ERSHOVi

Stanford University Lihrarii»sDept. of Spec n

Coll Title

BoxFol d Fol. titleFol. title

/>

80 ORIXF_MFqXIONI

A History of ComputingIn the USSR

Andrei P. Ershov

Though second generation machines still perform the bulk of the workload,a nationwide computer utility is planned for the 1980s.There are three specific fundamentalfactors which have greatly affected thedevelopment of data processing in theU.S.S.R.:

" The Soviet Union has been forcedto developall aspects of the com-puter business relying exclusivelyon its own intellectual and tech-nologicalresources.

" Computer development in theU.S.S.R. exists outside the scopeof multinational computer com-panics.

" The U.S.S.R.'s social system di-rectly and decisively influencesthe development and applicationof computers.

The first factor is not now as abso-lute as it was, say, 10 to 15 years ago.The normalization of international re-lations and growth of exchange intrade, science, and technology, led toa levelling of the scientific and tech-nologicalbase through a flow of tech-nical information. The extension andreinforcement of this base have beengreatlyaffected by processes of integra-tion in the national economies of So-cialist countries and, in particular,have been embodied in the develop-ment of es evm (the Unified Range ofElectronic Computing Machines)known better under the acronym"Ryad." It wasn't always that way.

The history and prospects of com-puter developmentin the U.S.S.R. canbe divided into four 10-year periods.The '50s werecharacterizedby the firstgeneration computers, their applica-tions to the most urgent scientific andtechnological problems, and pioneer-ing works in computer design, softwaredevelopment, and basic applicationsconcepts.

In the '60s, the second generationcomputers appeared, a foundation forthe regular production of computersand theircomponentswas built, and allbranches of the computer businesswere identified and formed.

In the '70s, a broad expansion ofthird-generationcomputers into all sec-tors of the national economy is takingplace. In all major organizations, com-puter applicationshave been developedto the point of regular service, anddata bases are being computerized.

In the '80s, a global network of

computers, terminals, and switching The main component of this boomcenters will be established. This net- is the production of es evm (orwork will allow an integration of data "Ryad") computers. These machinesbases and their transmission at elec- are produced through a joint venturetronic speeds. All business units of the of the seven socialist countries actingnational economy, without exception, as a kind of large-scale multinationalwill be involved in computer applica- company. Twenty thousand peopletions through a network of multi-ac- participated in es evm design and pro-cess computer utilities. Public com- duction preparation; 70,000 are in-puter applications (schools, medical volvedin production,service, public libraries, home termi- The second program is for com-nals) will start to grow on a large puterized automation of manufactur-scale. ing and management. Systems for

Computer developmentis subject to these applications are called Auto-a number of long-range national pro- mated Control Systems in Russian ter-grams with well-establishedgoals. minology. Industry is the main field for

The first program is for computer introducing them, and they are accord-production. Since 1971, computer ingly classified as:production has been considered a sepa- " automated process control sys-rate industry sector and separate statis- terns (apes)tics have been kept for it. The total sale " plant automatedmanagement sys-price of computers manufactured in terns (pams)1970 was 710 million rubles. The pro- " branch automated managementgram for production in the 1971-1975 systems (bams)period assumed a growth rate of 2.6 " regional automated managementtimes the 1970 level; the actual rate' is systems (rams)greater than this. (See Fig. 1.) The last three correspond to the En-

The

MINSK-32,

one of the U.S.S.R.'s first transistorizedcomputers has approximatelythe same computing power as the GE 400. novosti

from sovfoto

s 1

11lWe found the only thing mini aboutthe3000 was its price. J*fWhen we asked the EDP center manager

of another major manufacturing company aboutthe 3000, that was what he had to say. He alsohad this to say:

"Our computer needs include bothscientific and commercial applications. We werephasingoutourteleprocessingterminalandourEnvironmental MonitoringDivision's computer.So we started looking. We spent several monthsstudying computer systems, and rated them onspeed, versatility and ease of operation. Theresult of our study showed that the HP 3000provided these requirements and had the bestcost/performance ratio. We didn'tfully realizethe potential of the 3000 until we started pro-gramming it. We have experienced a significantcost savings in the seven months we've hadthe 3000 and we expect a greater savingsin the months ahead. We really like theinteractive CRTfor programming anddata input. Being a multi-programmingsystem we can have many users on at thesame time. The power andspeed of the 3000is equal to a largemachine. It'snomini. Callingit the Mini Data-Center is moreaccurate. I'd definitelyrecommend the 3000to other potential usersIn fact, we already have.We feel they wouldbe money ahead. \A WkmmKHH^kSeptember, 1975

We're gladtheseand otherusers of theHP3OOOCXset us straight. We called it a mini-computer because its state-of-the-art technologylets us sell it for a minicomputer price. Fromnow on we'll call it a Mini DataCenter.

