasedS. In.w Istruction, Andre Jones and
Harold Weinstock, editors, NATO Ad-vanced Study Instatute, Leyden, Neth-erlands, p. 21 (1977).
10. Cauchon, Pauf,-Wimistry with a Com-puter, Programs for Learning, Inc.,P.O. Box 954, New Milford, CT (1976).
11. Moore, John, Gerhold, George, Breneman, G.L., Owen, G. Scott, Butler,William, Smith, Stanley G., and Lyn -drup, Mark L., "Computer-Aided Instruc-tion with Microcomputers," Journal ofChemical Education, 56, 776 (1979).
Minority institutionoECM1 249
EDUCATIONAL USE OP COMPUTERS IN PUERTO RICO
Prank D. AngerDepartment of MathematicsUniversity of Puerto Rico
Rio Piedras, Puerto Rico 00931(OM 764-0000
ABSTRACTPuerto Rico presents a unique blend of
problems and promise in educational com-puting. Both the lack of an adequatetechnological base in the community andrestricted capital hold back the develop-ment and implementation of computerizedinstruction in all forms. Public schools,facing many of the same problems of main-land inner-city schools, are totally unableeven to start in this area. Nonetheless,since about 1974 there has been a rapidgrowth of instruction with and aboutcomputers in the universities, colleges,and private and parochial secondaryschools. This cowth promises to increaseradically during the eighties.
The educational problems to which thisdevelopment responds, the kinds of computerstrategies attempted, some interestingresults, and projections for the eightieswill be discussed.
262
Computer Laboratories in Education
MICROCOMPUTERS IN THETEACHING LAB*
Dr. Robert F. Tinker,Director, TechnicalEducation ResearchCenters, 0 Eliot St.,Cambridge, MA 02139:617) 547-3090
AN OVERLIOXED AREAWhile computers have many uses in sci-
ence education, their use in instrument-ation has been largely overlooked. Thispaper will show the many powerful thingsthat can be done with a properly inter-faced computer to gather, analyze, andpresent real laboratory data.
The major reason that this educationaltool has not been developed in the sameway as other educational applications ofcomputers relates to hardware. Twentyyears of experiments with educationalapplications of computers have almostentirely involved large time-shared main-frame computers, which are not welladapted to the timing requirements im-posed by real laboratory measurements.A time-shared computer determines whendata is going to be read from the term-inal, and so communications with it aredetermined by the computer's timing re-quirements. While that condition mayusually be adequate for human inter-actions with a terminal, it is not 'ade-quats for most laboratory interactions.In lab applications, the timing require-
*Some of the materials incorporated.inthis work were developed with the fin-ancial support of the National ScienceFoundation Grant Nos.7711116 and SED79-06101.
250
meets are often too stringent for simpletime-sharing systems. It is possible touse tine-sharing for laboratory measure-ments, but it is considerably more complexand expensive than using terminals fortime-sharing.
The arrival of microcomputers on thescene has completely changed the hardwaresituation. Many laboratory applicationsof computers require very little computa-tion and place only minimal requirementsop the computer. As a result, very simpleand inexpensive microcomputers can besuccessfully used in the laboratory. Nowthat the hardware question is being solved,it is time to begin a vigorous effort tomake up for lost time, to begin using com-puters in the lab, to develop relatedhardware and software, and to research theeducational implications of this tool.
SAMPLE APPLICATIONSOur group at TERC began experimenting
three years ago with applicatioac ofmicrocomputers in the laboratory becausewe were writing course material on instru-mentation as part of our modular electron-ics project, which is designed to prepare.tmdents to use instrumentation. It wasclear from the beginning that no suchcourse would be complete without a fairlythdrough coverage of microcomputer appli-
cations in instrumentatim.research labsthat are not already heavily dependent oncomputers for gathering and analyzingdata will certainly become so by the timethe current generation of students are onthe job for any length of time.
The Coolina Curve ExperimentHistorically, the first application we
developed used the KIM computer to recordtemperature of a liquid sample of naphtha-lene as it cooled through the crystalize-tiop temperature. This is the coolingcurve experiment familiar to those whohave taught introductory physical science.There is some nice physics in it because,as the sample loses heat to the surround-ing water or air, the temperature does notdrop uniformly. At the transition temper-ature, heat flows out but the temperaturedoes not drop, giving rise to a plateau-like graph of temperature as a function oftime. This shows very clearly somethiugthat many beginning students do'not appre-ciate! namely, that there is a distinc-tion between heat and temperature; thatat the solidification temperature, heatis flowing out without a temperaturechange.
There is some good science in thisexperiment. First of all, of course, theplateau temperature is a good way of de-termining the melting point of a sample.In addition, mixtures of napthalene andparadichlorobenzene (moth flakes and moth-balls) display some interesting meltingphenomena. In some ratios, the plateaudisappears, as it will for most mixtures,but at one particular ratio -- called theeutectic -- a plateau reappears at a temp-erature that is below the melting tempera-ture of either of the pure substances.
All these phenomena are accessible tosomeone with as little instrumentation asa mercury thermometer and a stopwatch.However, gathering and plotting the datacan be a tedious and boring task, parti-cularly when it is necessary to repeatthe experiment several times for differentmixtures. A typical cooling curve runtakes 20 to 30 minutes. This run can bespeeded up by using a Koller sample, butthen many fewer points are obtained andmany of the important features of thecooling curve are lost.
The microcomputer solves this problementirely. We developed # very simpleinterface that can be ad to record thetemperature of a sample at any ratedesired and to display on a simple oscill-oscope a graph of the resulting tempera-ture versus time. The microcomputer canlog in at a rate of one per second to gen-erate an apparently continuous graph,
Computer Laboratories in Education 251
which can be observed as it evolves. Thetemperature detector is an inexpensivesignal diode. Because the transducer issmall, a small sample can be used, andthe experiment can be speeded up so thatit only takes a few minutes. As a result,it is feasible fat students to obtain dataon mmy different samples and focus theirattention on the phenomena, rather thanon the boring details of gathering anddisplaying the results.
A Pot ourri of ApplicationsSStce this initial venture into micro-computer applications in the laboratory,we ham developed over a dozen other labapplications for simple microcomputers.These ire listed below with some of themore general applications described first,followed by some programs tailored to thespecific needs of certain laboratoryexper4ments:
Counter/Timer. In this program, thecomputer displays counts or elapsedtime. It has all the functions of anormal counter/time but with a lowerfrequency response. In addition, itcan record for later recovery the num-ber of times that multiple events hap-pened after an initial trigger period.IC Testin . ,This convenient programearns the logic expected !too an in-tegrated circuit (IC) by runningthrough the permutations of a func-tioning IC and than comparing this toquestionable ones.Function Generator. Using an analogoutput, this program generates a var-iety of functions that one would onlyexpect from a very sophisticated func-tion generator. Ten different func-tions, incliding white, pink, and bluenoise, are available with selectableamplitude, offset, and frequency up toa bandwidth of 20 kHz.Transient Recorder. This programtriggers when the transient starts anddisplays the result on a standardoscilloscope, thus effectively con-verting it to a storage-type oscillo-scope. Up to 256 samples of the inputsignal can be recorded at intervalsas fast as 20 microseconds. In effect,the cooling curve experiment is aspecial adaptation of this facility.Fourier Synthesis. This program canrequire: the amplitude and phase ofup to 30 terms in a Fourier synthesis.The resulting output waveform is dis-played On a common oscilloscope andcan be beard with the help of a simplepower amplifier.Fourier Analysis. This program usesthe idea behind the transient recorder
26,1
252 NECC 1960
to capture an input waveform, which isthen frequency analyzed. Using a FastFourier Transform algorith, the com-puter displays the power content of thefirst 128 frequencies in the capturedsignal.Radioactive Half-Life Experiment.Here, pu &Mt from a Geiger Tube arecounted over time, and the resultinghalf-life decay function is displayedon an oscilloscope.Pulse Height Analysis. This programuses a special interface to capturethe height of pulses from a standardphoto-multiplier particle detector.The resulting pulse height spectrum isthen displayed on an oscilloscope. Withless than a dozen integrated circuitsin the interface, a standard microcom-puter can be used to replace special-purpose PHA instruments costing tentimes as mach.The Computer of Average Transients. Wehave developed a geological soundingexperiment that requries careful pro-cessing of signals from a geophone thatdetects seismic waves generated by hit-ting the ground. The common techniqueis to use a compliter of average trans-ients, which the KIM simulates, using asimple program.Solar Collector Analysis. Temp4raturefrom a number of sensors distributedaround a solar collector, as well as alight detector measuring the inputlight levels, is simultaneously loggedand selectively displayed through useof this program.Rotational Dynamics. This program,written by Ken Flowers, measures theangular acceleration of a disk commonlycalled for in lab experiments involvingrotational dynamics.Linear Dynamics. An electronic versionof spark tape uses clear tape withblack lines. When attached to an objectand pulled through a special detector,the computer can display position,speed, and acceleration instantly.
EDUCATIONAL IMPLICATIONSThese examples illustrate that it is
technically and economically possible tohave students use the computer asa gen-eral-purpose laboratory instrument: thebig question in many people's minds iswhether such use is desirable. I willbegin addressing this point by firstdeveloping the skeptics' argument morefully. There are essentially three mainarguments against the use of this kind ofsophisticated instrumentation. First i*that it is too complex and difficult touse: secondly, it makes the laboratory
mysterious, thus weakening the connectionbetween reality and result; and third, itisn't really necessary at all. These ar-guments are taken up in order below.
ComplexityMicrocomputers introduce a whole new
level of complexity that requires master-ing sophisticated hardware, software'and operating concepts. In order to usecomputers effectively, the students willhave to investigate all these new ideasand will get sidetracked, so that anypossible advantages that the computer canbring will be offset by the enormous in-vestment of time and intellectual effortrequired.
This would certainly be a compellingargument if it were true, if students wererequired essentially to develop all thehardware and software needed in the labor-atory.. However, this is an inappropriateway to use a computer in the laboratory.The desirable approach is to use previous-ly developed canned.programs that arecarefully designed to require the minimumamount of additional knowledge. Lab pro-grams should operate just the way elec-tronic pinball machines do; anybody with aquarter can walk off the street and beginusing those microprocesmor-based machines.There is no reason that laboratory instru-ments using microprocessors should be anymore difficult to use. Admittedly, someof our first efforts in this area do notmeet this requirement. But there is nobasic reason that prevents microcomputersfrom being easy to use in an applicationarea. It is simply a question of goodprogramming; providing menus for selec-tion and help files to clarify any diffi-cult points.
MystificationThe skeptic would hold that an instru-
ment that instantly gives result is un-desirable when it is not clear to studentshow those results are obtained. If thereis no alternative way to cross-check the-,,cults, if the experiment must stop whenthe apparatus stops.fuectioning, and ifthe computations are too difficult to dis-cuss, then the results may as well bemagic.
Since the results appear to be magi-cal, the student loses any chance of hav-ing any intuitive understanding of themeaning of the results and connection be-tween those results and the physical phen-omena under investigation. For instance,you pull a strip of marked paper through alittle gadget, and the computer prints outsome acceleration data. The connection
between the way you pull the tape and theacceleration is incomprehensible and un-verifiable, and therefore, the students'understanding of the term acceleration isnot enhanced by this experiment. With aspark tape, the student can see the markson the tape, understand the relationshipbetween those and the motion, and tracethe calculation of the accelerationthrough a series of simple calculationsbased on the marks. Furthermore, the cal-culations are based directly on the defini-tions of velocity and acceleration, andtherefore, the student gets practice inapplying those definitions by performingthe calculations in the laboratory. allawareness of these relationships are lostwhen the computer automatically calculatesthe accelerations and graphs them.,An important part of laboratory exper-
iences for students is their enablingfunction. That is by learning how toframe and answer questions, students arethen better able to function as scientistsand as citizens. Understanding is the keyto this enabling function, and the inclu-sion of a major thing, the microcomputer,which is inherently incomprehensible, isdisabling rather than enabling.
These are important arguments and pointup some of the pitfalls of using the com-puter in the laboratory. Like any newtool, the microcomputer can be misused.However, the statement of the problempoints to the solution. Most people canlearn to use mysterious or incomprehen-sible devices -- so long as they have aclear idea of cause and effect. Manyaspects of diving a car are essentiallymysterious to drivers: the details of theconnection between the wheel and the dir-ection of motion, between turning the keyand the operation of the motor, betweenthe position of the accelerator and speed.The same arguments can be applied to mostof the technologies in our environmentstelevision, radio, calculators, tele-phones, and aspirin. People use them allmore or less effectively without any de-tailed understanding of their mechanisms.Through practice people have learned theconnection between inputs and outputs.They have learned what to expect whenthe accelerator is pressed, when therecord button on a tape recorder ispunched, whena telephone is dialed. Thereproducible and predictable nature ofthese technological aids, which can beunderstood at an intuitive level throughpractice, completely removes the mysterysurrounding their operation and obviate*the need for a detailed understanding ofhow they work. This process of gainingan intuitive appreciation of a connection
Computer Laboratodes in Education 253
of inputs and outputs I call "intuitioncalibration" because it is directly anal-ogous to the calibration one performs oninstruments.
The thesis, then,is that if a micro-computer is used to measure acceleration,this will appear to be mysterious onlyuntil the students' intuition is calibra-ted. Now is this done? The direct ap-roach is to give students this tool andlet them play around with it in a directedmanner in order to get a good intuitivefeel for what it is producing. In testswith students, we have found that ittakes very little time using this tool toappreciate, on an intuitive level, therelationship between the acceleration andthe change of speed, namely, that rapidchanges in speed results in large accel-eration and that slowing down is justanother form of acceleration with a neg-ative sign. These intuitive explorationsdemystify the tool, but more important,they foster an intuitive feeling that noother approach can create. Students comeaway with more than understanding of anisolated term, such as acceleration. Theytalk about acceleration in intuitive termsthat are directly related to its defini-tion. They see that it is position-indep-endent, that it only depends on how quick-ly speed is changinl. Thus, the measure-ment and the definition come together inthe laboratory without the use of compu-tation.
The one argument that can't be count-ered is that students do not gat as muchpractice performing computations in alaboratory where the computer does thecomputation. The obvious cure for thisis to assign some additional computationor to do the experiment once without thecomputer.
1 tt---TEirikepticts argument hers is that thecomputer, while attractive, is not reallynecessary in the laboratory. However,the microcomputer makes a number of thingspossible that could not otherwise be doneconvenientli or economically, and there-fore, unless cne takes the fan.estic pos-ition that there is no Seed to improvescience instruction, one cannot ignorethese rather substantial improvements.
The use of a general-purpose laboratorycomputer broadens the kinds of experimentsthat can be contemplated in a teachinglaboratory and also increases the rate atwhich data can be gathered and analyzed.For example, the dynamics measurementspermit a detailed analysis of the Connec-tic..1 between force and acceleration thatin simply tot possible through any other
2 c
.1MM.
254 NECC 1980
mechanism. This application alone opensup new and as yet, uncharted possibili-ties in the introductory laboratory. Sim-ilarly, the Fourier Analyzer, the PulseHeight Analyzer, and even the Computer/Timer open up possibilities that wouldotherwise not be considered in elementarylaboratories because of cost considera-tions.
The speed with which experiments canbe analyzed is often an important educa-tional factor. Instead of a simple fall-ing ball experiment occupying an entirelaboratory, many accelerated motions canbe studied together in one laboratory;instead of one cooling curve, a set ofcooling curves of various mixtures can bestudied in the same length of time. Theresult has to be that students will havemore raw data available to them on whichto fashion their theoretical model of thenatural world. Because less time is de-voted to the calculation and presentationof data, more time can be devoted to un-derstanding the scientifice phenomena. Inmany cases, the hand calculations are notonly time-consuming but distracting;because they occupy most of the lab time,students tend to focus on them and totallyforget the scientific phenomena being in-vestigated.
A secondary but important argument forthe inclusion of microcomputers in theteaching laboratory is that by the timeour students are employed, those that gointo the laboratory will find microcom-puters there.
THE FUTUREI feel that the microcomputer viewed as
a laboratory instrument is an emergingresource which is cost effective today incertain teaching situations and will beused in the immediate future in anextremely broad range of science teachingenvironments.
Current EffortsThe pestary impediment to the wide
scale use of microcomputers in the labora-tory now is that software and hardware arenot readily available from any one source.This, in turn, is partly because commer-cial vendors do not perceive the teachinglaboratory as a commercially viable areabecause there hus been no broad expressionof interest by laboratory teachers.
