DOCUMENT RESUME

ED 052 637 EN 009 118

AUTHOR Walker, Gerald S.; Gardner, Edward M.TITLE Application of Computers in Educational and Training

Systems: A Survey of Computer-Assisted InstructionalCenters.

INSTITUTION Air Force Human Resources Lab., Lowrey Air Fo:ceBase, Colo.

REPORT NO APHRL-TR-70-24PUB DATE Dec 70NOTE 32p.

EDRS PRICE EDRS Price MF-$0.65 HC-$3.29DESCRIPTORS *Computer Assisted Instruction, C Omputer Graphics,

*Demonstration Centers, Display Systems, *InputOutput Devices, *Programing Languages, Simulation,*Surveys, Training Laboratories

ABSTRACTIn an effort to gain an overviev, of the application

of computer technology in education and training, to identify areasrequiring additional research, and to identify/potential applicationareas for use in technical training, field viz its were made torepresentative military and civilian institutions which are using the

literature and computer equipment manufacturers, technical data wascomputer in instructional programs. A revi/.0 of the recent research

also conducted. The primary focus of the/study was on the terminaldevices available for use in a training/system and on the softwarerequired for effective use of a computer in a training environment.Several c(mputer terminals used in computer assisted instruction(CAI) are discussed and computer-controlled audio presentation isdescribed. Programing languages which have been used in CAI arecategorized into four groups: general-purpose problem languages,extended problem languages, interactive terminal problem languages,and author-student problem-oriented languages. A typical languagefrom each category is examined. On the basis of their survey, theauthors make some suggestions regarding terminal devices, programinglanguages, and areas for further research. (JY)

AIR FORCE

N-re\.0(NJ111

Lul

"0

AFPS A

iH

MAN

R

AFHRL-TR-70-24

APPLICATION OF COMPUTERS IN EDUCATIONALAND TRAINING SYSTEMS: A SURVEY OF

COMPUTER-ASSISTED INSTRUCTIONAL CENTERS

By

Gerald S. WalkerEdward M. Gardner, 2nd Lt., USAF

TECHNICAL TRAINING DIVISIONLowry Air Force Base, Colorado

10

December 1970

ORATORY

Approved for public Hease; distribution unlimited.

C

S LAB

AIR FORCE SYSTEMS COMMAND

BROOKS AIR FORCE BASE, TEXAS

NOTICE

When US Government drawings, specifications, or other data are usedfor any purpose other than a definitely related Governmentprocurement operation, the Government thereby incurs noresponsibility nor any obligation whatsoever, and the fact that theGovernment may have formulated, furnished, or in any way suppliedthe said drawings, specifications, or other data is not to be regarded byimplication or otherwise, as in any manner licensing the holder or anyother person or corporation, or conveying any rights or permission tomanufacture, use, or sell any patented invention that may in any waybe related thereto.

ir)fJ

.NN1 AFHRL-TR-70-24

N^

w

U.S. DEPARTMENT OF HEALTH,EDUCATION & WELFAREOFFICE OF EDUCATION

THIS DOCUMENT HAS BEEN REPRO-DUCED EXACTLY AS RECEIVED FROMTHE PERSON OR ORGANIZATION ORIG-INATING IT. POINTS OF VIEW OR OPIN-IONS STATED DO NOT NECESSARILYREPRESENT OFFICIAL OFFICE OF EDU-CATION POSITION OR POLICY.

APPLICATION OF COMPUTERS IN EDUCATIONALAND TRAINING SYSTEMS: A SURVEY OF

COMPUTER-ASSISTED INSTRUCTIONAL CENTERS

By

Gerald S. WalkerEdward M. Gardner, 2nd Lt., USAF

Approved for public release; distribution unlimited.

TECHNICAL TRAINING DIVISIONAIR FORCE HUMAN RESOURCES LABORATORY

AIR FORCE SYSTEMS COMMANDLowry Air Force Base, Colorado

December 1970

FOREWORD

This report documents the initial efforts of the Technical Training Division of theAir Force Human Resources Laboratory to assess the state of the art in the application ofcomp..ter technology in the educational and training process.

The authors wish to thank the personnel at the educational research centers visitedfor their cooperation and assistance in supplying information for this report. Thanksshould also go to the computer equipment manufacturers who provided the technicaldata used.