We want youto getthe wholestory- Write us foryour copy of our HP 3000CX Mini DataCenterbooklet. We know you'll find it interesting,informative, and maybe a bit surprising.

HP Mini DataCenters.They workfor a living.

HEWLETT MPACKARD

Sales and service

from

172

offices

in 65 countries.1501 Page Mill

Road,

Palo

Alto,

California 94304

79

,i'

CIRCLE

58

ON

READER

CARD

>

82 DHTfIMBTION

U.S.S.R

glish term "management informationsystems." pam operates within the lim-its of a single enterprise;

bam,

within abranch headquarters; and ram coverssome territorial region, like a city.

The program of automated controlsystems started during the eighth five-year plan, 1966 to 1970. Over 400systems in 30 sectors of the industrywere put into operation at a total costof over 1 billion rubles, plus indirectexpenditures for basic research. Inspite of the experimental nature ofmany of these systems, and their shortduration of operation, the net returnfor that period was 680 million rubles—mainly realized as a consequence ofraising labor productivity.

This program has been given realthrust now. The Directives of the 24thCongress of the U.S.S.R. CommunistParty prescribed the implementation,between 1971 and 1975, of about2,000 automated control systems—inparticular, 65 statewide and 1 50 re-public-wide

bams,

over 1,000 pams,and 670 process control systems. In1974, 614 new systems were put into

operation.Another long-term program is the

construction of the Statewide Auto-mated System of data collection andprocessing for accounting, planning,and management of the national econ-omy, the Statewide Automated Man-agement System (sams).

Sams

will beimplemented in two stages. The

first,

during 1971-1980, will be the globaldesign and the definition of standardsand procedures. Stage I involves thecreation of a Unified Automated Com-munication Network (uacn). Onebasic component of the system is theStatewide Data Communication Sys-tem to which any computer or ap-propriate terminal could be switched.The character of this work does notrequire much comment except that thegigantic size of the Soviet Union andits longitudinal geographic orientationmake the problem extremely difficult.

During sams second stage, a greatnumber of multi-access computer utili-ties will be constructed. According tosomeestimates made at the earlystagesof preliminary planning, some 10,000multi-access computer utilities are re-quired.

THE HARDWAREThis list of the main computer

models currently in active use in theU.S.S.R. also shows their Westerncounterparts. The majority of Russianmachines are still second generationones, including the besm-6, a reliableand powerful machine whose value isstill rising due to the growing pool ofsoftware available for it.

The minsk-32 is a very popular in- culations and analytic manipulations,expensive computer with reasonably it can also be used as an intelligentgood software; it is the primary ma- terminal,chine for management information The nairi-3 is very popular as asystems. small, compact configuration for engi-

M-220 and besm-4 are workhorses neeringand design departments. It hasof the '60s for engineering and small a large read-only memorywhere a sub-scientific applications. routine library and algol compilerare

The mir-2 has no direct counterpart stored. Recently it has been given anin Western computers; it combines a emulator for the minsk-22, the pre-microprogrammed byte-oriented mini- decessor of the minsk-32.computer with an alphanumeric disc s evm (Ryad) models do not re-play and a light pen. Besides being quire any specific comments. They useused extensively for engineering cal- integrated circuits and have perfor-

1960 12 3 4 1965 6 7 8 9 1970 12 3'Fig. 1. Since 1971, computer production has been tracked as a separate industryin the Soviet Union. The government program for production in the 1971-1975 periodassumed a growth factor of 2.6 times the 1970 level; the actual rate is greater thanthis, largely due to the ES EVM or "Ryad" Series. The figure for 1974 surpassed1,850million rubles targeted for 1975..