As a non-profit organization dedicatedto improving science instruction, ourgroup at TUC have undertaken a number ofprojects to speed the introduction of themicrocomputer into the laboratories. Wehave, of course, generated some sampleprograms that illustrate the kinds of
things that can be done, and we also eval-uate and distribute hardware and softwareso that schools and teachers can find mostwhat they need at one place. We have de-veloped three laboratory interfaces andare currently working to Interest somemanufacturers in these or other relatedproducts.
For a number of years, we have beengiving workshops for teachers on the use ofmicrocomputers in teaching. These work-shops have been very well received and,in fact, we face the problem that thereare more people interested in taking theseworkshops that we can possibly enroll. Tomeet this need, we are in the process ofdeveloping an exportable workshop -- a one-day workshop that former workshop partici-pants can themselves offer to others. Weare currently looking for groups ofteachers who will participate in the firstround of these workshops and be willing,then, to host these workshops themselvesin their immediate locale.
Long-range EffortsWe see a time in the near future when
microcomputer-based instrumentation willbe as widely used in introductory scienceteaching as microscopes, clocks, and ther-mometers. The microcomputer will not be afrill added on top of other instrumentation;it will be the primary instrument. Becauseof the ease of programming and flexibilityof input and transducers, grade schoolstudents could use microcomputers in muchthe way they currently use microscopes, toextend their senses, to help them perceivephysical phenomena that are outside therange of their immediate senses, and toprovide the questions and motivations forscientific studies. Most of the elementsnecessary to fulfill this vision are cur-rently available, and it is certain thatprices will drop over time to the point atwhich few schools could justify not havingmicrocomputers in the laboratory.
Meanwhile, however, we see four areasin which important work needs to be donebefore microcomputers will be widely uged:education research and development, improve-ments in the laboratory interface hardware,network software for sharing limitedresources,.and development of curriculathat use laboratory-interfaced computers.
EdttionResearchdDeveloent. Wesuggest t at t of
microcomputers in the laboratories, whilehaving many advantages, also creates cer-tain pitfalls. Both the advantages and thepitfalls need careful researching. In whatsense can students learn to master such apowerful tool? What intellectual pre-requisites must they have? The questions
2e'V
raised about mystification and enabling ofstudents must be studied in greater detail.There is a possibility that the introduc-tion of laboratory experiences that use themicrocomputer will be of benefit not onlyto science instruction but also to mathe-matics learning. When combined with thepower of the computer to analyze and model,the introduction into mathematics ,:tstruc-tion of computers that can deal with realphenomena will have an important role inmotivating students in building intuition;this area needs to be researched and de-veloped.
Interface Hardware. We see a continuingneed to keep abreast of hardware develop-ments and continually incorporate relevantinnovations into interface designs. Itis only by producing and distributing in-creasingly sophisticated prototypes thatwe will learn how these can be used in edu-cation, how costs can be minimized, andhow dissemination problems can be reduced.One of the benefits of our work in build-ing and distributing prototypes is that ithelps build a market we hope commercialcompanies will want to exploit.
We see a clear need to develop an in-terface that contains its own 16-bit CPUand communicates to a host computer overthe new general interface bus or overother standard communication lines. Sucha device would have greatly increasedperformance and would be compatible withalmost all commercially available compu-ters. While it may not be affordable bya large number of institutions this tine,it is important to learn, through limiteddistribution, what educational' advantagesaccrue from the increased performance thisunit would have.
Networking. The expensive part ofcomputers now is not the central process-ing unit or the memory but rather theperipherals: printers, disks, specializedinterfaces, graphic displays, and thelike. The appropriate way to maximizecomputational performance is to have ameans of sharing the expensive resourcesamong a variety of inexpensive computers.Thls is called networking and is themicrocomputer alternative to time-sharingwhich is really only appropriate forcomputers that have expensive CPUs andmemory'. We have some interesting net-working software and hardware currentlyoperating in our lab that need to bebrought into the teaching environmentand expanded to laboratory applications.We envision that the teaching laboratoryof the future will involve a network ofinexpensive computers at each laboratorystation, networked with a number of moresophisticated special-purpose microcom-
Computer Laboratories in Education 255
puters to be used for extensive dataanalysis, combining data from many studentsand generating complex displays. Thislocal network would be tied into a centralcomputing center, which could be used forlong -term, storage, high-speed printing,and the maintenance of large data bases.We are cooperating with a number ofschools to work towards an initial imple-mentation of this ideal.
Curriculum. The most pressing currentneorii-Wargeneral area is to developcurriculum material that uses laboratorymicrocomputers. One can view the programsthat we have developed so far as samples,a shotgun blast that illustrates what canbe done. There is an urgent need syste-matically to apply laboratory-based micro-computers to courses in various disci-plines. An important aspect of thiseffort is to try to define the electricalcharacteristics of the interface that isrequired, so that the curriculum materialdoes not haye to be tied to any particularhardware. We naturally feel that our labinterface defines this standard, but wewould be del.ghted to discuss any otherpossible standards with any educators.
The only way extensive curricula willbe developed is to use teacher-generatedmaterial. The major problek with thisapproach is the evaluation and distri-bution of this material. We have estab-lished an experimental MicrocomputerTeacher Resource Center that uses a com-mercial data base to store reviewed filesand programs. This data base is publicand contributions are welcomed. Ifteachers will evaluate what they usefrom this data base and contribute somenew material, great strides will be madein filling the missing curriculum gaps.
SUMMARYA microcomputer equipped with a labora-
tory interface offers a very exciting newresource for science instruction. Withthe proper hardware, the computer can beturned into an altogether general-purposeinstrument. It can replace a large numberof standard instruments at a fraction ofthe cost, and provide the major measure-ment and analytic tools needed in a widerange of science courses. Equipment andsoftware are available today that beginto do this, while future hardware develop-ments and declining costs will make lab-based microcomputers a necessity in naryteaching situations.
20 0w 0
256 .4E0C 1980
THE COMPUTER LAB OF THE 80S
Guy Larry BrownHead, Data Processing
Piedmont Virginia Community CollegeEt. 6, Box 1-A
Charlottesville, Virginia 22901(804) 977-3900
INTRODUCTIONDuring the past 20 or so years that the
computer has been used as an academic roof,there has been considerable discussionabout computer-assisted instruction (CAI).CAI, in fact, has been a luxury affordableby only a few. In recent years, with pres-sure being applied to educational budgets,it appeared that CAI would remain a dreamto be realized only if the good foivy god-'other waved her magic, wand in a class-room. However, the advent of the micro-computer witb its by cost and super cap-abilities has made it possible for thatmarvelous wand to be directed into eventhe most ispoverisbed academic niches.This paper describes an operating systemof microcomputers that is unique in capa-bilities and costs.
THE SYSTEMThe system in operation tonsists of 14 .
microcomputers eacb with 8,192 (8K) ofrandom access memory (RAM) and 8K of readonly memory (ROM). The RAM allows storageof instructions written in Microsoft BASICand data in the form of letters of the al-phabet, digits, and special characters.The ROM contain' an interpreter that tran-slates BASIC into the computer's language.Each micro has its own keyboard for inputinto the computer and a video monitor forvisual output.
Stored programs are available fromshared dual floppy disks, each capable ofholding up to 275K characters. The diskunit is part of the host computer tbreughwhich control of the micros is centrallyexercised. This host computer has 48K RAMwith its BASIC on disk instead of in ROM.Control of the bost.is through a CRT. At-tached to the host is a 100 character persecond (CPS) matrix printer which is alsoShared by the micros.
The program (software) that providesthe capability for the micros to share thedisks for input/output(1/0)and the print-
er for hardcopy output occupy only a smallpart of a single disk. The remainder isset aside to store student programs andfor other uses as may be desired. Forexample, a program can be,colled by a stu-dent to provide information about a problemhe does not fully understand. Or a seriesof tests can be stored for use by the stu-dent at the instructor's discretion.
Two of the micros received minor modi-fications to permit I/O througb a standardcassette player. This capability permitsthe units to be detached from the systemand moved out of the lab for demonstrationsor other purposes. Zr also allows for thetransfer of a program stored on a cassetteinto the micro's memory and hence onto adisk for storage and subsequent availabil-ity to all micros.
All computers were manufactured by OhioScientific, the printer is a Centronics779, the video monitors are Sanyo VM 4209,and the CRT is a TEC Series 500. The sys-tem is capable of operating with 16 of themicros; however, to keep total costs below$19,000, only 14 were purchased initially.This figure included all equipment,instal-lation, and softweref Equipment selectionwas influenced by the availability of areputable local dealer who services whathe sells. Also favoring Ohio Scientificis the large inventory of software,' smallbusiness system capability with hard disks,and design features making expansion andupgrading simple snd relatively inexpen-sive.
OPERATIONThe system ins operated almost flawless-
ly since installation in the fall of 1979.Typical micro systems use cassette I/Owhich is very slow compered to disk. Anda dedicated disk for stand -alone microsadds significantly to the costa the unit.The shared system allows each micro userto (1) load a program stored on disk intothe memory of his computer, (2) save a
2
program on disk, (3) print a listing ofthe program, or (4) print the output fromthe execution of a program. Only one microcan use the disk or printer at a time.The use of passwords for files, messagesending to the host computer, and lineadvancing on the printer are recent,locally added enhancements to the oper-ating software.
USESPresently the system is used only for
teaching the programming language BASIC.One 20-student class was taught in thefall of 1979, and two such classes arebeing conducted during the winter quarter.
During registration, several units weremoved from the lab for demonstration pur-poses. Curiosity was aroused in many whostopped to engage in conversation with thecomputer or play a game.
FUTURE PLANSIn addition to expansion to 16 micros,
there are plans for the acquisition of aquality character printer for word proces-sing. This capability will permit in-struction of secretarial science studentsand members of the staff in word proces-sing.
Expansion of courses to cover assemblylanguage is anticipated as well as a coursein the operationof small business computersystems. Longer range plans call forteaching COBOL.
Efforts are being made to make the labavailable to any of our faculty for use intheir courses. Programs will be developedfor math, chemistry. biology. and a varietyof business courses.
ADVANTAGESCost.Reliability.Graphics capabilitiesof 256 characters.Standardization of components.A single disk for program/data storage.Display of 64 characters/32 lines on
micro videos.Extensive software available from man-
ufacturer.Expansion of capabilities. and updating
inexpensive.Not dependent upon trained computer
personnel.Rapid transfer of programs/data through
a disk system.Hardcopy available to all micros.Control is facilitated thxdOsh the host
computer.No reliance on telephone connections.Students learn on a micro frequently
found in businesses.Micros still function independently.
Computer Laboratories in Education 257
DISADVANTAGESDisk and printer allow only one user at
a time (this is only a minor incon-venience).
Cost is so low that academic departmentscan afford procurement without in-volvement of computer management(maybe this is an advantage).
Documentation is not complete or errorfree.
Operating software lacked capabilityfor printing the execution outputof a program (locally modified toacquire this capability).
Service may not be locally or quicklyavailable.
Cost is so low that one cannot expectthe same level of manufacturer sup-port as previously experienced whensystems were priced much higher.
Demand for usage night be so higb thatthe establishment of priorities foruse may be difficult.
SUMMARY .
A computer lab for the &Os is availabletoday. It consists of up to 16 microcom-puters, each of which can bave up to 32KRAM. They have rapid access to programsand data through a shared dual floppy diskdrive which also provides for equallyznpidstorage. Hardcopy listings of programs andoutput from the execution of thoserrogramsare also available on a 'bared printer.The system, including the host computer andall other hardware, installation, and nec-essary software is available for around$20,000.
270
25$ NECC 1900
THE EDUCATIONAL TECHNOLOGY CENTER
Alfred Bork, Stephen Franklin, and Barry KurtzEducational Technology CenterUniversity of CaliforniaIrvine, California 92717
(714) 833-6911
This paper reports on the formationof the Educational Technology Center atthe University of California, Irvine. Theprimary focus of the Center is the use ofthe computer as a learning aid. TheEducational Technology Center was started onJanuary 1, 1980, with University fundsproviding staff support. The Center con-tinues the activities An computer-basedlearning conducted by the Physics ComputerDevelopment Project during the last elevenyears.
NEEDThe Educational Technology Center was
formed because we believe strongly thatthe next decade will be a critical periodin American education. Such centers areneeded to guide us toward a future wherette computer will play an extremely im-portant role An education. It is impor-tant to develop a number of continuinggroups that are not fully dependent ongrant funds but have an existence beyondsupport for particular projects.
We have pursued for some years withinthe University of California the possi-bility of one such Center. Estill pro-vide guidance to others working in thisarea. The Center will work on a widerange of research and development activi-ties leading to more effective use of thecomputer and associated technologies inlearning environments.
CURBINT ACTIVITIESThe Educational Technology Center
intends to engage in many activities con-cerning more effective and more efficientuse of information technology in learning,emaphasising learning materials on theWien.' computer. Some of the activitieswill be pure research, while others willhave& strong applied and developmentalcomponent. We shall work closely withindividuals and groups elsewhere, as in
the past, so that the Center has anationwide effect beyond its immediateactivities, materials, and publicity.
The Center will publish a newsletterreviewing the activities and results ofits projects. Although no set scheduleis planned, we expect this newsletter tobe published three times a year. Anyoneinterested in receiving the newslettershould write to the Center.
The following list gives the activeprojects at the Irvine Center. Furtherinformation about any activity is avail-able on request.
1. A. Testing and Tutoring Environmentfor Large Science Courses.Authoring for personal computersTesting environmentsPhysics - wavesStatisticsNational Science Foundation--Compre-hensive Assistance to UndergraduateScience Education (CAUSE)
2. Scientific Literacy in the PublicLibrary.Public libraries, shopping centers,
science museumsPublic understanding of sciencePersonal computersFund for the Improvement of Post-secondary Education (TIPS!)
3. Mathematics Competency Tests forBeginning Science Courses.University of California/CaliforniaState University and Colleges
4. Translation of timesharing materialsto personal computers.Universit of California/CaliforniaState University and Colleges
5. Biology materials ecology.University of California, Irvine,Committee on Instructtonal Develop-ment
27j
6. Development of Reasoning Skills inEarly Adolescence.Junior high studentsTransition to formal reasoningPersonal computersNational Science Foundation - Devel-opments in Science Education (DISE)
PRODUCTION SYSTEMIn addition to specific products,
such as those just mentioned, the Centerhas developed a production system forgenerating computer-based learning mater-ial. The emphasis is on both efficiencyand effectiveness and on techniques whichwill allow natural extensions to large-scale production of such models. Theproduction system is based on a systemsanalysis of the problem and on our manyyears of experience in producing a widerange of learning material. Literature isavailable describing the system and thesupporting software.
ISSUES FOR THE FUTURECurrently we can distinguish a number
of very important issues that will shapethe future of computer-based learning:these issues indicate directions theEducational Technology Center will pursue.No order of priority is intended in thislist.
1. Full-scale course development.At present, with a few notable exceptions,computer-based learning materials are sup-plementary to course structures. Very fewfull courses mike heavy use of computersto aid learninc. We need experience indeveloping such complete courses and inintegrating computer and other learningaids. We need additional experience incomputer-aided delivery of such courses.
2. Expanded acquaintance. Very fewteachers, and even fewer members of thegeneral public, have seen any effectivecomputer-based learning material. Oftenthe examples seen have been weak examples;so the learners have formed inaccurateopinions of the value of such material.We need more acquaintance with the fullrange of possibilities, more computerliteracy with a learning emphasis.
3. Research in learning. Presentlywe have conflicting theories about learn-ing. We need to know more about how stu-dents learn so that we can develop betterlearning aids.
4. Production techniques. Olderstrategies for developing materials oftenwere not suited for the large -scale devel-opment needed in the years ahead. Thetypes of systems approach followed atIrvine and elsewhere needs further explor-ation and refinement as the scale of
Computer Laboratories in Education 259
activities increases. We should aim forthe best possible materials at the leastdevelopmental cost.
5. Expanding technologies. Com-puter and associated technologies areevolving rapidly. We must learn quicklyto use an.expanding range of capability,developing materials which are not immed-iately outmoded.
6. Thec,terinewinterac-tive mediumW-ewearrrnigmedium, we must learn how it
differs from older media. For example,reading from computer displays has manydifferences from reading print medium, butthe empirical details are not known.
7. Dissemination. New media alsodemand new males of dissemination.
S. 'sew course and institutionalstructurss. As computers are more widelyused, they will have major effects oncourse and institutional structures.
The Educational Technology Centerintends to pursue these and other issues.