Every effort has been made to provide accurate information in the areas reviewed;however, in a field as complex and dynamic as the use of computers in instruction, theinformation reported remains current for only a very short period. Any references to aspecific computer system or component for illustrative purposes should not be construedto infer a recommendation of that item.

This research was accomplished under Project 1121, Technical TrainingDevelopment; Task 112102, Application of Computers in Air Force Educational andTraining Systems. Dr. M'irty R. Rockway was the project and task scientist.

This report has been reviewed and is approved.

George K. Patterson, Colonel, USAFCommander

4

ABSTRACT

This paper presents an overview of representative centers active in the research,development, and application of the computer in education and training. Field visits weremade to military and civilian agencies in order to obtain a cross-section of various typesof applications and research projects. The information presented is based on the fieldvisits and the most recent reports available on the activities of the various centers.Computer terminals used in computer-assisted instruction are discussed, andcomputer-controlled audio presentation is described. Programming languages which havebeen used in computer-assisted instruction are categorized, and a typical language fromeach category is examined.

lll

5

SUMMARY

Walker, G.S. & Gardner, E.M. Application of computers in educational and training systems: A survey ofcomputer-assisted instructional centers. AFHRL-TR-70-24, Lowry AFB, Colo.: Technical TrainingDivision, Air Force Human Resources Laboratory, December 1970.

Objectives

The objectives of this study were (a) to provide an overview of the application of computertechnology in the educational and training process, (b) to identify areas requiring additional research, and(c) to identify potential application areas for use in technical training.

Approach

Field visits were made to representative military r.nd civilian institutions using the computer in theinstructional process. A review of research literature and computer equipment manufacturers' technicaldata was also conducted.

Conclusions

In addition to the overview provided by the field visits, the primary focus of the study was on theterminal devices available for use in a training system and the software required for effective use of acomputer in a training environment. The following conclusions were reached in these areas:

1. A large number of new lower cost terminal devices are currently available and appear to havepotential for use in a technical training system.

2. Hardware interface for analog-digital applications can be simplified through use of small low-costprocessors of the mini-computer type.

3. Use of mini-computer interface expands the range of mainframe processors suitable for use intraining applications and provides greater flexibility for interfacing of nonstandard input/output devices.

4. Software for use in instructional systems should be designed to provide maximum flexibility inuse of all types of input /output devices.

5. Course writing languages must provide author support for on-line data entry.

6. Course authors should have minimal constraint in developing, testing, and implementing a widevariety of instructional strategies.

This summary was prepared by G. S. Walker and E. M. Gardner, Technical Training Division, AirForce Human Resources Laboratory.

v

I. Introduction

TABLE OF CONTENTS

Page1

Purpose 1

Background 1

Method 1

II. Discussion 2

Field Visits 2

Florida State University Computer-Assisted Instruction Center 2Human Resources Research Organization 3The University of Illinois Computer-Based Education Research Laboratory 3University of California at Santa Barbara Computer Center 4Naval Personnel and Training Research Laboratory, San Diego, California 5Other Navy Programs 5Stanford University Program in Computer-Assisted Instruction 6The University of Pittsburgh Learning Research and Development Center 7

Computer Terminal and Visual Display Devices 8

Teletype ASR-33 Terminals 8Dot Matrix CRT Displays 9IBM 2250 Vector Display Scope 9Digitized Television Display (CRT Terminal) 10Digitized Television Driver 10Direct View Storage Tubes (DVST) 10Computer-Controlled Optical Projectors 11Addressable Projection Display 11Owens-Corning Photosensitive Glass Terminal 12rortable Printing Terminals 12Portable CRT Terminals 12Color CRT Displays 12

Computer-Controlled Speech 13

Computer-Controlled Sequential Tape 13Computer-Controlled Random Access Audio Device 13Digitized Computer-Composed Speech i4Digital/Mechanical Time-Compressed Speech 14Touch-Tone Telephone Entry 14

CAI Languages 14

III. Conclusions and Recommendations 20

Terminal Devices 20Languages 20Suggestions for Research 20

References 21

vii

7

EXPLANATION OF TERMS

ADC - Analog to Digital Converter. Converts analog voltages to some binary number of correspondingmagnitude.

Alphameric terminal. A device which can display letters and numbers, but no graphic or photographicinformation.