APPLICATION DEVELOPMENTIN THE U.S.S.R.1951 Scientific application(MESM)1953 Math tables (BESM)1954 First public computingservice (STRELA)1 956 Language translation experiments( BESM )1957 Economic calculation (M-2)1958 Medical application (STRELA)1961 Process control (DNEPR)1964 Bank office (URAL)1966 Chess program(M-220)1967 Plant management information system (MINSK-22)

The first scientific computations were done during the checkout of the mesmcomputer. The economic calculations were also a self-serving step, from thehardware's viewpoint, as they were done as part of a study to introduce thecomputer into economics. Also of interest is the chess program; it beat anAmerican program running on a 7090 in a four-game match.

«

New from Datamedia.Conversational. The Elite .520Aspeaks your application language,and is ideal for interactive programming, data entry, information retrievaland simply anyplace the need for data communication is important.

Expandable. The Elite 1520Aoffers the standard 64-character ASCIIset displaying 1920 alphanumeric characters in a 24-line/80-characterformat, and is optionally expandable to a 128-character set, including upperand lower case. And now, APL/ASCII switch-selectable capability.

Separable. This newest video terminal from Datamedia is a stand-aloneunit, with the display and keyboard separated for greater applicationsflexibility.

Compatible. It's plug-compatible with Teletypewriter* Models 33and 35, either via a standard RS232C or an optional 20/60mA current loopinterface. Data rates from 50 to 9,600 bps may be accommodated.

Economical. No matter which performance yardstick or competitiveunit you stack it up against, the Elite 1520A's $1555 price ($1655with upperand lower case option) in quantities of 1 to 9 (20% lower in quantitiesof 100) is tough to beat.

NEW! Reliable. Our practical design approach is the answer, and our list ousers worldwide of other proven Elite video terminals can tell you whateliability is all about in a CRT. Or we can tell you.APL/ASCII is not one of those announced-today,

someday products. It's here. It's availablehe Elite 1520A

vered-.mised-tomorro

Phenomenal

Los Angeles,

Call

213/397/3556

UNITED

KINGDOM

Mellor DataWoodgates RoEast Bergholt,

WEST

GERMANY

Data Systems

GmbH

ederried 3

CIRCLE 41

ON

READER

CARD

letypewriter is a registered trade name

of

Teletype Corporation,

Skokie,

Illinois

r

September, 1975

84

mance characteristics approaching the quent maintenance. Recently the man-ibm 360 models 30, 40, 50, and what ufacturing ministries have established awould be models60 and 70. statewide networkof field engineer and

The es 1010,

electronika,

and M- software teams which, on a contract6000 are minis. basis, take full responsibility for the

In the early period of mass produc- installation and maintenance of com-tion, computers were sold, installed, puters. In the

U.S.S.R.,

computers areand put into operation by a manufac- not rented, but service bureaus areturer without any obligation for subse- quite widespread.

Developed in 1966, the second generation BESM-6 is still the most powerful Sovietcomputer. novosti

from sovfotoNOVOSTI from SOVFOTO

SOFTWARE DEVELOPMENTIN THE U.S.S.R.1953 Subroutine library (STRELA, BESM)1955 Algorithmic language compiler (PP-2, PP-BESM)1 957 Assembler (PAPA, SSP )1962 Algol 60 compiler (TA-1)1962 Batch processing monitor( AUTOOPERATOR)1 966 Delivery of first operatingsystems (MINSK-22, BESM-6)1967 Systems programming languages (EPSILON, ALMO)1971 Operatingsystem with batch and multi-access (AIST-O)1972 Delivery of time-sharingoperating systems (OS IPM, DNEPR-2)

The first Russian computers contained read-only memory with built-in li-braries of math and conversion functions. Compilers were the first bigadvance, providing Assignment,

Conditional,

For statements, subscriptedvariables,and, strangely enough, appearing before assemblers.

HARDWARE DEVELOPMENTIN THE U.S.S.R.1951 ENIAC type computer1952 First logical structure computer (BESM )1953 First serialproduction line computer (STRELA)1960 Transistorizedcomputers (RAZDAN, DNEPR, MINSK-2)1 966 Most powerful computer (BESM-6 )1969 Conversational computer with display(MIR-2)1970 Integrated circuit computer ( NAIRI-3, M-3000 )1 971 Time-sharing systems (AIST-O, SIRENA, BESM-6)1971 Minicomputers (ELECTRONIKA, M-6000)1972 ESEVM orRyad computers (1010, 1020, 1030, 1040, 1050)

Though built in 1966, the BESM-6 still remains the most powerful computerbuilt in the U.S.S.R. Its rate is 1 million operations per second. Third-generation computers did not appear until 1970, when the small (and verypopular) NAIRI-3 and the medium-scale 360-like M3OOO were introduced.The SIRENA system listed as "time-sharing" is for airline reservations.