2
Invited Session
MIS EDUCATION:INDUSTRY NEEDS AND EDUCATIONAL SOLUTIONS
Chaired By Eleanor W. JordanDepartment of General BusinessBusiness-Economics Bldg. 600university of Texas at Austin
Austin, Texas 78712(512) 471-3322
ABSTRACT7574FEhe past decade a considerablenumber of discussions in industry-orientedpublications like Computerworld havefocused on irrelevant education as a reasonfor prevalent software problems and DPpersonnel shortages. In this session,educators will discuss efforts made bytheir institutions to resolve the supposedrelevance problem in programs for businessapplication's software designers at thegraduate and undergraduate level. Twoindustry representatives will also partic-ipate in the panel.
200
PARTICIPANTS (Listed in order of Presentatzon
Marguerite Summers, ChairpersonWestern Illinois University
David NaumannUniversity of MinnesotaMBA and PhD PIS programs
Joyce J. ElamWharton SchoolUniversity of PennsylvaniaMBA concentration in MIS
Eleanor W. JordanUniversity of Texas at AustinUndergraduate DP program
James CookSouthwest Texas UniversityUndergraduate business CS program
Ken TruittARCO Oil and Gas Company
Willis WareRand Corporation
27ki
Computer Games in Instruction
SHALL WE TEACH STRUC-TURED PROGRAMMING TO
CHILDREN?Jacques E. LaPranccDcpt. of Mathematical
ScienceOral Roberts Univ.Tulsa, OK 74171
ABSTRACT--liZZT011ege programming experience hasbeen observed to be detrimental in manycases to college-level study of computerscience. The problem is the lack of un-derstanding of the principles of struc-tured programming. The solution proposedhere is the introduction of structuredprogramming games at the elementaryschool level which will prepare thechildren to do structured programminglater on with languages such as BASIC andFORTRAN that are not designed to promotestructured programming. An example ofdoing this called ANTFARM -s presentedand the results of using it with onegroup of children are discussed.
THE PROBLEMRUFF-7E7 more students are enteringuniversity computer science programs withsome form of prior.experience with com-puters. This is typically a high schoolcourse in FORTRAN or BASIC or experiencewith a personal computer. There seems tobe no corresponding exposure to struc-tured programming or design, however.The students believe they know a lot be-cause they have written programs in BASICor FORTRAN, but in reality they know onkcoding; they understand nothing of struc-ture, top down design, good style, ordocumentation. It is sometimes difficultto get them to break the resulting badprogramming habits. This problem is like-ly to be compounded by the increasingavailability of small personal computerswithout proper accompanying instructionin good top down programming. An in-creasing number of young people willlikely learn programming by reading theirmicrocomputer BASIC manuals. The learn-ing of coding will thus be encouragedrather than the learning of programming.Yet these students will be misled intothinking they know a lot because they are
able to code some substantial programs.The task of computer science educationwill be made more difficult by this back-ground experience because the studentswill have to do more unlearning than iscurrently the case and sometimes unlearn-ing is more difficult than learning.
THE PROPOSALA possible solution to the above prob-
lem would be to introduce structured pro-gramming concepts to children before theyare able to begin to use conventionalprogramming language. Other oulutionsmight be to replace BASIC wich PASCAL asthe most common microcomputer language orto have all manufacturers' BASIC manualsbased on structured programming. Neitherof these alternative solutions seemsfeasible, The simplicity of BASIC willcause it to continue to be poptIlr, es-pecially with the smallest computer con-figurations. The large number of pro-grams already coded in some dialect ofBASIC will also serve to keep support forBASIC strong. The manufacturers may bemore or less willing to include struc-tured programming concepts in their man-uals, but the bottom line will always bewhat enables them to sell their product.The general public will have to becomemore knowledgeable, discriminating, andparticular before the manufacturers willfeel any serious pressure to include ma-terial on structured programming.Although the first solution mentioned
above is not a shoo-in, it is one thatcan be developed more easily than eitherof the others and would be worth inves-tigating. It is not realistic to haveall elementary schools begin teachingstructured programming, but a few well-published success stories plus the avail-ability of well-written materials for usein schools would help to promote theidea. A few successful projects at
261
274
t,
282 NECC 1980
schools in several areas across the coun-try would encourage other educators . iosomething similar and inspire comp...ermanufacturers to include structured de-sign and programming in their manuals.Structmed programming 'concepts could
easily be introduced to elementary schoolthildren if they were presented in e geneform at their level of sophistication.Gamos could be developed for which thecon 11 language is inherently struc-tue and the children would be moti-val.ea by the attraction of the game tomaster the control language as well aspossible in order to do more with thegame. Because the language would be in-herently structured, they would automat-ically begin to develop structured prob-lem solving skills and structured expres-sion of their ideas. Manuals could beprovided that would describe structureddesign and programming in a way that mostteachers could follow and would alsodescribe how to use the games with thechildren to develop their understandingand skills.The LOGO system at H.I.T. Artificial
Intelligence Laboratory (12) already con-tributes to these goals. This system wasnot developed for introducing structuredprogramming concepts but aould be easilyadapted to that purpose. It is excitingfor the children to use (14) and there-fore has the necessary motivational char-acteristics. Several other tools forteaching structured programming could bedeveloped with different appeals, dif-ferent levels of sophistication, and dif-ferent organizational structures.
THE EXPERIMENT(71.:7taing of these ideas was made
in April 1979 by the author and Dr. MaryDee Fosberg of Central State University,Eemond, Oklahoma. To present the ideasof programming to a group of giftedchildren, we designed a gaze calledANTFARM. It was implemented on an IMSAIwith a Z80 processor, Digital Micro-systems disk, -Aid Infoton 200 CRT usingthe UCSD PASCAL system which we took tot' .f presentation and set up there.The ANTFARM program consists of drawing
in ant facing up in the center o2 the..creen and two rows of food ("e's) inthe upper left. The goal is to have theant move over to the food, eat it, andplant new food. The ant looks like:
\ *l *1 /0 or -0 or *00-.
0 / \71\ 1\
etc., depending on orientation
It accepts five basic commands: MOVE,TURN LEFT, TURN RIGHT, EAT, and PLANT.MOVE means move forward one characterposition in the direction the ant is fac-ing; TURN, LEFT or RIGHT, means a 45° ro-tation about the first body segment; EATmeans to consume whatever is under hishead; and PLANT means to plant a new seed
swith his tail. Seeds start out as ".";100 time units later they germinate(","); then after 50 more units theysprout ( " ; "). They grow into a stalk("1") after 50 more units and then into abranching plant ("Y") after 50 more. Atthe end of 75 more units, a flower: ("P")appears and then after 75 more or a totalof 400 units, the plant natures into food("S"). One time unit is the tire neces-stry to execute one operation. Such agrowing rate seems to be a satisfactorychoice though unrealistically fast.Each time the screen is updated, the
cursor is moved to the home position, andthe first line on the screen is erased toawait the next input. The user may typeany combination of commands on this line.When the return key is pressed the com-mends sze performed. We began by simplyshowing the children the effects of indi-vidual commands. Then we introduced theconcept of sequence by showing that sev-eral com.ands can be listed on the sameline.Since the sequence of commands cannot
be longer than one line, the limit onprogram size is quickly reached. Othercommends are then introduced to allow theant to do more things. From the author'spoint of view, however, additional com-mands are given to develop the additionalstructured programming concepts of itera-tion, selection, and refinement.The first additional command illus-
trates iteration: "DO a cc.amand sequencen TIMES," where n is an integer. Insteadof "n TIMES," there can be "TO ROWinteger" or "TO COLUMN integer." Therows ate numbered down the left of thescreen and the columns across the top onthe second row. This command allows theant to go through a sequence many timesbefore stopping for the entering of thenext command. (The "DO" command can beused in conjunction with the do nothingcommands "WAIT" and "REST" to watchplants grow.)The command to illustrate selection has
the form:
PLANTIF SEE (FOOD
,SPROU1LDIRT
[AHEAD)LEFT command.RIGHT
If the character space in front (or tofront left or front right) contains non-space for "PLANT," " @" for "FOOD," ";" or"1" for "SPROUT," or space for "DIRT,"the single command following will be exe-cuted; otherwise it will be skipped.This command allows for some interestingconditional tasks such as having the antlook for a row of food to eat or an openfurrow in which to plant more food. Al-though the chi'dret were able to use thiscommand it seems to need more developmentto be stronger. In its present form, itseems too restrictive.To encourage the development of hier-
archical structure and refinement, thecommand sequences are limited to oneline, and a command called "REMEMBER" isgiven. This command has tho form "REMEM-BER name command .sequence END" which al-lows the child to give numesto commandsequences for subsequent use as commandsthemselves. One popular sequence is
REMEMBER TURNAROUND DO TURN LEFT 4 TIMESEND
after which "TURNAROUND" can be used as acomman.l. This subprogram capability isespecially helpful when used in connec-tion with the selection command sinceonly one command may be selected orskipped. Tie commands "QUIT" and "STOP"allow the currant subprogram to be endedprematurely (or the ANWARM program it-silf to end if it is used at the command:t.ae level). This capability allows forpowerful search features when used insubmodules with selection and iterationcommands, such as
REMEMBER MOVETOFOOD DO MOVE IF SEE FOODAHEAD STOP 70 TIMES END
This will cause the an- to move forward71) times or until there is food right infront o' it, whichever comas first.
One of the tasks the children raced VASthe planting of a field of four rows often plants each. The following is an ex-ample of a structured development of asolution:
Computer Game': in Instruction 263
The final commands availab. are "FORGETname," which eliminates th- named modulefrom the table, and "nu," which listson the screen all the "REMEMBER"ed namesand their definitions. 14'...n the use"pushes the return key after the "TELL"display, the screen is cleared and theant's farm scene is redisplayed as it wasprior to the "TELL" display. A futureaddition will be commands to save and re-call the'set of defined modules to andfrom a disk file. Thece commands willallow a child to build his program fromday to day rather than having to typeeverything over from the beginning ateach session.
THE RESULTSWe had about 13/4 hours with a group of
gifted children between the ages of nineand twelve. Interest was universallyhigh throughout the session. The formatof the session was that for about 20 to30 minutes the children were shown someof the hardware aspec..,, of computers;then the ANTFARM was introduced. Sincethe Infoton has a detached keyboard, theauthor sat to the side and did all thetyping while the children looked at thescreen and said what to type. We intro-duced the ANTFARM features roughly in thesame order as in this paper, stoppingfrequently to ask the children how to getthe ant to do sou specific task. Thechildren picked up the ideas very quicklyand were soon telling us of th-ngs theywanted to see the ant do. By the end ofthe VI hours, three or four of the mostinvolved children were beginning to usetopdown design to achieve some specifictask they wanted the ant to do, definingtheir own modules and giving them names.The only weaknesses observed were the
selection commands and ,:he bugs. Theselection commands were too difficult forthe -hildren to use unaided in their pre-sent form, primarily because the condi-tion was something happening out in frontof the ant instead of right under him,and the seeing left or right does notcorrespond well to Where the ant isfacing when he turns. The bugs occasion-
REMEMBER FIELD TWOFURROWS TWOFURROWS ENDREMEMBER TWOFURROWS FURROW NEXTRIGHT FURROW NEXTLEFT ENDREMEMBER FURROW DO MOVE PLANT 10 TIMES ENDREMEMBER NEXTRIGHT DO TURN RIGHT 3 TIMES MOVE TURN RIGHT MOVE MOVE ENDREMEMBER NE.;TLEFT DO TURN LEFT 3 TIMES MOVE TURN LEFT MOVE MOVE END
A field could be planted in the upper right part of the screen by
DO MOVE TO ROW ? TURN FIGHT TURN RIGHT DO KOVE TO COLUMN 65 FIELD
2 76
264 NECC 1980
ally caused the program to abort, losingall the defined modules and all the farmdevelopment to that point. Actually theprogram worked quite well considering itwas developed and implemented in three orfour man-days and consists of over 400lines of PASCAL code. (This accomplish-ment in itself is a credit to the valueof top-down design, structured program-ming, and PASCAL.)
THE CONCLUSIONThe ANTFARM program has demonstrated
Chat it is possible to create tools withhigh motivational value for children thatcontain ell the concepts of structuredprogramming. Furthermore, the way thisprogram is designed. forces the childrento use hierarchical or structured devel-opment to achieve their goals. Even inone session we were'beginning to see someof the children using structured develop-ment, on their own. It seems reasonable
suggestuggest that continued use of toolssuch as =FARM over an extended periodof time could develop these concepts sowell in children that they would continueto use top-down design and structuredprogramming even with languages such asBASIC and FORTRAN which are not naturallyoriented toward them. We hope that wewill have the opportunity to continue toexplore these ideas and that others willhave similar opportunities. We needlong -term data on children using thesetools to determine the effects on theirfuture success in programming or computerscience.
13. Papert, Seymour and Cynthia Solomon,"Twenty Things to Co with aComputer," MIT, a. I. Memo 248(June 1971).
14. Papery, Seymour and Harold Abelson,Jeanr0 Bamberger, AndreadiSess, Sylvia Weir, "ZnterimReport of the LOGO Project inthe Brookline Public Schools:An Assessment and Documentationof a Children's Computer Labor-atory," MIT, A. I. Memo 484,(June 1978).
15. Perlman, Radia, "Using ComputerTechnology to Provide aCreative Learning Environmentfor Preschool Children," MIT,A. I. Memo 360 (May 1976).
16. Ragsdale, Ronald G., "A Program. Package for Introducing the
Top-Down Approach to ComputerProgramming," SIGCSE Bulletin,#11-1 (Feb. 1979174. 113-117.
17. Solomon, Cynthia J. and SeymourPapert, "A Case Study of aYoung Child Doing TurtleGraphics in LOGO," MIT,Artificial Intelligence Memo375 (July 1976).
18. Watt, Daniel R., 'A Comparison ofthe Problem - Solving Styles ofTwo Students Learning LOGO: AComputer Language for Child-ren," Proceedings of theNational EducaticTarini'eutingU5EfiFiFice7-11707-pp. 255-260.
19. Weinberg, Gerald H., The Psychologyof Computer Promammjag,RFw York: Van NostrandReinhold, 1971.
2 :7S
Computer Games in Instruction 265
266 NECC 1980
STRUCTURED GAMING: PLAY AND WORK IN HIGH SCHOOL COMPUTER SCIENCE
J. M. Moshill, G. W. Amann(The University of Tennessee)
W. E. Baird(West High School)
Knoxville, Tennessee
PROLOGUEQuestion 1: When, is a computer,gare not agame? When is it an ok class-activity?Question 2: So what's wrong with games,anyway?Exasperated answer to 2:Students won't work on programs when theyhave access to games. Games are fin andprograms are work.Reflective answer to 2:
The usual computer games are eitherhand-eye (with occasionally some small ,
amount of brain-) coordination contests,such as "Lunar Lander," or fantasy-land,interactive do-it-yourself storybooks suchas "Dungeons and Dragons.* These activi-ties take part in the allure of broadcasttelevision: namely, they involve thestudent kinetically and emotionally, butthey do not have a cumulative component.You can walk in on television (or computergames like "PONG" or "SPACE WAR")anytime; no prerequisities or logical-deductive skills are required. (For anexcellent exploration of this theme, seePostman, 1979.)
We cannot call this kind of attentionpassive -- observe any kid watching anaction TV show or playing a video game.Nevertheless the interaction is non-ana-lytical. It has more in common withbaseball than with reading, more of re-.cess than of curriculum. No wonderteachers of computing have game troublewhenever interactive terminals or microsbecome available.
The problem this paper explores isthe development of an introductory cos-puting curriculum built around a kind ofstruc*ured gaming. The computing commu-nity has begun to understand that care-fully chosen programming language featurescan guide our thought in ways that makecode work across time, that. is, remainadaptable, comprehensible, repairable.
lualr74.4
ThruatiallyeupEortedbytSF GrantSED-79,-91
We propose that a similar choice of aminfeatures can foster the development ological problem-solving skills, whileretaining the kinetic/esthetic motivationalstructure of video games. (We have allknown programming hacks who have made thegame/program connection.) We want to usemidrecomutercolorgmapidestomakeonoputingmore Mtn color crayons and less likearithrettc.
Having said that much, we will answerQuestion 1 and then flesh out our answerwith a description of the curriculum weare developing.Question 1: When is a computer game an ok
class activity?Answers 1) The game must be designed with
a set of concepts and skills inmind and a plan for how thegame teaches theses
2) The things learned in the acti-vity must contribute toward acumulative body of knowledge,a toolkit that the student canperceive and make use of, astoys and toy-making tools, and
3) The game must be su erseded bya more interesting, more nter-active game, chosen with extremecare to be cnplayabln unless thestudent has mastered the skillstaught in the previous lesson/game.