ASCII - American Standard Code of Information Interchange. A type of binary encoding used to representcharacters sent over telephone or telegraph lines.

Audio response unit. A computer peripheral which stores words or phrases of human speech on a drum ordisk device which allows the words to be retained and played, possibly to several peoplesimultaneously, under computer control.

BCD - Binary Coded Decimal. An alternative 6-bit code for encoding characters. Allows 64 characters to berepresented. Commonly used on second generation computers.

CAT - Computer-Administered Testing. The presentation and collection of answers to questions presentedon a computer-controlled terminal device.

Computation. Performance of mathematical processes using the computer as a calculator.

Computer-managed instruction. Use of the computer to aid the instructor in guiding the student throughthe instructional process and in scheduling his academic activity.

CRT Cathrode Ray Tube Display Unit. A device whose display is on a phosphor coated evacuated tube ofthe type used in television receivers.

Cursor. A movable point of light giving visual indication of a point which is being referenced on a screen.

1)AC - Digital to Analog Converter. Converts binary numbers to some analog voltage of correspondingmagnitude.

Data channel. A device which reads from or writes data into core memory of a computer. Typicallyconnects peripherals to mainframe of computer.

Dataphone. A Bell System device which allows direct connection of teletype compatible equipment to atelephone line without an acoustic coupler.

Delta modulation. The process of representing the magnitude of an analog signal by periodic statement ofits relative change in value since the last sample rather than by periodic statement of its absolutemagnitude.

Drill and practice. Use of the computer to guide, control, and monitor by repetition of a specific task or setof tasks.

DVST - Direct View Storage Tube. A phosphor screen storage device which stores its display as an internalstatic charge in the tube. Contents stored are viewed directly as points of light on screen.

EBCDIC - Extended &nary Coded Decimal Interchange Code. An alternative 8-bit code for encodingcharacters. Allows 256 characters to be represented. Commonly used on third generation computers.

Emulation. The process of duplicating the input-ouput response pattern of a hardware device by means ofother hardware which interprets some and duplicates other parts of tlk' emulated system.

Graphic tablet. A two-dimensioned surface and a light pen type device which inputs the location of the penon the surface to the computer.

Graphic terminal. A terminal which can display not only letters and numbers, but also lines and figureswhich can be composed of illuminated lines and points.

Joy-stick. A lever controlled input device which is used to position a cursor on a graphic display screen.

Light pen. A probe device which is touched to a screen in order to select a point on the screen. The selectedpoint is conveyed to the computer to which the probe is attached.

ix

8

Logical problem solving. Use of the computer for performing a series of computations necessary to arrive ata solution to a problem.

Mainframe. The processor, channels, core meraory, and control unit of the computer. Does not includeperipherals.

Mini-computer. A class of small primarily arithmetic processors addressing no more than 32,000 16-bitwords and accessing a limited number of larger peripherals. Can usually be connected to largerprocessors and often interface between large systems and various slower or non-standard devices.Priced normally less than $15,000, as low as $4,000.

MODEM - Modulator-demodulator. A device to encode binary computer information into audioinformation to be transmitted over telephone lines. Usually a higher-speed more complex device thanan acoustic c aupler.

Mouse. A graphic input device in which the device is rolled around on a plane surface to position a pointeron cursor on a graphic display screen.

Multiplexer. An electronic device used to pack data from several slow devices into a single high-speedtelephone 1.ine. Reduces line cost when several devices are to be communicating with the samecomputer since one high speed line is cheaper than all of the individual lines which would have to beused if no multiplexer were available.

Peripherals. Input and output devices attached to the mainframe for information transfer and storage.Typically, tape, bulk-storage, card reader, line printer, communications handler.

Simulation. The process of duplicating the input-output response patterns of a hardware device by means ofa software program designed to interpret input to the simulated system in a manner analogous to thehardware being simulated.

Simulation and gaming. A model of a real life situation represented by a given set of circumstances andparameters stored in the computer.

Telecommunications (teleprocessing,!. Connection of peripherals to a remotely located mainframe overtelephone lines of some type. A MODEM is normally required for connection.

Time-sharing. A simulated environment in which each of several users is allowed access to a single computerand is generally unable to distinguish any interactions of the system with other uses.

Tutorial dialog. A technique based upon interaction between the student and the computer in which theresponse of the computer varies to fit the individual characteristics displayed by the student.