Second Generation MachinesSoviet Model U.S. CountU.S. CounterpartBESM-6 CDC 3600MINSK-32 GE4OOM-220 ] ,„,. ,-.„BESM-4 ) IBM 7040MIR-2

Third Generation MachinesSoviet Model U.S. CounU.S. CounterpartNAIRI-3

IBM 360ES EVM1020 301030 401040 501050 "60"1060 "70"

ES 1010 MITRAISELECTRONIKA DEC PDP-8M 6000 HP-21168

THE SOFTWAREFor every Russian computer, there

is both "official" and "unofficial" soft-ware. Official software is deliveredby amanufactureras an integral part of theproduct. This software is not separate-ly priced and its cost influences theprice of a computer quite indirectly.Unofficial software usually appears asa result of some user activity and isaccumulated at a particular installationeither through exchange agreements orthrough user association efforts.

Official

software,

in turn, consists ofbasic and secondary software. Basicsoftware is part of the design specifica-tions prescribed by a general con-tractor.For national economy applica-tions, such a contractor is the StateCommittee for Science and Technol-ogy. The basic software is usually acompromise between users* insatiabledemands and manufacturers' availableresources. The manufacturer gets bigpieces of the basic software from sub-contractors—some academic or educa-tional institutes.

The secondary software has usuallybeen successfully developed by oneuser for his own purposes, and later, asa result of formal recommendation bysome authoritative committee, is ac-cepted and ratified by the manufac-turer*

In the established Russian terminol-ogy, a well developed software systemconsists of three parts: the operatingsystem, the programming system (in-cluding compilers, debugging means,and other man-machine interfaceaids), and application packages. Database management is not yet technicallystable and falls into operating or pro-gramming systems depending on thedesigner's philosophy.

The minsk-32 has a sound exampleof a well integrated software system ofreasonably high quality. Most usersadopt the official software withoutmodification. The size of the system isover 225K words, available in 175modules. A high integrity also charac-

%

Soonerorlateryou'll switchto

key-to-disc.It's in

thecards.■M:.-

0m- nv^^

0*"

-■

MM g .

■"~liS£m> ■■■■■""-■

'"»>mm*v:;m

m

86 DnTHMRTiaN

U.S.S.R. Programming systems France, had chances to run his pro-It should be noted that the Soviet gram without changing the basic al-

Union has become the home for gorithm through several different com-terizes the software systemfor es com- ALGOL . Algoland alpha, its extension puter

S

and compilers. The results ofputers. Its first release, delivered in into complex and vector-matrix arith- the tests are shown in pig 21972, consists of

■

600K instructions metic, are the primary languages for

'

A few words about' the fates 0fana is close to dos/ 36U.

scientific,

engineering, and educational Fortran and COBOL are appropriate.Dnpratinp cVctpmc

purposes. Its popularity is depicted in a

FO

rtran appeared in the U.S.S.R. as an« " . systems breakdown by language type of all the direct result of scientific exchanee butOfficial operating system versions

■

. .v, „

t

J ~ . faireci resuu or scieniinc excnange, oui

provide support for batch multioro J°b& mn at the ComPutat,on Center of not until the second half of the '60s! Itsprovide support tor oaten mult pro- th_

siberian Division of the Academy first imDiementations were for thegramming on sing c processor, large .c " / v- v v -. nnn . impic_n_ni_.uun_ weie iui uic

and medium scab computers. Thefe f Sc"

C

% Si MINSK

"

22 ***

BESM "

6: they ""*operating systems are of reasonably during Apnl 1974: in response t0 the demands of physi-high quality with respect to effective alphaperformance. The overhead time for assemblerthe dispack operating system for algol

besm-6 is no more than

9%;

the idle Fortrantime is no more than 8% and the epsilonmultiprogramming factor is over 1 .6. Others

cists who had visited

Cern, Switzerland,

and were eager to use programs devel-oped there.