CHALLENGEOur mission, in the University of
Tennessee/NSF High School Computer ScienceCurriculum Project (HSCS), is to make com-puter skills available to average students.Computers may indeed become as ubiquitousas telephones and televisions, although webelieve that the introduction of anothertechnology as soporific, captivating, andanti-thinking as television could be amajor social disaster. M. hope, rather,that computers will become convivial toolslike the telephones "convivial" means thattheir use is determined by the user, not
by some central leastcommon denominatorsuch as broadcaster. We don't have utopianideas as to what future generations willdo with computers. (Who could have pre-dicted in 1915 what problems we'd have withautomobiles?) We do, however, have astrong feeling that the question ofwhether individuals will be able to programtheir computers, or merely ha Programs, san open and important questiori. The chal-lenge, then, is to give every citizen whocan dial the telephone some ability to pro-gram a computer.METHOD
This section will be brief; we havepublished elsewhere (Aiken, Hughes, MOshell1980) the nuts and bolts description ofHSCS. We are using a cartoon-animationsoftware system called RASCAL which runsas part of UCSD PASCAL on the APPLEmicrocomputer. The basic installationcosts about $3200, including a singlefloppy disk, color television, and 100character-per-second printer. Each lessonin a one - semester (18 week) course consistsof approximately a week of work, dividedinto these parts:
A class consists of about 15 students percomputer (our collaborating schools haveonly one APPLE each we hope to try thecurriculum in multi-computer classes later),Five groups of three students alternatecomputer use with planning work usinggraph paper and marker pens. The off-linestudents ere planning their strategies,doing hand simulations and observing theon-line students, fur to graduate to thenext activity, a group must successfullypredict the outcome of an assigned "seed"(e.g., geometric pattern, algorithm. pro-gram). The activities develop during 18weeks from a non-linguistic, color-patternprocess called "quilting" (next section),through immediate-mode and straight-line-code entry of TURTLEGRAPHICS (Papert,1970)commands and the introduction of PASCALcontrol structures such as REPEAT...UNTIL and IF...THEN, to the creation ofcartoon characters and their animationwith complex programs using the RASCALanimation system. The output is alwayscolor graphics and music; the curriculumsteadily increases its interaction asstudents learn how to use the joystickto control motion. There is always anunderlying lesson about how programswork. Al. code is in a completelystructured language (PASCAL) and istaught from the inside out. Only at laterstages do environmental details such as
Computer Games in Instruction 267
declarations become of concern. A PASCALinterpreter is used which scroUs thesource psgram being executed on the bottomof the screen (at a controllable rate)while the program produces its output onthe top part of the screen. A workingsystem will be on exhibit at NECC2.
We will conclude with fairly detailedexplanatiqn of the first two lessons,quilting and TURTLEGRAPHICS.TWO EXAMPLE LESSONSLesson ls WM-W.--Behavioral Objectives: Students shouldlearn how to insert a disk and start thesystem. They will run the Quilt program,explore its features and limits, and useit to generate :static and moving colorpatterns.Conce ts: Students will learn the follow-ng:
1) how to enter an initial pattern ofinformation into the system whichthen controls the repetitive be-havior of the computer(Introduction Activities).
2) 116:7; to construct simple experimentsto determine the parameters of thecomputer's operation, such as howmany colors it has, how big thescreen is (Exploration Projects) .
3) how to hand simulate a formal pro-cess of discrete steps to predictthe computer's behavior and under-stand it(Skill-building Projects).
Resources:IT-The Quilt program. Students insert
a disk and on the computer.When the greeting message appears,thuv type X (execute). Thecomputer asks EXECUTE WHICH FILE?The student types QUILT and a re-turn. QUILT then prints the fol-lowing message:THIS PROGRAM LETS YOU DRAW A PIC-TURE.AND THEN MAKE A 'QUILT' OF ITBY REPEATING THE PATTERN. TO DRAW:SET PADDLE 0 FOR A COLOR;TYPE KEYS ARCUND K TO MOVE THE DOT.
FOR INSTANCE, 'I' MOVES THE DOTUPWARD.
TYPE 'R' TO SEE THE PATTERNREPEATED.TYPE '0' TO HALT THE REPEATING.TYPE 'C' TO CLEAR THE SCREEN.TYPE 'Q' TO END THE PROGRAM.NOW HIT THE RETURN KEY TO BEGIN.When the student hits the returnkey, the following reminder messageremains on the bottom of thescreensREMINDERS:KEYS AROUND,MOVEsIstUP,J=LEFT,U=DIAG,ETCC) LEAR REPEAT MALT DE REPEAT=QMSITTETWOUSPROMM
2S9
268 NECC 1980
The game paddle supplied with theAPPLE computer consists of a knobcontrolling a computer input; the studentturns it and observes that the dot in thecenter of the screen changes color. Shetypes an experimental sequence ofkeystrokes such as LLLI and observes thatthe dot moves and leaves a colored trailshaped like loigure 1. She than tries typing R.The coseuter immediately reproduces the pattern,starting where the cursor ins left and finis; thescreen with a periodic pattern. (Since the screenis waived, when the cursor dot falls offscreenleft, it returns on the right, and similarly fortop-betben.) If you RP:levet the. above seed, you
get Figure 2.2) Miscellaneous resources: cusbzweede
40x40 graph roper, five or sax sets ofcolor crayons.
Sequence of Events:The first day of this lesson consists of
introducing the program and allowing students toplay with it for a few minutes each. They are sentany with copies of a sheet of exploration projests to be undertaken the awn day. These pro-jects are carefully ordered by increasing difficultyso that the first group on the machine will havesane chance of success, while the second group 411have to do ,scale planning before getting on the on:-puter Mile the first group is working) in orderbD succeed. The exploration projects are:Group 1: 93W many colors are there?
Are they distinct, or are some repeated?How are you going to be mare?
Group 2: Can you paint witieblack, or is it likethe paintbrueh not touching the painting?WM high and wide is the screen? Howmany colored blocks are there?
Geoup 3: How cony blocks (pixels) per second canthis =cuter draw?
Group 4: Can you fire a pattern (seed) belch, whenrun, 411 fill the entire screen with onecolor? Whet is the shortest or* you canf ird?
Group 5: Let's say you were allowed only sevenkeystrokes to make your seed. If I stucka piece of tape on the screen somewhere,can you make a seed which will zap a linethrough my =irked text? (It might goaround several times; that's ok.)
The second and part of the third days arespent giving each work group (two or three students)tea minutes access time to the caqouter, while theother groups catch or plan. The teacher :rovesabout observing, providing hints, and showingstudents low to pretend to be the ocmputersusinggraph paper and crayons. We turn the scarcity ofassurer access to our advantage here by requiringthe students to be careful and precise in theirhart-ainulation of the repeat-process in order topredict :bat their prcgrams will do.
During the third day the teacher deteridnes
which groups have caplets) their tasks and givesthem the next challenge sheet: the skill-buildingprojects. Since the first groups got easy explor-ation projects (and finisher] earlier), they aregiven tougher skill-building projects. Groupsthat haven't figured out their ecplorations by themiddle of the third day are helped to fine theanswer and roved on.
The object of skill-bsilding exercises is tobe able, to accurately predict what a given seedwill do. Two parts are given. In pert I, eachgroup is given a different list of three or fourseeds and asked to hand-sisolate the result andthen to try it. when they can do this assignmentcorrectly, they move to part II. Here, the work-sheet gives them the finished pattern, and they areto find the seed that creates it. In this exer-cise, the gang really Incases like a game. Movingpatterns are tensible, for instance, if the seedgoes basic over itself in the background color beforecontinuing. Again, because the students are havingto hand-simulate durino their off-machine time,skills are being built ladle students are planning=ewes. The teacher nqa-only give very shortmachine wows times during this phase, requiringthat a filled-out crayon simulation be presented asa ticket to allow its being tried on the ccmputer.
On the fifth day the class is given over to ashow-and-tell, in which each group explains km itfound the mow to its exploration question andshows off the most interesting seed. Polaroidphotos or slides of each group's best pattern aretaken for use in the et-of-course Parents' Dayed-Abit.Lesson 2: ACSBehavioral Objectives: Students will learn how tocontrol the screen tte-tle via individual amendsno the TWERP progran. They will explore the fea-tures of =ME:GRAPHICS and of teSTE, a neusic-outputfunction.Concepts: Students will learn the folleArsg:
1) the idea of using cateends that are weird*to get desired output. They should under-stand that connaais consist of operators(verbs) and (nouns or adjectives).
2) the conceptscperirent constructionand hand-simulation (which are introducedby Lesson I and reinforced here).
Resources:nINTERP, the PASCAL interpreter program.
MEM will be used throoehout the Munn., In this lesson, the insediate :rode will be
used; as each command is entered it isexecuted. The canaries used this week(listed below in Day 1 Evens) control theposition of an imaginary screen turtlewhich leaves a colored trail as it moves.A musicetaking canard is also explored.
2) Six small protractors.3) Six pre-made, but unmarked, cardboard
rulers long enough to reach diagonallyacross the Te screen.
Sequence of Events:bo the zirst day, the teacher shows students
2S3 4.
how to start INIERP and enter the immediate mode,then types the following six lines, pausing aftereach while the Class observes the effect:
PILISCFEEN (HIES)FILISCSEEN(BLACK)PBCOLORHereretmow (20)Itge (50)HOVE (20)These contaands are written on the blacklaard.
one sore emend is added, AVVE'IO(20,20), to beused if the line being drawn goes off the screen.Students are allowed two minutes each on the con -
pater to try these cowards. The concept of can -mends (operator, operands) is briefly explained.
Fbur discovery questions are then given to theclass:
I) *at is the size of the screen? Where isthe center?
2) Had Bony angles are there? What do nega-tive angles do?
3) How many colors are there? What are they?What happens when various eeneolors andtee/wound colors are used together?
4) Try the camend BYTE(nuater, timber) . Whatdoes it do? Oat is the effect changingthe first number? The second number? Whatare the legal ranges of_each Tarter?
Each group is given responsibility for one question.Tie groups get question 4.
Cat the second day, work groups attempt toewer their questions. They are given a descrip-tion of the concept of sheeting a dot, which theymay try when they have answered their questions.(They need those answers to sheet the dot.)
'lb shoot a dot, a piece of colored tape isstuck to the television. The students try to passa turtle track through the cot. Doing it with oneTURN and one WIVE is a successful shot. The teacherwill provide protractors and (blank) rulers andwill suggest that students make television oilers,but without telling them lame
After having deronstated the ability to shoota fixed det on the screen (beginning of Day 3), theteam will graduate to production geometric figuresof the Collaring types increasing in difficultyt
1) square2) rectangle3) isosceles right triangle4) equilateral triangle5) rectangle with diagonals6) trapezoid7) 5-point starB) Star of David9) "naked dandelion" (radial lines from a
cam= point)10) a short Mlle, in block lettersThe first five designs above will be attempted
by the teens on Day 3, with the first teem to thetezminal getting design 1, the second tarn gettingdesign 2 and working first at their desks with graphpaper, protrictors and rulers, and so on through allthe teens. On Day 4 or after they have mastered allof the first dive designs, the teens will tackledesigner 6e10. on Day 5, the class as individuals
Computer Games in Instruction 269
will he given IURITEGRAFH/CS-oarmands to produce adesign at abet* level S above and asked to draw thefigure on graph paper as it would appear when drawnby IVICETAWAPHICe.THE Ox IT& STYLE OF THE HSCS CURRICUILlet
At idusTeint-the reada may ask, lefty do youcall these activities games at all? What have yourlessons in cannon with Blackjack or lunar Lander?
The fact that unifies Quilting, TURTLEGRAPHICE,and more traditional interactive gazes is that theysell thenselves. No one has to canpel students todo their assignments.
The point at which our curriculum divergesfran closed genes is that the only real opponent intraditional games is a pseudo - random amber gener-ator or perhaps another Inman. In a cognitive gamethe onaenent is the rich structure of our ownignorance. The excitenent of being able to createpattern and order is as old as the wall paintingsin the caves of France. It is an essentiallyhewn activity, one at which all players can win.It is also a meta-gene; in which an infinite numberof specific games like shoot-the-dot can beexpressed. The relationship between cognitivegenes and traditional genes is analogous to thatbetween a set of blocks and a preassembled toy. Adifferent order of learning becomes possible.THE DEEM STILE C*? THE HSCS CURRICUILI4
The rundenetwardesign princWre have fol-lowed is to attewpt bomake.each lesson augment thestudents skills in three areas: discovery, con-trol, and design. We elide,/ students to play withthe* system as each new feature is introduced, butthe? have discovery questions Onee answers theyseek as they mass around in more or less structuredways. They need to find the answers to be allowedaccess to the next level of the system. Studentsdevelop discovery skills by experimentally answer-ing questions nee "what does this cameed do?"
Ile ask students to undertake specific chal-lenge, such as the shoot -tine -dot game, to developtheir ability to control the receputer by selectingthe correct cartrand and pzovieing correct valuesfor its operands. Their understanding of the systemis built by simulation exercises, which alive theeto predict the behavior of a catmand and ttete tochoose the right command.
Later in the sweeter, students will beginwriting prcgrantsr but even at early stages there isthe impetus to desi input sequences to producethe desired pate. Students must be able to pro-luoe a sequence of cayman& which produces the pre-natal output on first submission in order tograduate to the next level of the system.
Another principle we have followed can be sum-med up in the phrase "design from the first experi-ence." Mt believe that =touter science (or any-thing else) should be taught firm the inside out.That is, first experiences mist incorporate theheart of the trebter at hand, with as little ettrai.nexus matter as rossielc. For instance,teaches the frrlamental core of the computingexperiences in zepetitiotrif a =trolled processthere is greet weer, The Quilting lesson is taughtwithout introducing a word of jargon, pi vices
2S:3
270 NECC 1980
assigements, or complex camend sequences. alin-ing, and its fundamental message, can be taught toilliterates. The second lesson similarly teachesthe relationship between operands, operators, andresults. Only after students have firm operationalskill with a given tool, do we introduce terminol-ogy, written reference materials, and the ultimatelynecessary environmental details such as datadeclarations and control statements.
This paper has addressed the important ques-tion of "Om Vans" microcomputer education? Weare excited by the prospect of transforming gaming,a traditional problem area for computing teachers,into one of their primary tools.
Ihe authors aclarmaedge and appreciate theassistance of their collaborators: R. M. Aiken,C. E. Hughes, C. R. Gregory and J. A. Ross(University of Tennessee); L. Demarotta (H. C.Maynard High School); E. Miner (Alcoa High School).
REFERENCES
Aiken, R. M., Hughes, C. E., and Mbsbell, J. M.,"Computer Science Curriculum for High SchoolStudents," Proceedings Prbi/SIOCSE Conference,Kansas City, Montana, February 25, 1980.Papert, S., "Teaching Children Thinking," Proceed-ings WW World Congress on Computers and Education,Preswerdam, 1970.
Postmen, Neil, "The First CUrriculons ComparingSchool and Television," Phi Delta Kappan, 61:3,November 1979.
Figure 1: The Seed Pattern
.. ;',1-.-i- .1
- ., .--:-.;-.. -, .r , I
. o 13-42:-.--
, --r-T---.--111-1 ,l.e.I-1- , I'l ""."11-1-1-1- r r --1--1-; +-Ealmi-4
417'1- --i-7 17:::-.-t-rrrrizi-r.---i-n-r.' :- :- I l I .1-- i-
Around the time when computers beganappearing in classrooms and ceased beinga novelty, the theory arose that theideal way of teaching with a computerwas to put instruction in the form of agame. Actually, this teaching methodhas been around for years; pioneers inthe Dewey teaching method used manygames in their curriculum, and a numberof Montessori techniques also involvegames.
Probably one reason the game idea socaught and held the attention of educa-tors with computers was the mania forcomputer games in general. Really bril-liant students (and instructors) spentan inordinate amount of time designingand programming new games. Other in-structors and students spent a hugeamount of time playing these games. Whyfight it? Design instruction to takeadvantage of the built-in motivation;there is a pool of course developersanxious to develop instruction (i.e.,games) and a large audience of enthusi-astic students (players).
What Happened: Expectations Denied
With a few exceptions, the resultswere notably disappointing. Problemsarose on two fronts when games were madean integral part of computer-aided in-struction:
The game quickly took overthe instruction and frequeAtlybecame more important than thecontent of the lesson.Language-oriented, subjects
required a huge amount of pro-gramming effort to bend to an
effective computer game. Humanlanguages and thought processessimply are not easily trans-ferred to a computer, except onthe simplest level. (Note theword "easily " - -there are bril-liant exceptions to the abovestatement.)