Video (photogruphic) terminal. A terminal which can display information with varying gray tones such asphotographs, graphic, and alphameric information.

x

9

APPLICAI ION OF COMPUTERS IN EDUCATIONAL AND TRAINING SYSTEMS:A SURVEY OF COMPUTER-ASSISTED INSTRUCTIONAL CENTERS

I. INTRODUCTION

Purpose

The purpose of this report is to document theinitial efforts of the Technical Training Division toexplore the application of computers as an aid inthe instructional process. The use of the computerin an instructional system sears to have long-range potential when select rely applied to AirForce technical training.

This report is part of the preliminary researchrequired to develop the selection criteria for amulti-media computer-based instructional system.The system should be capable of providing aresearch tool for exploration of the variousinstructional strategies available in a computer-based system controlling, or used in conjunctionwith, other instructional media. Development ofthis system is a part of the Air Force advanceddevelopment program "Innovation in Training andEducation." Unit B of the program calls for thedevelopment of a functional prototype of anadvanced instructional system. The system will bedesigned to integrate computer-assisted instruc-tion, automated training, and audio-visual displayswith information handling techniques all con-trolled by a computer.

Most applications of the computer as a tool toaid in the instructional process can be categorizedas Computer-As'sisted Instruction (CAI) orComputer-Managed Instruction (CMI). CAI gen-erally involves the computer as an active elementin the learning process, thereby providing sometype of man-machine interaction. CMI is an aid tothe instructor in guiding the student through theinstructional process. The computer monitors thestudent's progress through his materials but doesnot interact directly with the student. Thesegeneral categories of applications, combined withan appropriate audio-visual presentation systemand management information system, will form apart of the prototype advanced instructionalsystem.

Background

The Air Force used computers in training forthe Semi-Automatic Ground Environment System(SAGE) in 1955. The system provided automatedon-the-job training via simulation. An updated

1

version of the training is currently in use with theBack-Up Interceptor Control System (BUIC). TheAir Force has also sponsored CM research on theSocrates system developed in 1963 by Bolt,Beranek and Newman. The system used a DigitalEquipment Corporation PDP-1 computer with on-line typewriters and auxiliary drum storage. Appli-cations were made to electronic troubleshooting,medical diagnosis, and accounting, (Feurzeig,1964).

Since 1967, the Air Force has also jointlysponsored research by the System DevelopmentCorporation on development of a natural languagecapability for CAI.

Currently, the Air Force is field testing acomputer-directed training subsystem developedby the System Development Corporation for useon base-leve Burroughs B-3500 computers. Thissystem is designed to interact with the user bymeans of a remote keyboard/printer device whichsends or receives one line of data at a time. Thesystem is not designed to control any other on-linedisplay devices at the present time. Two courseshave been developed for use on the system, aB-3500 Computer Operator CAI course and -acourse on Operation of the Base-Level MilitaryPersonnel System. The CAI system is programmedin an adaptation of System Development Corpora-tion's PLANIT CAI language. The system has beendesigned to meet a variety of training needs, butthe current course material is primarily for on-the-job training. Another on-the-job training appli-cation of computers was developed in 1966 on alimited basis for Intelligence Training on use of theformatted file system for intelligence data re-trieval. The project developed and tested thecourse at the Intelligence Data Handling SystemDirectorate, Headquarters, Continent'al Air De-fense, Colorado Springs, Colorado. The system,although not currently in use, did tiutonstratethat CAI could be effectively used for on.the-jobtraining in an operational environment.

Method

Field visits were made to representative educa-tional and training research centers using thecomputer as a tool ia aiding the instructionalprocess. These centers were selected to get as,broad a cross section as possible of the hardwire,software, and instructional applications. The visits

1 0

provided current information on the variousprojects and also established communications forfuture exchange of information.

A review of the most current literature on theuse of computers in training was conducted toprovide a background on the system requirementsfor educational use. Computer manufacturers'brochures and equipment specifications were alsoused to determine the status of computer tech-nology relative to possible use in a computer-basedtraining system.

In addition to describing the activities of theresearch centers visited, this report examines twoof the most critical factors in the development of acomputer-based training system: (a) computerterminal devices and (b) computer languages usedin instructional systems.