35%20%13%

The sale of western computers to theU.S.S.R. also resulted in the gradualgrowth of the use of Fortran. I amnot sure we are experiencing a

12%9%

11%Some actively used unofficial sys- not sure we f re experiencing a

terns have much broader capabilities. one of the reasons for

ALGOL

'

S

and

f°RTran

boom, but Fortran as wellThe operating system of the Institute ALPHA'S great popularity is the fact that as algol and cobol are 'by law ob-of Applied Mathematics for besm-6 they nave lent themselves to the con- hgatory parts of every basic softwareprovides a combination of multipro- struction of so many efficient Russian set for generalpurpose computers,grammed backgroundbatch processing compilers. Although I have no elab- The need for developing a commonwith a multi-access foreground and orate statistics to prove that efficiency, business-oriented language became ap-adopts a multiprocessing architecture. a case study of one program for parent in the U.S.S.R. in the beginningA universal time-sharing operating svs- meteorological calculations supports of the '60s. The first responses to thetern for the minsk-32 has been devel- my c laUTI - Dr - G- R- Kontarev from demand were algol adaptations toopedbut not yet ratified tne Novosibirsk Computation

Center,

data structure description known un-Operating systems for small cpus durinS his visits t0 the U -S- and der the names of ALGEC and ALGEM-

and minis provide a combination of ■»—— — — „_... „....■F . , . ... PERFORMANCE

PROGRAM

RUNthe main program calculations with language computer thous. op/sec time mm.operator dialogue. The official os forthe dnepr-2 also has some multi-accessfeatures. There may be only severaldozen time-sharing systems in opera-

OCTAL

M-220 40 45

ALPHA

BESM-6

1000 2

3000 6

FORTRAN CDC

6600

tion, but their number is growing very optimized... „,

_i

7 " i

FORTRAN CDC

6600 3000 4rapidly. The most usable terminals arestandard

rft

teleprinters from

ddr,

1000 12

ALGOL Oil

terminalsbased on a cssr electric type- p ig 2 . The relative efficiencies of various compilers was tested by running a meteoro-writer

"Consul,"

and Hungarian al- logical program through several machines without changing its algorithm. The ALGOL-phanumeric displays called "Video- based ALPHA compiler on the second-generation BESM-6 stacks up nicely, eventons." against optimized FORTRAN on the CDC 6600.

Some implementations of these lan-guages are still in use.

Meanwhile, patient work on COBOLand its adaptationto the Russian lan-guage led to a "minimal" cobol forbroadly used implementations on thednepr-2 and minsk-32. Now a draftof Russian standard cobol has beenprepared which corresponds to theansi version.

Pl/ 1, algol 68, and

Simula

67 lan-guages are not yet in use, but aboutready. And almost all time-sharing op-eratingsystems have Russian as well asEnglish versionsof joss and basic.

Applications packagesThe first applications packages were

in the lines of standard subroutine li-braries. Later, large application pro-grams were created in the "black box"style with a monolithic structure whichmade adaptationand modification dif-ficult. Recently a new concept hasbeen accepted, that of a structured andopen applications package taking theform of a dedicated operating system

The ES EVM or Ryad series of third generation computers is reportedly comparableto the IBM 360 series. This model 1050 has a main storage capacity of 128-1024KB;the larger 1060 has twice this amount.

*Foryears,keypunchand cards havebeen a perfectlyfine combination for

dataentry. Butnowadays,throughputis thename ofthe game. Andif you're stillplaying withcards, you're not holding a winninghand.

Enter Inforex.Inforex is theworld's leading manufacturerofkey-to-disc dataentry systems

Sofar we've installed over 3000systems throughout the world. Andalmost all of themwereconversions from keypunch.

Thereasons forour success are simple.Our "nofrills" designapproach andourenormousamount of experience

allow us to make theconversion fromkeypunchtokey-to-disc at thelowestcost ofanymanufactureranywhere. And doit with a minimum of troubleto theuser.

Once the system is installed, we back it up with an extensiveand highlyeffective fieldservice organization. So if youneedhelp, we're nofurther away thanyourphone.

And thanks to theremarkable efficiency of the Inforexsystems, yourcom-puter canreceive purer data, errorscan be virtually eliminated,and throughputcanbe increased to a level thatkeypunchers can only dreamabout.