The pendulum recently has appearedto swing in the opposite direction:games are now anathema and viewed as afrivolous pastime at best. But thisapproach ignores that games are immense-ly popular and have a tremendous at-traction for game players and designersboth. This paper proposes to extractsome of the good points about games andthen apply them to real computer-aidedinstruction.
Salvaging Valuable Parts
What points are worth saving? Whatcan games do that other methods of in-struction do not do as well? The obser-vable characteristics of computer games(and computer game players) are:
A high level of motivation.Game players and designersspend hours working at theterminal.Clear and consistent goals.
All true games have a clearending in mind. How much in-struction becomes bogged downbecause of muddy goals andobjectives?A high amount of player in-
teraction. Game players aredoing something: they aremanipulating the terminal, thecomputer, and the game struc-ture itself.
284
272 NECC 1980
Maximum choices for the Play-er. Studies of children's toys'Rive shown that the flashiesttoy is not the one that isplayed with most often. Thetoy that gets the most atten-tion is the simplest one thatallows the child to use imagi-nation. The most popular games(on or off the computer) arethose with the simplest cC1.-
structs and widest range ofchoices.Simplicity. How much in-
struction fails because thestarting sequence for the stu-dent is too complex? Complex-ity in itself is not the hor-ror--the problem is that no onebothers to explain all the de-tails to the user. Most goodgames come with very detailedinstructions. (Note: why doesthe programmer who balks atproviding documentation churnout reams of instructions forthe game he just designed?)Creativity. Many of the as-
pects of computer games-- wheth-er they are traditional or in-vented by the user--have highlycreative parts. An axiom amonggame designers that is fre-quently used to justify theirinterest in games is that someof the most highly creativeideas and programs are tirstdeveloped in a gene.
Applied Gaming Principles
How can the principles from games beapplied to instruction? It turns outthat the items that make games so ap-pealing are intertwined, and by incor-porating several facets of games intogenuine instruction, several objectivescan be achieved at once (e.g., high mc-tivation, strong interest,. good inter-action).
The remainder of this paper is achecklist of items to look for in anycomputer-aided instruction. These areitems which have been found to most re-liably, increase the quality of computerinstruction and, it is hoped,' the skillsof students taking the course. An im-portant point about this list is that itIs not meant to be exhaustive--it is astart. Persons wanting to use the ques-tions as a checklist may also have theirown principles and procedures to add tothe list.
Questions for the Developer
Is the method of presentation con-sistent with the material? The bestcomputer games require player actionsthat fit the total concept of the game.Players move tokens by giving directionsto the computer in a manner similar topicking up the piece by hand (for boardgames). If the game is a thought-typegame, the player makes decisions thatare consistent with real-world decisionmethods. The computer in both instancesis a referee, informing the player ofthe consequences of his actions.
Poorly designed instructional gamesor simulations use methods of advancingplayers (students) that are not relatedto real actions. In a number of compu-ter board games, students race cars,horses, or other items around a track.But instructional racing games make thepieces move around the track by answer-ing questions. These games are not no-tably successful because the action istoo slow and because races are not nor-mally run by answering questions. Thisdesign error occurs because the devel-oper has inferred that actions that gameplayers enjoy in ale setting must begood in all settings. This mistake re-sults in a product that is neither goodinstruction nor a goo'' game. Lesson de-velopers can avoid this trap by chosinga presentation method without regard tothe observed popular appeal of a methodin a different setting. A natural, orconsistent, presentation enhances theappeal of a lesson considerably. If thesubject is math, the game should use ma-thematics naturally, not as a means ofmoving a token. The Minnesota Educa-tional Computer Consortium has developedeconomic models that allow children torun simulated businesses (lemonadestands, bicycle factories); both mathe-matics and economic fundamentals aretaught through these games. Englishteachers can use word processors avail-able on most computers in their clansen.Students take naturally to word proces-sors--they enjoy seeing their workturned out in a professional format.There are many other examples--the keyis to ensure that the presentationmatches the material.
Is there a large amount of playerinteraction? f'requently the chargerated against television can also beraised against education: the studentsits and absorbs the material passively.The lecture format certainly has itsplace, and brilliant and stimulating
lecturers are rare individuals thatshould be sought after. But computer-aided education as a lecture or elec-tronic page-turning format usually doesnot rise to these heights--and is also abad use of the medium.
An examination of the most populargames shows a high amount of player in-teraction. This interaction is not sim-ply key-pushing; some of the populargames are also limited to single-keyresponses after lengthy actions by thecomputer (football, adventure, road-race). But the player is constantlythinking -- decisions must be made, pathschosen, strategies worked out.
Again, an objection will be raisedthat the material in education cannot beadapted to such a lengthy decision-making format. But an examination ofthe actions that students take whenstudying shows several possibilities forincreasing the student's participation:
1. 'ath choice. When a stu-dent Tritairii away from theclassroom, he at least has theoption of choosing which pageor chapter to start on. Why notinclude this choice in the com-puter session?
2. Notes. It is relativelysimple to provide an electron-ic scratch pad--each studentcan be provided with a commentor note option during the les-son. At the and of the lesson,the student can either storethe notes or (if disk space islimited) take a printout of thenotes away.3. More inquiry. This does
not mean more questions. In-quiry can be a form of path-taking, or an invitation tospeculation. The studentchooses which subject to takefirst. Incidentally, this alsoanswers the frequent questionof the correct order to studysubtopics. The student canchoose the topic and also skipfrom topic to topic if neces-sary. The same material iseventually covered, but alongpaths that the student has cho-sen.
Is the material creative? Onepuzzling problem in computer-aided in-struction is that the same individualsthat produce ho -hum instruction also areturning out brilliantly designed games.Obviously the same effort is not going
Computer Games in Instruction 273
into both projects. This problem is amanagement, not a computer, problem.There can be several reasons for thisdichotomy when it appears:
1. Freedom to experiment andmake errors. If you insist on
-)r-fand perfect instruc-tion from your developers, thatis exactly what you'll get- -with all the creativity and or-iginality of the Saturday-morn-ing cartoon shows. A perfectlesson is not necessarily agood lesson. In fact, lessonsthat have incomplete sectionsmay even be more useful, sincethe student must fill in themissing portions.
2. Appropriate comments atappropriate times. Kibitzersseem to abound in educationaldevelopment areas. The discus-sion of approaches and criticalanalysis of lessons under dev-elopment is an important partof the creative process, butthere are many cases where apromising bit of instructionhas been discarded becauseconstructive criticism washeaped upon the material--be-fore it was developed and be-fore advice was requested.While a developer is writing agame as a hobby, no one isleaning over his shoulder.Leave the developers aloneuntil there is something tocriticize other than a firstdraft.3. Responsibility. Managers
frequently lick confidence inthe persons working under them,and hope to avoid errors bygiving very detailed instruc-tions to the developers. If adeveloper is presented with acut-and-dried package, there islittle room for a new and crea-tive approach. For someone tobecome truly involved with aproject, she must feel that theproject is at least partlyhers, and must have some in-volvement with planning. Andresponsibility extends down tosome very low levels. In alanguage arts project in Nor-folk, Virginia, the computeroperators were employed toenter some of the questionsstudents would be given. Theywere asked to use their imagin-
2S
274 NECC 1980
ations when entering the praiseresponse students would receiveon correct answers. The re-sults were a series of ques-tions with highly original re-sponses that the students couldrelate to. The operators alsotook a personal interest in thedevelopment of the course, andfrequently monitored studentprogress and asked to changeLessons that were not popular.
Is everyone using the material en-thusiastic? Bete is another factor thatis a function of the people involved andnot the computer. 'You WILL be enthusi-astic:" is obviously an unworkable ap-proach (but one which is sometimestried:). But if everyone involved withthe development of a particular piece ofinstruction finds it tedious and diffi-cult to work with, then perhaps the en-tire instruction set should be lookedat. If persons who are working in theirsubject area find the topic boring, thenwhat about the students? What is neededis either a fresh and creative approach(see above) or a different topic. If asubject is completely boring, then itsvalue is open to question. Tepid topicsusually result if the individuals cansee no value in them whatsoever.
Enthusiasm, once engendered, tendsto be catching. Enthusiasm most fre-quently stems from freedom and responsi-bility--two subjects discussed undercreativity and maximum choices, above.Freedom and responsibility extend to alllevels: supervisors, developers, tea-chers, and students. Teachers and ad-ministrators tend to be wary Of allowingstudents to take part in the teachingprocess, and for good reason: studentfreedom is difficult to manage and con-trol. Real participation in their owneducation is also a novel idea for moststudents; they have been passive consum-ers of instruction for so long that theymay not be able to handle such a respon-sibility without careful pre-training.But the results can be spectacular.
Here are some areas in the instruc-tional process where personnel at alllevel: can directly participate:
1. Level of instruction. Toohard? Too easy? Should it bepresented to a different gradelevel or in a different subjectarea?
2. Method of presentation. Isthe format (question and ans-wer, screen display, essay andchoice) appropriate for thesubject?3. Effectiveness. Does the
content of the lesson stay withthe student?4. Correctness. Is the con-
tent free of grammar, syntax,spelling and content errors?(Here is an area where youngerstudents delight in showingoff. If you choose to releasethem on this one, prepare foran avalanche.)5. Additional lessons. What
other material shodIrle addedor changed?
6. Design of instruction it-self. Developers can work withMir students, or students whohave just completed the course,to redesign and evaluate theinstruction.
Is the course simple and direct?One fault of some of the indiVrecourses is that they tend to be immense-ly complex. While at some time in thefuture humans may learn to think recur-sively and in complex algorithmic pat-terns analogous to the machines theyuse, presently people still think in thesame old way. Presenting highly complexstudy structures, maps, objectivefields, and learning paths can set up aforest that quickly discourages even themost determined students.
If the structure is too complex,there are three solutions:
1. Check the presentation. Incomplex structures, a simple,direct path can frequently befound in the material.
2. Turn routing over to thestudent. If given a list oftWand objectives, the stu-dents can make their ownchoices of what to study first,and when.
3. Ride some of the complex-ilx. taiTrUistructure maybe-due to the enthusiasm of thedevelopers for marvelous struc-ture. A lot of the design canbe safely hidden by having thecomputer do the routing work.
Great advances have been made in theuse of computers in classrooms in thepast few years, and the pace of develop-sent and discovery increases with each
2 =,
Computer Games In Instruction 275
new application. The procedures dis-cussed in this paper are a startingpoint for developing better computer-aided instruction. The key to success-ful computer-aided instruction, as toany other form of instruction, is thepeople involved in the project. Thereis no substitute for a creative, dynam-ic, and highly qualified individual inthe right position at the right time.Fortunately, these persons appear to beavailable because computer-aided in-struction has made great gains in recentyears. The continuing analysis of whatconstitutes good technique in the use clcomputers will continue this trend.
Computing Curricula
AN EDUCATIONAL PROGRAM IN MEDICAL COMPUTING FOR CLINICIANS AND HEALTH SCIENTISTS
Albert HyblDepartment of Biophysics
(301) 528-7940
James A. ReggieDepartment of Neurology &Department of Computer Science(301) 528.6484
UNIVERSITY OF MARYLANDSchool of Medicine
Baltimore, Maryland 21201
ABSTRACT
The growing importance of computers in medi-cine implies that health professionals lacking ru-
dimentery knowledge of their potential use will behandicapped in the future. In this paper we dis.cues the content and implementation of a curriculumin medical bomputing that we have introduced at ourmedical school to remedy this problem. In additionto formal classroom instruction our proves inchides such innovative features as a demonstrationlaboratory for exhibiting computer applications inmedicine, the use of a family of knowledge mange.sent languages that are designed for computer -inexperienced clinicians, and the involvement ofinterested individuals in ongoing researoh in nor-puter applications in medicine. Our description ofthis curriculum and its implementation should be ofinterest to anyone involved in tesohing computer.inexnerienced individuals about the potential usesof computers.
INTRODUCTION
In recent years the role of computers in medi-cal education has been receiving a great deal ofattention. The vast majority of this interest hencentered on oomputer.aided inatruotion for a wide
range of olinioel and pro.olinical topics (3, 4, 8,9, 14). At the present time many medical schoolsuse such autcsated instructional material tosupplement conventional teaching methods.
In this article, however, we will discuss acomplementary aspect of computers in medical educa-tion: introduoing medical students, prastioingphysicians, end other health-related personnel tothe applications of computers in medicine. Iduoat.ing individuals in the health professions about
278
medlOal computing has recently been singled out asan area of great importance that deserves more at -tentiOn than it hes received (7, 12, 18). Verylittle literature has addressed this gojl and therehas been no clear consensus about which aspects ofcomputer science should be taught to medical stu-dents and physicians. How this material should betaught and how it should be introduced into analready information.dense medical education processhave been given little if any attention.
It is our belief that the growing influence ofcomputers on medical care makes some familiaritywith their uses and potential important to healthprofessionals. We have thus designed and imple-mented a program tc supplement conventional medicaleducation with instruction on medical computing.This program has two major OnjnotiVen:
(1) to provide health professionals with a broadunderstending of current and potential usesof computers in medical researoh and clinicalpractice; and
(2) tc give health professionals the ability todirectly use available computer reeouroea forsolving clinical and researoh problems.
In this report we begin by out ..ing the con-tent of a curriculum on computer applications inblomedisine that we believe achieves these objeo.Elves. he then disouse how we have approached theimplementation of this program at our institutionand the response it has generated aeons medicalstudents and fatuity members. Finally, our expert.
*noes are reviewed in the context of other resent .
eadOevors in this,area.
2Q9
A MEDICAL CONFUTING PROGRAM
The content of our medical computing programreflects our feelings about ghat topics are cur-rently important or will become so in the future.03r 00100 of topics is based largely on modioal(imputing systems that are presently used and a re»view of recent literature on biomedical computingresearch. Our emphasis is on the use of the com.voter as a tool to accomplish various tasks ratherthan the fundamental conoepts of computer science.The materiel in our program can conveniently be di-vided into five sections:
O General ConceptsO Soientifio and Laboratory ApplicationsO Clinical ApplioationO Educational ApplicationsO Administrative Applications
The section on general oonoepts is tailored toincrease student awareness of the oapabillties ofcomputers and to provide then with the fundamentalknowledge needed to use a computer. It includes:
O Essentials of interactive Computing* In ord-er to entourage students to become computerusers, we begin by explaining the mysteriesof account numbers and password*, the tech.niques for signing on and off, the idiosyn-crasies of using certain terminal*, the mar-vels of the executive system, and how to savedata and/or progress in retrievable files.(Batch processing or programs is disoussed
but not used by students.) .
O Document Prooesaing and Text Editing: Thesesystems are easy to learn and iamediatelyuseful to the students. Their mastery faoil-itates the learning and use of other oomputerresources.
O General Information Processing: Programminglanguages for intonation prooessin4 and forgeneral problem-solving are surveyed. Guidelines are presented for selecting an seem-priate language for various tasks.
O An Overview of Applications: An initialoverview of the range of scientific, labora-tory, alinioal, and educational applicationof computers in medicine is given. This in.eludes SISOSSSIOO of boil working systemssuch as a whole.body CAT scanner; a cranialCAT seamier; two nuclear scanners; computer-ised baste solemn research laboratories; theQ8.2 System for ICU patient data colleetion,storage, and mitering (at the MarylandInstitute for Emergenoy Medioal Services); apartially implemented PROM'S system (17) (atthe Baltimore Cancer Research Center); and
the Stroke Database Center (located adjacentto the Neurology Mari, this unit serves as anentry point to the sultioenter Nations'Strike Database OM.
Computing Curricula 277
O Social and Legal Issues: A discussion or thepresent and future effects of computers onthe practice of medicine is given anoeludes material on ethical issues, standard-ization, and the economics of medical comput-ing. The Privaoy Act of 1974 (PL 93-579),the Federal Brooks Bill (FL 89-306), and fed-eral information processing standards areexamined.
The section on scientific and laboratory ap-plications contains the following material:
I Real-Time Instrumentation: Configurations ofmicroprocessors and minicomputers for dataacquisition ranging from stand-alone dedi-cated facilities to hierarchical computernetworks are discussed. Techniques for col-lecting and communicating data to the oosput-
er are stressed.
O Graphics: The use of terminals with graphicscapabilities and plotting systems is ex-plored. Their ability to extend the effec-tive use of computer resources by improvingthe presentation of certain types of informa-
tion is covered. The plotting of resultsfrom the digital simulation of a biophysicalsystem is used as an example of the suppor-tive role of graphics.
O image Prooessing: The role of computers invarious clinical scanning systems (e.g., CAT,nuclear) is discussed. Attention is given tothe tuture..pntential of image analysis in thelaboratory (e.g., ohromoscse analysis, celloountins).