II. DISCUSSION

Field Visits

The information presented in this section isbased on observations and documentationsobtained during field visits conducted fromNovember 1969 through May 1970. The visitsprovided an up-to-date assessment of the field andan in-depth understanding of the problems associ-ated with the varied applications and systemsobserved. A summary of the activities of each ofthe research centers visited is presented in theorder that the visits were made.

Florida State University Computer-Assisted Instruction Center

Since 1964, Florida State University (FSU) hashad an active program for research, development,and operation of computer-assisted instruction.Starting with one terminal in 1965, the center nowhas an IBM 1500 CAI Instructional System. Thissystem operates 32 cathode ray tube (CRT)terminals which have both light pen and keyboardresponse devices. The system also has a tele-communications capability supported by a DEC680 switching system controlled by a PDP-8computer. This system supports eight teletypeterminals located in the Wakulla County schools,approximately 20 miles from FSU. These termi-nals are used by elementary and junior high schoolstudents for reading, mathematics, and oral lan-guage CAI materials. The FSU center hasdeveloped its own Data Analysis and ManagementSystem for the 1500 CAI system. Lack of a suit-

able data reduction and analysis capability for useby faculty and students doing educational researchmade it necessary for them to provide their ownsoftware capability. Efforts are also being made tointegrate the CAI software with the data manage-ment system to provide an operating system forCAI research. Work has also been done to extendthe research capabilities of various CAI languagessuch as COURSEWRITER II, BASIC, APL, andFOCAL.

The hardware system approach at FSU has beentoward small dedicated CAI systems such as theDigital Equipment Corporation PDP-8 TSS and theIBM 1500 CAI system. This approach may proveto be the only practical cost-effective way CAI canbe introduced in the near future on a broad scale.The introduction of so-called mini-computers withlow-cost time-sharing capabilities and CAI soft-ware may be the beginning of the technologicalbreakthrough required to broaden the implementa-tion of CAI.

In addition to computer systems development,FSU is working on the following CAI researchtopics .(Hansen, O'Neil, Brown, Rivers, & King,1970):

(1) Identification of those instructional processesand learning behaviors which we most efficientlyoptimize by computer support, (2) investigation ofappropriate instructional strategies that implementoptimal learning, (3) investigation of learner stra-tegies as they are revealed within CAI trainingsystems, and (4) investigation of the reliability,effectiveness, and validity of each new computer-based instructional activity (e.g., computer-managed instruction).

The FSU center also assists in CAI curriculadevelopment for use in university courses. Thecenter also provides instructional services to othercolleges and public schools. An undergraduate CAIcourse in physics which had been offered in thepast for credit provided excellent results with test-review exercises. It is planned to continue to usethe test-review exercises and to expand the prob-lem solving units as a service function.

A partial list of projects illustrates the varietyand scope of the research program conducted bythe FSU CAI center (Hansen et al., 1970):

1. Social Science Information Retrieval.2. A Rural County Computer Related Instruc-

tional Technology Project.

3. The Mediated Transfer of Words.

4. Preventative Maintenance Systems.

1

5. Learner Control in Automated Instruction.

6. Effects of Memory Support on Anxiety andPerformance in Computer-Assisted Learning.

7. Development of a Dynamic Decision Modelfor CAI Utilizing Regression Analysis Tech-niques.

8. An Investigation into the Effects of Re-tention on Differential Feedback DelayIntervals.

9. Effect of Sex and Stress on State Anxietyand Pe for mance in Computer- AssistedLearning.

This brief summary of activities of the CAIcenter at FSU demonstrates the potential for boththe research and the service function of a relativelysmall yet effectively managed computer system.

Human Resources ResearchOrganization

Project IMPACT is an advanced developmentproject sponsored by the United States Army andconducted by the Human Resources ResearchOrganization (HumRRO). IMPACT is anacronym for Instructional Model PrototypeAttainable in Computer Training. The projectis designed to

. . . . provide the Army an effective, efficient,and economical computer-administered instructionsystem. The objective is to (a) develop two genera-tions of a prototype CAI system with (b) accom-panying prototype multipath (branching), individ-ualized programs of instruction.

The system of instruction is to be capable ofadapting to the capabilities and characteristics ofeach individual trainee. This adaptiveness will bebased both on the "entry characteristics" of thetrainee and on his long term and immediateresponse patterns within the course, so that eachstep in the instruction will be fitted directly to hisneeds at that point in the instructional process.The instruction will also be made directly relevantto his specific job requirements (Seidel, 1969).