There's a whole series of Inforexkey-to-disc systems to choose from:Systems 1301, 1302,and 1303.One ofthemis justright foryourpresentneeds. Andasyougrow,your Inforexsystem will growwith you. Because we have aline ofcompatiblesystems thatwill take youas far as you want to go.

For more information,give us acall ormail us thecoupon today. Because ifthere's ever been atimefor switchingfromkeypunch tokey-to-disc, this is it.

Like we said, it's tathecards. r*l INFORiIi^C21NorthAvenue,Burlington, Massachusetts01803.

-.7_SW_R. , l';::

";

——□ Iwould likemore informationonyourkey-to- j*disc dataentry systems, Series 1300.

._.—

->.». ■

_. :,■.-_.. ,.',;...„.,„,,„„,„,,„ ,1 D Please havean Inforexsalesrepresentative.:«_■- _;..-.-„ contactme. UmMnnira mi

i______ Name' "'" ■'' '" "-—-——»——.

■■■—

m

mmmawmammrim.*.,,.

, .1

j TitleOrganization

Addressi! City. State Zip.i

TelephoneDD4 l

Officesinmajor cities throughout theUnited States,Canada,andEurope.Distributors worldwide.

CIRCLE

31

ON

READER

CARD

tU

U.S.S.R.

either occupying all computer re-sources or acting as a subsystem of ageneral-purpose operating system.

With our current knowledge, suchan applications design is obvious toeverybody, but it poses a problemwhich hasn't been solved in the

The NAIRI-3, developedby Armenian sci-entists, is used primarily to solve engi-neering and design problems. This thirdgeneration integrated circuit machinehas 4K words of memory.

NOVOSTI from SOVFOTO

U.S.S.R. either. We can establishmodules of a quite different nature insuch a system:

" data access (search,

fetch,

store,sort)

functions (problem-oriented)

" communication with a user (in-put data translation, text editing,output display)

" control (calls, storage mapping,parameterprocessing)

" communication with a designer(modification, adaptation, exten-sions, statistics)

This variety requires different special-ists in the design team, and differentimplementation languages. If we cansometimes find the right language tosolve the second half of the problem,we often cannot find the right mix ofskills to solve the first half.

Research and developmentSpace limitations prevent a detailed

exposition of the state of research, buta few interesting facts are easy to pre-sent. Since an analysis of publicationsin the field reflects reasonably well thetrends in r&d, I have looked through800 research and technical papers ap-pearing in publications in Russia andthe U.S. datamation was one of theU.S. publications. The percentages ofpapers on various topics were closelyrelated between the countries. The big-gest differences were nearly traditional—the Soviet papers had more on pro-gramming systems, math methods, andmath theory of computations; U.S. pa-pers had more on computer equipmentand operating systems. On the otherhand, these traditional differences are

88

diminishing (and the comparison isgetting tougher, now that datamationis being widely circulated within theSoviet Union).

Also, in the Soviet Union there is notradition of holding annual gatheringsof thousandsof professionalssimilar tothe U.S. National Computer Confer-ence. Again this might be explainedbythe size of the territory. However,smaller thematic symposia and confer-ences are held regularly. And Soviet dpexperts travel widely. (For example,the full staff of the Novosibirsk Com-putation Center numbers 630; they ac-counted for 600 business trips and 40visits abroad in a singleyear.)

Education and trainingHigher education, at least from the

organizationalpoint of view, effective-ly responded to the appearanceof thenew branch of knowledge called com-puter science. Nevertheless, it tookabout 10 years to establish the neces-sary basis (computers and professors)in educational institutions for provid-ing mass education in computerscience.

At first I should make a few generalremarks. In the

U.S.S.R.,

education isprovided through universities and "in-stitutes." Universities roughly corre-spond to schools or faculties of sciencein European and American univer-sities. Institutes are divided into poly-technical institutes (which correspondto schools of engineering or technicaluniversities), economic institutes, andbranch institutes which prepare spe-cialists for specific industries or othersectors of the national economy. Eachstudent graduating from a universityor institute gets a kind of degree of acertified specialist.

In the early '60s, the major educa-tional events were the opening of adegree in economic cybernetics, inuniversities and in engineering math inpolytechnical institutes. But the keyissue in computer related educationtook place in 1969 when a formal pro-gram in applied mathematics (for uni-versities and polytechnical institutes)and in automatedcontrol systems wasestablished. The applied math is ouranalogy to computer science and in-formatics in Western schools. The spe-cialty in automated control systemsprepares specialists for designing andmaintainingsuch systems.

