CSC 8570 -- USI

Class Meeting 4September 14, 2010

Topics of the Day

• From last time• Research project issues– Bibliography– IRB form– Theory and models

• Good and bad interfaces• Lee & Zhai. Extracting design principles• Cataloging design principles• GOMS review

Where is Gregor?

One thousandth of a minute= 6 x one hundredth of a second

Diameter of earth = 7926 miles, approximatelyClaim: Assuming the earth is a perfect sphere,

the arc of a circle with central angle of 1/100 second is 1.0+ feet long. (latitude)

Claim: At 40 degrees north, the arc of longitudinal circle with central angle of 1/100 second is 0.77+ feet long.

Hand In

• Everyone: GOMS model of table creation• Each research team:– Hypotheses, independent and dependent

variables– First draft of IRB form– Printout from EndNote of research bibliography

Research Project

• Hypotheses (tonight)• Independent variables (tonight)• Dependent variables (tonight)• Bibliography– First entries (tonight)– Continuous process

• Background • Formal setting (let’s talk)

Research Project (2)

• Short presentation– One spokesperson– Five minutes– Hypothesis, variables, initial experimental design– PowerPoint

Research Project (3)

• Experimental design– Between subjects– Within subjects

• Sources of bias• Sources of subjects• Logic of conclusions

Lessons from Lee & Zhai

• Subjects: 13 total – 9 men, 4 women• Overall design: practice, test,

questionnaire/interview• Experiment structure:– Block– Trial

Experimental Design

• Within subjects– vs. Across subjects

• Randomized order of conditions– Generating random numbers

• Balanced Latin square design– Think Sudoku

Experiment 1

• Conditions (10): button type + feedback type• Entry set (5): 1450 X 9276 = ; 8327 – 7231 = ;

etc.• Order of entry set randomized• Block (3): Instances of ordered entry set– How many entry set arrangements are possible?

• One subject generates 3 x 5 x 10 sets of data– Time, # of corrections

Hard ButtonsBlock 1 - A Block 1 - N Block 2 Block 3

Subj 1 5 trials 5 trials

Subj 2

Subj 3

Subj 4

Subj 5

Subj 6

Subj 7

Subj 8

Subj 9

Subj 10

Subj 11

Subj 12

Subj 13

Inferring Design Principles

• Augment soft buttons with synthetic feedback• Make soft buttons at least 10 mm wide• Provide automatic error correction for text

entry using small buttons

GOMS

• Goal• Operators: atomic actions available in the

system• Methods: each method is a sequence of

operators• Selection Rule:

GOMS Examples

• Deleting a message from a mail system folder• Setting a clock or watch– one hour ahead– one hour back– to a particular time

• Creating a KWIC index of a list of research paper titles

Clock Setting

• Set the clock one hour earlier, the common action at the end of daylight savings time.

• GOMS model– Goal: stated above– Operations:– Methods:– Selection Rule:

How Was Your Week?

• Write down two examples of system interfaces that were very helpful as you tried to complete a task.

• Write down two examples of system interfaces whose design made it difficult for you to complete a task.

Good Interfaces

• Who says?• On what basis?• Why were your choices (of helpful interfaces)

good?• What were the design principles that

promoted “goodness”?

Good Interfaces (2)• iPhone tells when apps to be updated and can be updated all at once:

decreases work on part of user, simplicity, allow same operation to be applied to multiple targets in one go (select all)

• toolbar shortcuts on a browser (IE): simplifies task, decreases keystrokes, allow accessing “favorites” with simple actions, provide the capability to create shortcuts

• purchasing groceries at self-check asks whether cash back w/o drilling down: time-saving (keystroke saving) choices, based on smart choices, combines tasks

• using EndNote to add and format references: simplifies (understanding of) task, • Firebug allows manipulating and debugging web page programming• Minipen gives easy translation• Embed a map in a text message• Redbox for renting a DVD• Autocompletion on Garmin GPS

Bad Interfaces

• Same questions as for “good”– Who says?– On what basis?

• What design principles were violated?

Bad Interfaces (2)

• System did not follow standards • ABG antivirus—getting free version (web interface)• Importing songs to iTunes: inconsistent behavior,

are there hidden states?, remove all hidden states• Windows XP gives “random” errors• MS messenger• Makefile in C didn’t run – excessive white space: • Moving app from computer to iPod• Selective startup for Microsoft Windows

Good and Bad (2)

• What do users want?– Easy to learn– Consistent with what users know already– Quick response time unless perceived as complex– Minimum level of effort to interact– Good information about use– Powerful systems– Accomplishes the task at hand– Compatibility, consistency across systems and platforms

• How do designers know?• Do designs match mental models?

– What if they don’t?

Good and Bad (3)

• Generative User Engineering Principles (GUEPs)

• Cognitive Dimensions (CDs)– of notation– of representation in an interface– of understanding

NB. In this context a “dimension” is a factor (distinct from others factors)

Next Time• Read Edge and Blackwell, Cognitive Dimensions Tradeoffs in

Tangible User Interface Design– You find this paper by going to

www.cl.cam.ac.uk/~afb21/CognitiveDimensions– Then choosing the link for the 10th Anniversary Workshop on CDs

research• Create a concept map relating the ideas of the paper to one

another– Use the concept map software– Make sure you include at least 12 concepts and maybe many more– Consult the papers by Thomas Green for more information

• Research project presentation

Research Team Meetings

Modeling Actions

• Task analysis: temporal (time) issues• GOMS analysis: method (how to) issues• Keystroke level analysis: operator (widget)

issues

Theory Collection

Visual processing• Three-stage visual system• Preattentive processing theory (Triesman &

Gormican)• Structured object perception theory

(Biederman)• Scientific color theory (CIE)

Theory Collection (2)

Motor Behavior Models• Hick-Hyman Law: choice time• Keystroke-Level Model: error-free task

completion time• Three-state model of graphical input (Buxton)• Fitts’ Law: human movement• Guiard’s Model of Bimanual Skill

Theory Collection (3)

Models of Computation• Finite state machine• Grammar in Backus-Naur form (BNF)

Theory Collection (3)

Task analysis• Describes the process the user chooses to

reach a goal in a specific domain• Hierarchical model using ConcurTaskTreesDesign principles• GUEPs• Cognitive Dimensions• Mental models