Statistical Analysis of Dates The basic cm-oepts of data analysis are introduoed andavailable packages for statistical processingof data are desoribed (0.g., SPSS, MOP).Example applications :mob as epidesiologicdata analysis Sr. provided.
The seotion on clinical applications coverstopics dealing with the use of computers in patientoarc and clinical research including*
O The User-Computer interface: The use of com-puters for obtaining s patient's history andthe problems of making computer facilitiesdireotly usable by physicians are explored.A brief introduction to natural languageprocessing by aaohine is provided.
O Medics' Information Systems: This inoludesthe use of registries, data bases, and hospi-tal information system for storing and re-trieving manic:al records for use in patientcare and clinical researoh. Aspects of seourity and integrity ere examined in detail.The oonoept of a query language is introducedand ourrently available systems are reviewod
(e.g., COSTAR, PROMS, ARAMIS).
290
278 NECC 1980
Computer-Assisted Medical Decision-Making:Tne potential uses of computers for assistingphysicians with diagnosis and patient manage»sent are discussed. Concepts of knowledgerepresentation and alternative inferencemethods (s.g., statistical pattern classifi-cation, production rule systems, cognitivemodels) are covered.
The section on educational applications coversmaterial relating to any aspect of the health pro-fessional's training. It includes:
O Computer-Aided Instructions The availablesystems for computer -aided instruction (e.g.,PLATO, ASET, MEN) are reviewed and informa»tion on how to author suitable lessons ispresented.
O Education in Medical Computing:. The atti-tudes of health professionals toward usingthe computer for medical applications areexplored. The concept of a curriculum onmedical computing and a survey of existing orproposed programs are presented.
Referent:. Systemss Computerized literaturesearches (e.g., MEDL1NE) and more advancedmedical knowledge bases are described.
The section on admiestrative applicationssurveys the use of computers in hospitals end phy»sician offices for scheduling, accounting, analysisof polioy-related issues, and other related tasks.
CURRICULUM IMPLEMENTATION
Currently we have implemented formal course-work that encompasses most of the material outlinedabove. This coursework began with only a singleone»credit minimester course that covered essen-tially all of the material listed above at a fairlyeuperfioial level. We have recently added a secondone-credit @animater course and divided the ma-terial between the two courses permitting it to bepresented in greater (loath. Although both coursesbegin brfatiliarizing students with the essentialelements of interactive computing, they are largelycomplementary in what they cover. One course ad-dresses general concepts, scientific and laboratoryapplications, educational applications, and ad-ministrative applications. The other course Con-centrates on clinical applications. Both coursesOre offered twine a year and are taught in a coor-dinated fashion. Students are permitted and en-couraged to take both courses.
Since biomedical computing has not been gener-ally recognized as pert of the medical curriculum,our approach has been to gradually introduce this;material into the medical school program. lie feel
that the development and evolution of our programhas profited from an incremental implementation.The gradually increasing amount of ooursework ap-propriately parallels the amount of computer use by
the medical community.
To maximize the student's experience, we naveadopted a teaching approach that uses severaldifferent instructional methods. Much of the ma-terial is presented as formal lectures that arecomplemented by related reading assignments from
the current literature. The use of lectures andreading assignments allows us to cover the widerange of concepts outlined above in the brief timeavailable. However, to give mediCal-students amore intimate acquaintance with the impact of thecomputer on health care and medical research, wehave supplemented the traditional teaching approachin three ways.
First, students are made cognizant of the di-versity of applications through tours to on-campusfacilities where dedtcated computers are in activeuse (the sites are listed above under "An Overview
of Applications"). These visits effectively empha-size the present role of computers in the healthprofessions and contribute to student motivation to
learn more about biomedical computing.
Tne second way we are supplementing classroomteaching is through the development of a laboratoryfor the demonstration of applications of computersto specific research and to clinical and education-al problems that are not currently in day -to-dayuse at our hospital. EKO analysis (currently beingimplemented) is especially suitable for demonstrat-ing the major components of laboratory computing:data acquisition (analog - digital signal process»ing), feature extraction (R -wave recognition, ONSTclassification, power spectrum analysis), anddecision- making (presence or absence of myocardialinfarction). Several computer-aided decision-making systems are complete (e.g., diagnosis ofthyroid disorders and stroke; prediction of outcomefollowing cardiac arrest) and are used to illus-trate the current level of state-of-the-art medicaldecision support systems. Feedback from medicalstudents and physicians is useful for guiding the
development of the experimental systems that arebeing demonstrated.
The third aspect of our instructional approac:
involves hands-on experience using the computer fa-cilities at our university. In one of the mini -
wester courses students are expected to use theDocument Processing System to prepare a critique ofa current journal article of their choloe. Thestudents are also given projects involving digitalsimulation, computer graphics, general problemsolving, and using statistical packages. In theother course that covers mainly clinical applica-tions, students are asked to develop two smallcomputer -aided deoision.making systems using a fanily of knowledge management languages designed foruse by computer - inexperienced individuals (15).The systems supporting these languages provide aoomplete computer -aided decision - making system whengiven a description of the problem to be solved andtha relevant knowledge noeded to solve it. Dy
Ins the computer for these projects students are
able to learn directly about the capabilities andlimitations of the computer for assisting them witha wide range of tasks.
At the present time our elective minimestarcourses are reaching about 5 10% of the freshman
and sophomore *lasses. Students taking these
courses have a wide range-of ccaputing backgrounds,varying from essentially no previous computerscience to an occasional student with an undergrad-uate degree in a related field (e.g. biomedicalengineering). The variety of ihstruotional ap-proaches used has been fairly successful in inkingthe material presented understandable to cosputerinexperienced students. At the sue time the broadscope of the material has provided a challenge to
even the most advanced students. Student evalua-tions of the courses implemented so far have beenessentially positiv4 other faculty umbers andu niversity administrators have begun to take aninterest in the program.
DISCUSSION
Previously proposed or implemented programsfor education in medioal =outing fell primarilyinto two ostegories. First, and most common, arethose representing complete curricula that consistof interdisciplinary studies in medicine and com-puter science. These ourrioula are designed toproduce individuals who are specialists in biomedi-cal computing and thus typioally had to graduatedegrees. Several examples of such ourricula havebeen described (1, 6, 13, 19) and a resent reviewof relevant progress in the Federal Republio ofOermsny has addressed this issue (10).
The seoond category of educational progress inmedical oomputing include those consisting of onlya single *Ours* designed to supplement the educa-tion of health professionals (e.g., nurses, plume-elate). For the soot part, this courses have coo-*red a very limited range of topics (2, 5, 16).
The medical computing ourriculum that we havedescribed ih this report lies in between the twodifferent oetegorles outlined above. On the oneband, it is designed to supplement the traditionaleducational experience of medical students and oth-er health professional' stellar to the singlecourse programs. On the other hand, it consiats ofmore than a single course and oovers a wider rangeof topic". The depth to whioh these topics areoovered is gradually increasing es the availableoouraework 'woes's.
In this report we have attempted tc desoribe
n ot only what the content of a medical °elutingcurrioulus should be, but also how it night beimplemented within the constraints of the tradi.
medioal school experience. In particular,we have presented our belief ih the need for anincremental implementation and an argument forsupplementing oomventional lecture material with
Computing Curricula 279
hands -On computer experience, visits to relevantfacilities, and the use of a demonstration
laboratory.
Our present plan is to continue the evolutionof our minisester elective program through addi-tional courses and by instituting a more formalevaluation of their success. The demonstrationlaboratory needs further work; we need to developand acquire additional software and hardware.Long-range goals include the establishment of medi-
cal student externships in advanced topics in medi-cal computing, integration of at least some form ofrequired instruction on biomedical computer appli-cations into the standard medical ourriculum, andthe creation of a continuing medical educationcourse to reach practicing physicians.
Atanaldikilltlitat he thank the Departments ofDiagnostic Radiology, Pharmacology and ExperimentalTherapeutics, and Radiation Therapy, the Marylandinstitute for :urgency Medical Services, and theStroke Databank for graciously accommodating ourclass tours of their facilities. Computer time forthis project was supported in full by the ComputerScience Center of the University of Maryland. Dr.
Reggie gratefully ackLooledges support for thiswork from an NIB Teacher-Investigator DevelopmentAward (5 KO7 NS 00349).
REFERENCES
1. Ackerman, L. V. & Harris, D. K. (1975).rohitecture for a Graduate LevelEducations/ Program in the Area of ComputerSystems in Nedlotne,wAuggproopentwol sziMagma Lansaw cont*ranat VP 765468'
1. Buokwell, L. J. (1979). *Education of HealthCare Students to Acoept and Use the
Cominutar, ItHOgullnii Set Ina 3tst., 1280111111Soimakoz 1881liaLlos An Manisa Sam
edited by Dunn, R. A., 206-209. Nov York*IBEB.
3. Dora, J. G, Ne;4an, A., Nasmond, V. E. 6Haynie, C. (1978). "Computer-sidedInstruotion in Clinical Neurology,* journal
sitilesilaalignaatia.., 23(8): 693.
4. Deland, E. C. (1978). IducgatiggIgghgalgggAn AIWA &WM llibioAtutog New !oft: PlenumPress.
5. Dumas, B. P. (1979). *A Computing Course forPharmacists,* lamunliana sit ant .3rsL.Zionntan an Smut= laalinattaa jii Salm/SAKI, edited by Dunn, R. A., 210.213. NewYorks IEEE.
292
280 NECC 1980
6. Duncan, N. A., Austing, R. H, Katz, S.,Pengov, M. E. & Wasserman, A. I. (1978)."Health Computing: Curriculum for an EmergingProfession, Report of the ACM CurriculumCommittee on Health Computing Education,"
Pr dinar of Ing eGli 1.921 &lianaConference, 277.285. New York: A.M.
7. Duncan, K. A. (1979). "Educational Programsfor Health Care Computing," Prooeedinen sit
the 3rd. =min an Daunt= lantiantina Ingattratfarg, edited by Dunn, R. A., 204-205.New York: IEEE.
8. Halverson, J. D. & Ballinger, W. F. (1978)." Computer- Assisted Instruction in Surgery,"Surgery, 81(6): 633-638.
9. Kenny, G. N. & Schmulian, C. (1979)."Computer-Assisted Learning In the Teaching
of Anaesthesia," inalabgaii, 352): 159-162.
10. Kocppe, P. (1977). "Education in MedicalInformatics in the Federal Republic of
Germany," of Infarnalina In Mining,16(3): 160-167.
11. Kunitz, S. C., Havekost, C. L. & Gross, C. R.
(1979). "Pilot Data Bank Networks forNeurological Disorders," prooeedluna slum
IYERoilum an raw= Anginal= InMedical fare, edited by Dunn, H. A., 793-797.New York: IEEE.
12. Levy, A. R. (1977). "Is Informatics a BasicMedical Science?" Yroceedlnea.MC1101=2.7..edited by Shires, D. B. & Wolf, H., 979-981.New York: North Holland.
13. Levy, A. H. & Chen, T. T. (1977). "Plana fora Program in Medical Information Science,"
14. Madsen, B. W. & Bell, R. C. (1977). "TheDelopment of a Computer - AssistedInstruction and Assessment System inPhareacology,"Mdicalsdueation, 11(1): 12 -20.
15. Reggie, J. A. (1980). "A Domain-IndependentSystem for Developing Knowledge Bases,"
Iracanslinna Ord.] Wang Sanatanneaf Ma amain And= lac faunatntlanaMalin of latsaLlinannt, victories B.O.'Canada. (in press).
16. Rowley, B. A. (1979). "Computer MedioineEducational Program for Medical Students and
Others," Proceeding ithi 3Eg, $ ignaly. anSamna Anoliettian In Mina am editedby Dunn, M. A., 214.216. New York: IEEE.
17. Schultz, J. R. & Davis, L. (1979). "The
Technology of PROM1S," ISSE/tooeedIngn,61(9): 1237-1244.
18. Shannon, R. H. & Duncan, K. A. (1978). "Whya Curriculum in Health Computing?"hsasslinta of Ihn Ardi 142 liatinnaSPererena, 273-276. New York: ACM.
19. Wasserman, A. I. (1979). "EducationalPrograms in Medical Computing Their Basisand Implementation," hstnesunna,g,m3rg,Asonalui fagot= Andinatinn In MedicalSlign, edited by Dunn, R. A., 217-222. NewYork: IEEE.
2° r)
A SECONDARY LEVELCURRICULUM INSYSTEM DYNAMICS
Nancy RobertsLesley College
Graduate School of Education9 Mellen Street
Cambridge NA 02138617-547-8844
Ralph N. DealChemistry DepartmentKalamazoo College
Kalamazoo MI 49007616-393-9452
SYSTEMS THINKING IN EDUCATION
For quite some time, the influenceof science on society and education busbeen primarily reductionist, seeking tounderstand phenomena by detailed studyof smaller and smaller parts. Over-emphasis in this direction has led to aseparation of allied wciences like physicsand chemistry. "Learning more and moreabout less and less" (Odum, p.9), as onecritic described reductionism, has takenplace with almost the complete absenceof a movement in the opposite directionto integrate knowledge. Education tendsto follow the pattern of parts withinparts. Knowledge is divided into dis-ciplines, and disciplines are furtherdivided into subjects. The student isseldom exposed to materials which seek to_integrate the various disciplines andsubjects.
Education ar well as the everydayworld of social and economic analysis hasa pressing need for synthesis and holism.The systems approach seeks to understandhow parts fit together to form a wholethat functions for a common purpose. ItIs in the world around us that moststudents will be making decisions for therest of their lives. Explicitly understanding the nature of these systemsfrom a structural and a behavioral pointof view constitutes a most relevant anduseful education.
THE APPROACH: SYSTEM DYNAMICS
System dynamics is a method for under-standing and managing complex social systems.It is built on traditional management (infermmotion, experience, end judgment) and feed-back theory or cybernetics (principles of
Computing Curricula 281
structure and selection of information).Mathematical models of varying degrees ofcomplexity are built to reflect the systemout of which a problem grows. Computersimulation is used because behavior implic-it in the structure of complex systems istoo involved to be solved by direct math-ematical methods. Model building and com-puter simulation is therefore an integralpart of the system dynamics problem-solvingapproach.
System dynamics starts from the preetical world of observation and experience.It does not begin with abstract theory noris it restricted to the limited informationavailable in numerical form. Instead ituses the descriptive knowledge of the oper-ating arena about structure, along withavailable experience about decision. making.Such input* are augmented. where possible bywrLtten description, theory, and numericaldata. Feedback principlee are used toselect and filter the information that givesthe structure and numerical values.
Systemwhen Professor Jay W. Forrester of UIT pub»lishe :ta 6:Industrial
1961
field of system dynamics broadened in itsapplications, it also broadened in itsstudent audience. Initially taught as agraduate level course at MIT, it has spreadto the undergraduate curriculum in manyuniversities. In 1975, ae a dissertationproject, Roberts introduced the basic con-cepts to fifth and sixth grade students(1978). The evidence that fifth and sixthgrade students could understand and applythese basic concepts lad to the developmentof the warrant secondary level curriculumproject being funded by the Office ofEducation. The project is based at Lesley
294
282 NECC 1980
College Graduate School of Education,Cambridge, Massachusetts, under thedirection of Nancy Roberts.
A SECONDARY LEVEL CURRICULUM ZN SYSTEMDYNAMICS
The current curriculum developmentgroup is writing and pilot testing a setof six self-teaching, introductory learningpackages. This material will make it possi-ble for a teacher in any discipline tointroduce model building and computersimulation as a ptJblem-solving method,
The titles of the six learningpackages are:
I. Basic Concepts: DynamicProblems, Systems and Models
II. Structure of - Feedback SystemsXII. Graphing and Analyzing the
Behavior of Feedback SystemsIV. Understanding Dynamic ProblemsV. Introduction to Simulation
VI. Formulating and AnalyzingSimulation Models
An overview of the content of eachlearning package,follows.
I. Basic Concepts: Dynamic Problems,Systems, and Models
Embarking on a course of study insystem dynamics, students need to under,stand the basic concepts that underliethe field. Three concepts are central;the focus is on dynamic problems; theintent is to consider the whole systemof interacting parts from which a problemarises; and models are explicitly employedto carry out the analyses. Bach of theseconcepts is explored briefly in learningpackago I to provide a foundation for thestudy of feedback systems which follows.
A problem is dynamic if it changesover time. Urban crime rates, for example,rise; the economy cycles, as do pendulumsand people's feelings of depression andelation; central city populations decline,and so do reserves of natural resources:
A system is defined as a collectionof parts operating together for a commonpurpose, but the concept is sometimesbetter left undefined to be inferred frog;examples. The notion should connotecomplexity, but it should also suggesta wholeness of perspective and the 'feelingthat the whole is greater than the sum ofits parts.