In 1972 a formal course in appliedmathematics was opened at 40 univer-sities and institutes, and one in auto-mated control systems at 50 institutes.Besides this regular education, two andthree year courses were opened in 23universities and institutes for getting asecond diplomainthese subjects.

Basic courses in computer scienceinclude programming and languages,computer architecture, information

processing methods, numerical meth-ods, optimization, and operations re-search.

Students practice on computersthree and a half years, typically usingan installationprovidingbatch process-ing with algol and Fortran com-pilers, and, to a lesser extent, assem-blers.

In addition to the formal schools,there is a "scientific and technologicalschool" in every center of an adminis-trative region in the country. Theseschools charge up to 40 rubles forshort, intensive courses or seminars ona variety of subjects, including com-puter subjects. Their average enroll-ment is between 100 and 200.

ConclusionI apologize for a fragmentary pre-

sentation, but I deliberatelypreferredto present a collection of more or lessconcrete information rather than gen-eral statements.

On the whole, Soviet computerscientists and engineers are faced witha great deal of work to respond effec-tivelyto those challenging problems inthis sector of science and technology.Nevertheless, premises to solutions tothe problems are being established, thegoals have been

formulated,

and theprogram of action determined. Wehope that this program will be a con-siderable contribution to world tech-nical and social progress. We are surethat this program of computer de-velopmentcan be a broad platform forintensive international cooperation inscience, technology, and trade. $

A leading

Soviet

authority on com-puters, Dr. Ershov is a professor

of

Computer Science at NovosibirskUniv. His professional associationsinclude membership in

ACM,U.S.S.R.

representative on theIFIP Technical Committee on Pro-gramming, member

of

the IFIPWorking Group on Algol, and edi-torial board member

of

several

Soviet

computer journals. He hascontributed to

many

projects in-cluding the ALPHA programmingsystem

for

the M-20 computer,and the BETA universal program-ming processor.

The information here was

first

presented at the XXXI Diebold

Conference

in Rome in 1974.

DfITRMRTION

.%

September, 1975 87

CIRCLE

65

ON

READER CARD

Expert adviceisn't expert unless you'vechecked out Basic/Four.

Next time someone who's not an expert in thedata processing field asks for yourendorsementof a small business computer, you're not anexpert unless you're thoroughly familiar withthe Basic/Four® system.

After all, how good is your endorsementif the pioneer and innovator in thefield hasn'tbeen checked out?

Sofar, over 1,500 businessmen are usingBasic/Four mini-based business computers. AndBasic/Four business computers are not onlydoing their accounting and bookkeepingjobs, but generating information they neverdreamed possible. Like sales and managementreports, instantaneous credit checks, billing, andinventory control.

And doing it all at far less cost than theaccounting machine or service bureau way.

Indicative of Basic/Four's leadership,consider the fact that the industry giant justentered the small business computer market.Imitation is the sincerest form of flattery.

Next time someone asks you for yourendorsement, be ready. Check out the Basic/Four small business computers. For a completebrochure, write Basic/Four Corporation, 18552MacArthur Boulevard, Irvine, California 92707.

/Four corporation joA Subsidiary of Management Assistance Inc. (MAI)

basic

SEATTLE■

SAN

FRANCISCO■ LOSANGELES

"

ORANGECO.■ SANDIEGO■ PHOENIX■ MINN./ST

PAUL "

ST

LOUIS "

CHICAGO

■

INDIANAPOLIS■ CLEVELAND

■ NEWORLEANS

■

BOSTON

"HARTFORD■

N.Y

■

NJ.

■

BALTIMORE

" WASHINGTON

■ PHILA

■

ATLANTA

■

MIAMI

■

VANCOUVER

■

TORONTO

"

MONTREAL

"

OTTAWA

"

MEXICO CITY

"

PARIS

"

LONDON

"

BRUSSELS

"

FRANKFURT

"

HAMBURG

■

STUTTGART

■ DUSSELDORF■

MUNICH

■

ZURICH

"

AMSTERDAM

"

VIENNA

"

MADRID

■

COSTA RICA

"

SAN JUAN

■

CARACAS

■

PANAMA

■

NICARAGUA

"

SANTIAGO

"

MANILA