A model is a representation»usuallya simplification--of some slice of reality.Pictures, verbal descriptions, graphs, setsof equations, and laws are models. Think.Jug can be characterized as the manipula.-tion of mental models; the real system isnever in one's head. Such a concept mayseem too generalized to be useful, but
11
realizing the central role models play inour thinking (especially in the scientificmethod) allows asking the, proper question-not whether to develop a model, but whatkind of model is most helpful for a givenpurpose.
System dynamics helps us to under..stand problems arising in dynamic systems,first by making our mental models explicit,second by incorporating into them feedbackrelationships (learning packages If, III,and IV of the curriculum), and third byproviding means for developing them intounambiguous mathematical models (learningpackages V and VI) when the complexitiesbecome too great for mental models tohandle.
II.Stucture of Feedback SystemsThis part of the sequence introduces
a way of thinking holistically abortcause and effect. The principal toolemployed to analyze a system is the causal,.loop d.agram. Cause-and-effect relation-ships are symbolized with arrows, formingchains of causal links. Loops result whensome or all of these chains return totheir starting points. These loops ofcausal influences are called feedbackloops; they are the central focus andfoundation of system dynamics.
Causal»loOp diagrams are visualmodels of feedback systems. In partof the curriculum they are used in a gonerally descriptive way to summarize the com-plex interactions in a variety of stories,problems, and systems with which the stu-dent is fel:tiller. The following is one ofthe storiea from this learning packagefollowed by the possible csusalrloop die.grams that students might develop.
MS OIL CRIS2S
2():;
One aspect of the oil crisis, asexplained by an economist, was the startingof a vicious circle. This vicious circlewas begun by agreements made by the Araboil-producing countries in 1971 called theTeheran and Tripoli Agreements (named forthe cities in which the meetings were held).Here these countries agreed to raise theprice of oil. The rise in oil pricesmeant that these countries made more money.They made so much more money that theycould not possibly spend it all. Real-,ising this, these countries decided not toproduce as such oil. They knew that eventually their oil supplies would run out sothey might as well make thelplast as longas possible.
Because there was less oil beingproduced in the world and more oil wasneeded every day, a scarcity developed.This scarcity of oil forced the oil pricesto go up even higher, continuing thevicious circle.
The following causal-loop diagramexplain he economist's vicious circletheory.
dOil Produki) of Oil
refits 4.
Figure 1. The Vicious Circle
In causal-loop diagramming, as inmultiplicatin, two negative elementscancel each other out, creating a positivefeedback effect.
The diagram below expands on theinformation given in this story and mightresult from a class discussion.
Amount of+ t Amount
Oil Left Oil Price of 40Used byIn Ground Pr cord Oil Cons eers
...
I.JAPProfits
Figure 2. The Oil Crisis
The principal skills the students willdevelop in learning package II are theability to represent the essential Le»finances in a problem or system as acausal-loop model and the habit of search.Leg for feedback influences which closecausal loops. if students couplet* thispart of the curriculum but do not cony.tinue on in the sequence, they willhave ,gained a powerful tool for under,standing complicated interactions, butthey will have only begun to understand
Computing Curricula 283
the connection between feedback loops anddynamic behavior.
M. Graphing and Analysing the Behaviorof Feedback SystemsDynamic behavior of variables in
systems is the focus of this learning pack-age. The purpose is to understand how thestructure of feedback loops is responsiblefor the behavior of variables in the systemover time. Graphs are introduced and usedintuitively, often without specifyingnuaegical scales. The amte of the graphof a variable over time is the concern.
Behavioral characteristics of positivefeedback loops and negative feedback loopsare explored in simple situations and thenexploited to analyse more complex multiple-loop structures. Positive loops are shownto be responsible for uncontrolled behaviorsuch as exponential growth, while negativeloops are shown to produce goal?seekingbehavior, though often with disturbingfluctuation* and cycles. S»shaped growthin several apparently different systems isshown to occur when a quantity is influencedfirst by a positive loop and then by a neg-ative loop; a shift from dominance by thepositive loop to dominance by the negativeloop produces behavior which looks initiallylike unrestrained exponential growth butthen becomes goalvmeeking. An importantprinciple will appear as different systemsare explored: systems with the same feed-back structure tend to behave the same overtime, in the sense that the shapes of theirgraphs over time are essentially the same.
The role of delays in goalvseekingsystems is explored intuitively. Studentswill become familiar with the principlethat delays can cause a variable to oscil-late around itmgoal.
The approach in this part of thecurriculum is essentially nonvquantitative,except that some quantitative graphing willbe done to provide the tools required tograph intuitively the behavior in causal-loop diagrams.
Students completing learning packageIII will have a foundation for the principlethat the bshaVior of a system is a censequence of its feedback structure. Theywill be ready to try to apply their under-standings to more complex situation*.
IV. Understanding Dynamic Problems'A number of recurring themes in real.,
world problem' are uncovered in this partof the sequence. Each theme is exploredin the oontext of several apparently dif-ferent dynamic problems, making use ofcausalvlocp models and the properties offeedback systems developed in Sections IIand III. The structure of feedback loopsresponsible f.; the thematic behavior is
296'
284 NECC 1980
exposed in each system, provi'ding a commonfocus for understanding the differentproblems sharing that theme.
One of the themes explored here issometimes referred to as the counterintu-itive behavior of complex systems=intentioned policies often tend not toproduce the behavior expected, occasion.ally producing results opposite to thoseintended. The phenomenon is tracedinitially to the distinction between openloop and closed-loop thinking. The formeroverlooks feedback in contrast to thelatter, which incorporates within theboundary of the system all the essentialfeedback influences. Counterintuitivebehavior appears again as the significanceof feedback structure is explored; feed.back systems tend to resist certain kindsof change, unless the actual structure ofthe system is affected. In feedbacksystems the short -term effects of apolicy may be different from, even opposite to, its long-term effects. The tend.:ency of complex systems to become dependentupon external controls is explored underthe theme of "shifting the burden to theintervener". The concept of tradeoffs isintroduced. In complex systems, policiesrarely improve all aspects of the systemat once; usually a policy improves someareas and is deleterious to others, remquiring policy-makers to teke explicitaccount of the tradeoffs.
The problems explored in this partof the sequence range from peer pressure,cramming for a test, and mowing lawns, tocriminal justice, pollution, urban growth.;stagnationi.decay, drug-related crime, andglobal population and food needs.
Learning package IV is the canineation of the essentially now.quantitative_part of this introductory system dynamicscurriculum. It shows the power of attempting to understand complex dynamic problemsby focusing on feedback loops, to illustratesome themes which recur so frequently incomplex systems that they may be called"principles", and to leave the studentswith a balanced view of the power and limitrations of their understandings at thispoint. as learning packages II, III andIV have progressed, the need for knowingthe relative strengths of feedback loopsin a system will have arisen at varioustimes, leading naturally (but not ;moos.eerily) to the next part of the sequencein which methods are developed for makingthe assumption. embodied in a causal-loop diagram unambiguous by quantifyingthem.
V. Introduction to SimulationComputer simulation in system dynamics
becomes necessary when the implications of
a structure of feedback loops are indoubt. Greater precision is required thana causal-:oop diagram can provide. SectionV develops the skills needed to translatesimple causal-loop models into quantitativemodels which a computer can trace throughtime, simulating the behavior of the actualsystem.
Two critical notions form the t 4481the concepts of a level (or stock) ada theconcept of a rate (or flow). Level var-iables are pictured as rectangles; therate adjusts the flow of something intothe level with which it is associated,just as a faucet adjusts the flow of waterinto a tub, changing the water level.
Sires hpu- Birth PopulationRate V Y
Figure 3. Causal-Loop Diagram as It Relatesto a Flow Diagram
This part of the Curriculum returnsto the work of learning packages II andIII, developing quantitative understandingof positive and negative feedback Loopsand the behavior of simple systems. Stu-dents will expend their skills in under-standing and interpreting graphs of var-iables over time. In addition, they willdevelop abilities to write general levelequations and the elementary rate equationsfor exponential growth and decay and sig-moid (S-shaped) growth. Exercises includefirst-hand calculatin7 and graphing themodel variables, then simulating the modelswith the aid of a computer. Familiaritywith computers and programming is not re-quired; the introduction to the simulationlanguage DYNAMO is self-contained. Theprogramming is presented as a means to anend: making assumptions sufficiently pre-cise and suitably coded that a computercan trace out their implications over time.Having hand-simulated the models first, thestudents will know what the computer isdoing.
At the close of learning package Vthe students are ready to understand morecomplex simulation models, they will havebegun to see the power of simulation, andthey will have solidified their understand-ings of the behavior of feedback loopscovered intuitively in learning package IIand III.
VI. Formulating and Analyzing SimulationModelsSeveral of the problems addressed in
learning package IV are reconsidered, gen-erally in greater detail. For each, in
2 Qt
turn,a quantitative model is developed inDYNAMO, and explorations of the system arecarried out by simulating different con-ditions in the 'model. The central focus,besides the significant problems them-selves, is the understanding of complexsystem behavior. Where in a given systemdoes intervention have the most effect?Why does the system behave as it does?What policies actually improve the be-havior of the system? Why does onepolicy have a desirable effect while otherswhich initially appear promising have littlehelpful effect or may even prove to beharmful?
Explorations of the behavior of asystem are carried out, and alternativepolicies investigated by changing,numer-ical relationships in the model, alter-ing, or adding equations. The computeris shown as an obedient servant tracingout the implications of a modeler'sassumptions over time. Each simulationmodel and each simulation run appearnot as ends in themselves, but as meansto understanding,the dynamics of a cer-tain problematic system. The goal isunderstanding, and feedback models helpus to understand certain kinds of problems.
Learning package VI completes thisintroduction to system dynamics. Stu-dents continuing through all six sectionswill have a new understanding of the causesof dynamic problems and the beginnings ofa set of tools for analysing and under-standing them. They will have learnedto look at problems holistically and tosearch for feedback loops responsiblefor the behavior of the system. Theywill understand the role of models inapproaching problems and what is re-quired to develop such models. The stu-dents will also,have an introduction tothe meaningful role computer models andsimulations can play in helping peopleto cope with the complex dynamic problemsthey face.
THE CURRENT CURRICULUM PROJECT
The project is scheduled for comple-tion in August 1980. During the spring of1980 it will be pilot tested in six highschools. Two teachers from each highschool will integrate the curriculum intotheir courses. These teachers are fromthe disciplines of Computer science,mathematics, chemistry, physics, English,history, environmental studies. And biology.
The targeted audience of secondarystudents need not be the only user, of thematerials. The materials that were createdfor fifth and sixth graders have been usedin high schools, universities, and evenin the MIT Sloan Fellows Executive Develop-
Computing Curricula 285
ment Program. The secondary level designshould also be appropriate for many under-graduate college classes, as well as forother possible learning settings. Thelearning packages have been planned tomeet the following criteria:1) Interdisciplinary and Unifying. Thematerials will provide examples from anumber of subject areas: physics, econ-omics, biology, ecology, sociology, anthro-pology, social studies, and literature.The generic similarity of structure andbehavior in these fields will be illus-trated.2) Relevant. The materials will deal withsignificant world and national problemsas well as problems encountered by studentsin their own lives.3) Hands-on Approach. The set of learningpackages will teach students the modelingprocess through a series of exercises.Initially, the students will 4evelop ontheir own a set of feedback diagrams de-rived from the verbal description of aproblem. Then they will go on to quantifythe elements of the problem through graph-ing, and finally write the equations re-presenting the structure of the model.The students will then hand simulate theirmodel and finally simulate the model withthe aid of a computer.4) Supplementing Rather Than Replacing
Existing Subjects. The curricularmaterials are designed to supplement ratherthan replace current courses. For example,the materials will be used to demonstrateintegration in a mathematics course, popu-lation growth in a biology course, factorsunderlying social problems in a socialscience course.5) Reiterative. Roth the problems selectedfor study and the skills taught will re-appear in several learning packages. Thestudents therefore will study the sameproblem areas with increasingly more detailsof problem elaboration and increasinglymore sophisticated methods as they proceedfrom causal-loop diagramming through graph-ing, to equation writing, hand-simulation,and then computer-simulation. Each skillthat is taught will be recalled for use inlater exercises that focus ostensibly onother skills development. Further, thestudent's sense of the similarity of under-lying dynamic structures will also be in-creased as problems from a variety of dis-ciplines are studied from a system's per-spective. The students will begin tounderstand how identifying the underlyingstructure of a problem gives one a strongsense of comprehending the problem.
The curriculum development group.hopes to do extensive field testing overthe next few years. Should anyone be interested in field testing the learning
25
286 NECC 1960
packages or wish more information on theprojec-, please contact either of theauthors.
minim1. Forrester, J.W. Industrial Dynamics.
Cambridge, MAt MIT Press, 1956.2. Odom, X.T. Energy, Power and Society.
New Yorks Wiley-Interscience, 1971.3. Roberts, N. "Teaching Dynamic Feedback
Systems Thinking* An Elementary View."Management Science, Vol. 24, April 19794pp. 836 -943.
2919
THE COMPUTER SOFTWARE TECHNICIAN MANI
AT PORTLAND COMMUNITY COLLEGE
1980
,David M. HataPortland Community College12000 S.W. 49th AvenuePortland, Oregon 97219
(503) 244-6111
ABSTRACT-This paper is a status report on a new voca-
tional program being offered at Portland CommunityCollege entitled "Computer Software Technician."The curriculum was developed jointly by PortlandCommohnity College and an advis'ry board composedof representatives from local electronics manu-facturing companies such as Tektronix, Intel, andElectra -Scientific Industries and is targeted atsoftware/hardware technician positions in theO.E.M. environment.
INTRODUCTIONPortland Community College is a comprehensive
community college serving all of Washington Countyand parts of Multnomah,Columbia, Clackamas, andYamhill counties, a region of some 679,000 people.Within the geographic boundary of the communitycollege district are located several large manu-facturers of electronic equipment -- Tektronix,Intel, E.S.I., and others.
Courses offered at Portland Community Collegeare organized and integrated into a broad varietyof programs ranging from associate degree, certif-icate career, and college transfer programs tospecial interest and enrichment courses, appren-ticeship training, and high school completion.The philosophy of Portland Community College isto offer learning opportunities to everyone re-gardless of prior educational experiences, aphilosophy which has earned the school the name.*the Educational Shopping Center.* It is in thisenvironment that the Computer Software TechnicianProgram is growing.
PROGRAM DEVELOPMENT
e need for computer softwaretechnicians began in late 1977. A survey con-ducted within various groups at Tektronix showedwidely varying skills among those polled. Forexample. the Information Display Group's skillset was much more computer oriented, greater than40 percent software, as opposed to less than 20percent for the Test and Measurement Group. Itwas also noted that groups with more softwareengineers had fewer technicians. It became
Computing Cunicula 287
evident that the software equivalent of the hard-ware technician did not exist.
With the increase in the number of micro-computer-based systems being manufactured, shiftsin the skill sets required in the manufacturingenvironment were projected to move trward increas-ing emphasis on software skills. Translated intofuture manpower requirements, an addition ofseveral hundred software professionals was pro-jected
.--
over the next five years, an impossibletask in the face of a critical shortage of tech-nical people. Using past experience in developingelectronic technician programs at the two-yearlevel, the two-year concept for training softwaretechnicians began to grow among managers.
Development of the program began in March1978 when Tektronix formally approached PortlandCommunity College with the idea of offering a two -year associate degree program to train computersoftware technicians. An ad hoc committee wasformed and instructors from the Electronic Engin-eering Technology and Data Processing Departmentswere assigned to course development tasks.
PROGRAM GOAL-----urrart programs in electronics and dataprocessing/computer science emphasize either hard-ware-related or software-related topics. The goalof the Conputir- Software Technician Program is totrain a technician who has the skills to write anddevelop microcomputer applications software underthe guidance of a software engineer and can bridgethe hardware/software gap by posses:Mg-skillsthat will enable him/her to verify correct systemoperation.
THE COMPUTER SOFTWARE CURRICULUMThe proposed Computer Software Technician
Program consists of 101 credit hours for the ass-ociate of applied science degree. The first yearconsists of perallm eardware and software coursesequences combinpr ith mathematics and communica-tions courses. C: wads develop basic problem-solving skills whi:m Wilding a foundation inbasic circuit theory and device operation. Thesoftware and hardware areas merge in the fourth
300
r-
288 NEcc
term as courses begin to integrate both softwareand hardware topics into a unified presentation.The program concludes with courses and on-the-joblearning experiences, to create the ability tofunction as part of a development team and learnthe art of project management.
TERM I
CST 2.211Lecture Lab. Hrs.2Software Programming I
$TN 125 Computer Oriented Mathematics I 5 5
WR 2.301 flatness Communications I 3 3
EL 6.117 Basic Electric Circuits 6 4 6
CST 2.221 Software Programming I 4 4 2
ITN 126 Computer Oriented Mathematics II 5 5
WR 2.302 Business Communications II 3 3
EL 6.127 Fundamentals of Semiconductors 6 6
TERM IIICST 2.231 Software Programing III 4 4 2
CS 233 Intro. to Numerical Computation. 4 4 2
WR 227 Technical Report Writing 3 3
EL 6.137 Digital Logic Fundamentals I 6 4 6
TERM IVEL 6.248 Introduction to Microprocessors 4 3 3
CST 2.241 tow Level Languages 4 4 2
SP 100 Basic Communication 3 3
EL 6.247 Digital Logic Fundamentals II 4 3 3
TERM VCST 2.123 Language Processors 4 4 2
CST 2.136 I/0 & Data Cbsmunication Prog. 4 4 2
PSY 1.546 Asychol &ammo 3 3
EL 6.257 Peripheral Circuits 4 3 3
TERM VI
CST 2.126 Project Management 4 4 2
CST 2.132 Operating Systems 4 4 2
CST 2.141 Field ProJect 6 1 18
EL 6.267 Advanced Micro Systems 4 3 3
IMPLEMENTATIONlaboratory Facilities:
Parallel efforts to upgrade existing electron-ics laboratory space used by the electronic engine-wing technology program and to acquire the spec-ialized equipment necessary to implement the Com-puter Software Technician Program have produced awell equipped set of laboratories to support bothprograms. Our initial goal to equip each labora-tory with industry-standard equipment has beenrealized. Specialized equipment for the ComputerSoftware Program purchased for laboratory use in-cludes logic analyzers, microcomputer treinerc, andsoftware development systems.
Digital System Laboratory:Dual-trace oscilloscopeDigital voltmeterFunction generatorTriple power supplyLogic analyzersMicrocomputer trainers
-so
Computational Systems Laboratory:Microcomputer systems capable of running
BASIC and PASCALMicrocomputer development systems
Cooperative Work Experience:A cooperative education program has been
established with local companies, and ten CSTstudents are currently participating in the pro-gram. Each cooperative education student is re-quired to carry twelve credit hours of academicwork while working twenty hours per week. Closesupervision is maintained between the cooperativeeducation coordinator and managers of the coopera-tive education students.
The response has been favorable among parti-cipating managers; student response has beenpositive although many have expressed the feelingthat the academic work load combined with workassignments has been demanding.
Faculty Upgrading:Program development has reinforced the need
for a teaching faculty having a skill mdx thatspans the hardware/software boundary. To date,faculty upgrading has been accomplished by tech-nical seminars, independent study, and interactionwith industry. Keeping pace with an ever changingtechnology is a continual struggle, and incorpor-ating new teaching ideas into courses remainsbefore our faculty.
Course Content Guides:Course content guides are being reviewed
for content and updated. These course contentguides are available upon request by writing tothe author.
IMPACT ON COMPUTER EDUCATION-nate processing and computer science pro
grams to date have been very business applica-tions oriented. Curricula are heavil), weightedtoward programing languages and applicationsthat do not fit the treditional product develop-ment cycle.
The Computer Software Program was a totallynew program rather than a modification of anexisting program. The electronic engineeringtechnology am data processing Programs wereleft intact since they are currently serving awell-defined industry need. in this manner, acurriculum was developed to meet a future needfor technicians with both software and hardwareskills.
With the first gradmaiing class in June1980,'software'technicians capable of fittinginto systems development and manufacturing groupswill be entering the Job market. It has beenevident from the ftedbeck obtained through mana-gers of our cooperative education students thatthe skill mix will be right for entry-level
posngititheseons in software groups currently employ -
programs of this Ups are opening up new
Computing Curricula 289
areas in which persons interested in computers,computer programing, and electronics can enter:Programs of this type have not been available inthe past. To industry, these technicians willprovide the analogue of the hardware-orientedhardware technician that has been missing.
CONCLUSIONThe need for computer software technician
programs will grow, and the Computer SoftwareTechnician Program at Portland Community Collegecan serve as a model for programs at othercolleges. It is the result of close industry -college interaction and is currently being eval-uated through a cooperative work program withlocal electronics companies.
302
290 NECC 1980
A COMPUTER SCIENCE MAJOR INA SMALL LIBERAL ARTS COLLEGE
Joerg MayerDepartment of Mathematical Sciences
Lebanon Valley CollegeAnnville, PA 17003
(717) C67-10SS
This presentation examines the diffi-culties, both practical and philosophical,which stand in the way of offering a com-puter science degree at a small four-yearliberal arts college. Also I will presentthe compromise which was struck at ourcollege.
Until recently, computer science as anundergraduate degree program was almostentirely restricted to the larger uni-versities. Such programs were expensive,both in terms of teaching personnel andequipment, and required an environmentwhich included an engineering school.These conditions are easing somewhat, andit is becoming increasingly difficult fora small college to resist the incentivesto enter the cornuter science market.Computer hardware is now within the meansof most institutions, and for those withinadequate funding there are ties into acommercia2 or educational time-sharingnetwork. Also with declining overallenrollments on the horizon and an at beststable interest in most traditional sub-jects, the continuing shortage of college-trained computer personnel seems to offerthe chance to shore up at least some ofthe drain in the student population byoffering a major in computer science.
Encouraging though the emerging con-ditions may be the arguments againstadding computer science to the program ofa liberal arts college are rather unset-tling.
Despite the efforts of many, thereremains a sizeable gap between the defend-ers of liberal education as the attempt tocapture the essence of the complete man,and the proponents of an introduction toscience and technology as the only meansof preparing our youth for the technolog-ical future. The present mixture of majorprograms in most liberal arts institutionsreflects more an uneasy truce than theresult of mutual understanding. The
separation between the two cultures showsno signs of dissolving, at least not ineducation.
Computer science as an undergraduatemajor brings this dichotomy in educationinto sharp focus, perhaps because thisdiscipline is predominantly machine andprocess oriented. The aim of computerscience is not to seek truth or to under-stand man and nature, and in that senseit is anti-humanistic. Computer scien-tists are as seen by the humanist,essentially non-reflective, meaning thatthey do not concern themselves with theethical and moral implications of theirwork. (Weizenbaum's book on that issueis not widely known; and in any case,it does not seem to be very popular amonghis colleagues.) Therefore, computerscience is considered by many to be anti-thetical to the basic philosophy ofliberal education. *One only has to ob-serve the nearest computer freak to findproof for the dehumanizing, indeed deper-sonalizing influence of that machine.*Computer science, as a relatively youngdiscipline--exuberant, irreverent, and insome ways irresponsible and antagonisticto deeply held cultural values- -does noteasily fit into the age-worn quilt ofliberal education.
Even if the misgivings were somehow tobe overcome, however, the practical prob-lems of offering computer science in asmall liberal arts college are most for-bidding.
To begin with financing the neededtechnical expertise is beyond the capa-bilities of most such institution.. Itwould be unwise, both politically andfinancially, for the college to hire abeginning Ph.D. in computer science for$25,000 when the average salary of theassistant professore in that college is$14,000 and that of full professors is$22,000. Another manpower problem arises
from the ACM standards set in Curriculum'78. There it is recommended that *ap-proximately six full-time equivalentfaculty members are necessary. ...* Witha student :faculty ratio of 14:1 thistranslates into roughly 80 computer sci-ence majors, a number not likely to bereached in an institution of 1500 students,which is the typical enrollment of a smallfour-year college.
Another difficulty lies in the courseswhich should be offered. According toCurriculum '78, a minimal program in com-puter science consists of twelvecourses in computer science and sevencourses in mathematics, making a total of57 credit hours in the major. Rare is \the liberal arts institution that wouldallow such a concentration of a student'stime in his or her major.
Finally, there is the need for hard-ware and software. Again, Curriculum '78established certain benchmarks. "It isessential that appropriate laboratoryfacilities be made available that arecomparable to those in the physical . . .
disciplines . . .. The initial budgetarysupport . . . may be substantial.* Alsomentioned are extensive software systems.The whole package, even if conservativelyinterpreted, presents a financial burdenthat reaches well into six-digit numbers.
Realistically, then, it is impossiblefor all but a handful of the best endowedsmall liberal arts colleges to find thetotal budgetary support for the start ofa computer science major which approachesthe standards set by Curriculum '78. Soa student who wishes to major in computerscience must limit his choice of a col-lege to those with an enrollment of atleast 5000. It means that the smallercollege cannot enter the market of com-puter science education. Finally, itmeans that employers and graduate schoolsof computer science cannot reach into thatrich pool of above-average, motivated,and broadly educated graduates.
At Lebanon Valley College we wore drawninto the computer science field more byaccident than by design. In the earlyseventies when it became clear that themajor in actuarial science was veryattractive to prospective Students andcould be so designed as not to violatethe precepts of liberal education tooseverely, the Department of Mathematicsdecided to expand its offerings into theapplicable areas. The department grewsteadily, so that in 1978 there were 70majors in a school whose total enrollmentwas 950. In 1975, enough computer sciencecourses hadbeen added to offer an inform-al,concentretiOn in computer science. The
Computing Curricula 291
technical courses included programming(advanced), computer architecture andassembly language, data structures, andan independent study. Of the 36 mathe-matics graduates since 1977, six tookcomputer-oriented jobs, seven becamesystem analyst/programmers, and fourentered highly respected computer sciencegraduate departments.
Encouraged by these results, andstrongly persuaded by the AdmissionsOffice, a major in computer science wasadded to the programs of the Departmentof Mathematical Sciences. Its structureis as follows. All students in the De-partment, regardless of their major, takethe same core curriculum which consistsOf&
1Analysis 13 hoursFoundations of Mathematics 3 hoursDifferential Equations 3 hoursIntroduction to Computer Science 3 hoursLinear Algebra 3 hours
During the difficult sophomore year,the students assess both their capabili-ties and their strongest interest.Accordingly, some will drop out of com-puter science and others will shift intoit. For that major the remaining re-quirements are:
The internship must be taken in somecommercial or industrial computer opera-tion, usually during the summer. Thelanguages taught during the last threeyears are BASIC-PLUS, FORTRAN, andAssembler (PDP11/40). We strongly en-courage the mastery of one additionallanguage, such as RPG or COBOL. .
There are other requirements. To givethe students some bacuground in the basiccomponents of computer hardware, theymust take a year of physics and work in aBeall computer science laboratory. Alsorequired are six hours in psychology, andthree hours in an ethics course designedto deal with the ethical issues inherentin modern technology, and computers inparticular. Finally, having observed theoften abominable technocratese of hand-books and manuals, we require that thecomputer science major take a three hourwriting workshop.
Thus, the major consists of 31 hoursin mathematics, including the Curriculum
3u4
292 NECC 1980
'78 courses MA 1, MA 2, MA 5, MA 3, CS 17,CS 18, with MA 6 strongly encouraged;-twelve hours in computer science, includ-ing CS 2, CS 3, parts of CS 4, also CS 7,and CS 81 18 hours in what we consider re-lated topics; and three hours internship.The totals are 46 hours in technicalcourses and 18 hours in supportive disci-plines.
These requirements are in contrast toCurriculum '78, which sets a minimum tech-nical requirement of 51 hours, of which 30are in computer science.
The disadvantages of our approach areat least the following. It makes us Un-comfortable to be in such disagreementwith Curriculum '78, and we would not beprepared to adjust if the suggested pro-gram of Curriculum '78 were to developfrom recommendations to quasi-accredita-tion requirements. Secondly, the rathernarrow scope of the computer sciencecourses must result in a limited appreci-ation of the broad sweep of that field.And finally, the substantial as difficultrequirements in mathematics will mostlikely lead to a higher dropout rate thanwe are accustomed to because many studentsinterested in computer science are mathe-matically not very gifted.
The advantages are that we will be ableto increase the enrollment in our depart-ment and that all the students in thedepartment are being exposed to the newdiscipline. For them, the main and over-riding advantage is the flexibility whichthey enjoy during the first three years.To be able to switch, without loss ofcredit and required information, betweenmathematics, actuarial science, computerscience (and probably operations researchin the near future) is valuable to themathematically gifted, who did not intheir high sohoo1.4ears have any indica-tion of the many job and graduate schoolopportunities for which they may be pre-destined. For me personally, the greatestadvantage of our approach is that it showsour students, and not just in theory, thegreat breadth of mathematics and the mutualdependence and influence of the variousdisciplines we incorporate in our totalprogram.
Two conclusions appear to be clear.First, a small liberal arts college canoffer a computer science major only if itdoes not adhere too closely to the recom-mendation of Curriculum '78. And second,such a Major has any chance of successonly if it is incorporated in the oper-ations of an already existing department.Its44Seeultyembers must be willing towork hard mastering most of the coursesin computer science while they are still
teaching the normal twelve hour load.The Computer Center personnel must besympathetic to the reality that the newmajor will tie up the system more oftenthan they are accustomed to. And theadministration must be willing to acceptthe fact that one cannot start a tech-nology-based major without improving theexisting hardware and hiring at least onespecialist. It is well to remember thatmore often than not, the necessary sup-port will come at least two years afterit is needed.For those who may be contemplating
implementing n variation of our a2proach,let me close on a bright note. Withintilt first semester of the new major, fourstudents transferred into it from otherdepartments in the college, and two morestudents transferred into it from otherschools because they wanted both computerscience and a small liberal arts insti-tution.
3 06
Abernethy, ,Kenneth 179Allard, Me 230Alpert, Elizabeth 96Amenn, G.W. 266Anger, Frank D. 171, 249Aronson, Michael 1
Baird, W.B. 266-/"4"4".Baldwin, R. Scott 46Bass, George H. Jr. 38Beck, James D. 245-it,Seidler, John A. 143Bishop, Judy H. 147Bork, Alfred 258Bowman, William R. 16Boyle, Thomas 214Brown, Bobby 199Brown, Guy Larry 256.
Brown, Thomas 237Burrows, Robert L. 220Bush, Steve 19
Caldwell, Robert H. 31-A4-6'1-Collins, Ronald W. 74Corbet, Antoinette Tracy 54Czejdo, Bogdan 112
Davis, A. Douglas 86Deal, Ralph H. 281DeKock, Arlan 138Dempsey, Richard F. 152Dershem, Herbert L. 65-Aw444.Dorn, William 75
Effarah, Jamil E. 25Ellison, Robert J. 68
Franklin, Stephen 258Friedman, Frank L. 103Fdhs, F. Paul 205
Garraway, Hugh 200Carson, James W. 42Oats, A. John, Jr. 76Gordon, Sheldon P. 169Oruener, William 198Groaa- Thornton, Joan C. 54
Hagee, Michael W. 90Haiduk, H. P. 222Hannay, David O. 119Hata, David H. 287Hausmann, Kevin 2, 170Hazen, Paul 73Hopper, Judith A. 62Hunter, Beverly 168Hybl, Albert 276 "A.e..4
Author Index
293
Johnson, Dale M. 46Jordan, Eleanor U. 7, 260.44"'
Kneifel, David 37Kurtz, Barry 258
LaFrance, Jacques E. 261Little, Joyce Currie 220Lim, Tian S. 90Lovis, Prank 75
McVay, Peter O. 27101.46.0e40,
Hagnant, Peter 214Manor, Walter 221Marshall, Sr. Patricia 238Mayer, Joerg 290Heinke, John G. 143Mebane, Donna Davis 197Mebane, Rodney 197Hocciola, Michael 236Morgan, Catherine E. 168Hoshe1/4 J.H.Hoursund, David 125
Oliver, Lawrence 137, 169Olivo, Richard F. 81 - 442749,01w.404
Piegari, O. 179Poirot, James L. 130Pollak, Richard A. 90Powell, James D. 1So
Rahmlow, Harold 1
Reggie, James A. 276--.44"4Roberts, Nancy 281Rodriguez, Rita Virginia 171
Schimming, Bruce B. 58Shelly, Gary 3
Sobol, Thomas 155/14.0.
Stokes, Gary 4Stoutemyer, David R. 194 - 'e'~4'
Sustik, Joan 199
Taylor, Robert P. 130,155Taylor, Timothy 223Thorkildsen, R'n 230Thorsen, A.L. 179Tinker, Robert F. 250Turner, A.J. 6
Varanasi, Murali R. 5
Ward, Darrell L. 19Wehrle, Howard F., III 12Wet lore, David E. 139 A4'4-'4V,Whittle, John T. 65/..r...et..