One of the most important aspects of theproject is the design and development of theInstructional Decision Model (IDM). The model isto be a set of rules for "matching presentation ofspecific content (selecting and sequencing) withtrainee capabilitites (student characteristics andresponses to earlier material)." The developmentconcept of IMPACT requires the testing of a seriesof IDM iterations to develop the most effectivemodel for a given training situation. The conceptof the instructional model in the IMPACT projectis that of "an open information-exchange system,

3

in which there is continual interaction between thestudent and his environment, and in which theeffective environment for the student changes withcontinued experiences (Seidel, 1969)."

The project's current computer system is anIBM 360 Model 40 with 256 bytes of core storage,an IBM 2314 disk drive, four 2401-5 tape drives, a1403 NI printer, a 2540 card reader/punch, a2501 reader, and three 1050 typewriters (oneremotely located). Of special interest are the 12student terminals which contain a CRT displaywith light pen, keyboard, visual display (rearprojection) capable of presenting material in stillor motion picture form, and an electronic(Sylvania) tablet that permits a somewhat stylizedhand-written input to the computer.

The software efforts in this project have beendirected toward the expansion of the IBMCOURSEWRITER III CAI language. A list proc-essing capability has been added and anothermodification allows the use cf CRT terminals overtelephone lines at any distance from the centralprocessing unit.

Efforts are also being made to establish criteriafor design of a CAI language that would be largelyindependent of the computer system on which it isused.

A CAI course in COBOL Computer Program-ming is being developed as the initial instructionalmaterial to be implemented. The course content isbased on a field survey and task analysis of the jobfunctions of COBOL programmers in the Army.

Project IMPACT is an excellent example of thetype of interdisciplinary approach required in CAIsystem development. The staff is composed ofbehavioral scientists, mathematicians, engineers,instructional programmers, and computer softwareprogrammers. Progress in this complex develop-mental effort has been substantial and shouldproduce an optimal CAI system suitable for large-scale implementation.

The University of Illinois Computer-BasedEducation Research Laboratory

One of the oldest CAI research facilities wasestablished iu 1959 at the University of Illinoisunder the PLATO program-PLATO is an acronymfor Programmed Logic for Automatic TeachingOperations. The PLATO system has grown from asingle terminal connected to an ILLIAC I com-puter to 20 graphic pictorial terminals driven by aControl Data Corporation 1604 computer.

In 1963, a development effort was started on alow-cost, interactive graphic terminal combinedwith a random-access image selector. The terminalis based on the use of a plasma display 'panel whichis under development at the University of Illinois.It is hoped that the panel which has the propertiesof memory, display, and high brightness can beproduced commercially at low cost. The panelitself is a thin glass structure filled with gas cellsthat are capable of being selectively ignited undercontrol of the computer. Production of a low-coststudent terminal with the characteristics of aplasma-type terminal will constitute a major costbreakthrough in the development of economicallyviable computerized instruction. The possibilitiesof a large-scale application of computer-basededucation have been summarized as follows:

1. Gradual abolishment of lock-step schedulesand narrowly specified curricula in formal educa-tion. Students could proceed at a pace determinedby their own capacity and motivation.2. Provision of remedial instruction or tutorialassistance during regularly scheduled courses forstudents with insufficient preparation.3. Reduction in the number of large lectureclasses at the college level, in favor of small instruc-tional groupings and seminars.4. Special instruction at home for physicallyhandicapped students.5. Development of arithmetical or other skills, atthe elementary level, away from the exposed andoften competitive environment of the classroom.6. Effective job training or retraining for anyemployee group especially affected by expandingtechnology.

7. Continuing education for professional person-nel, permitting the updating of knowledge andskills in their own offices and on their ownschedules (Alpert & Bitzer, 1969).

Some of the areas in which course material hasbeen presented have been electrical engineering,geometry, biology, nursing, library science,pharmacology, chemistry, algebra, math drill,computer progranuting, and foreign languages. Inpresenting these materials, a wide range ofcomputer application modes have been usedranging from drill and practice to student-directedinquiry. The use of the inquiry mode is anextension of the tutorial made. This methodallows the student to control his interaction withthe system, investigate, and gather additionalinformation to solve a problem presented to himby the computer. This mode has been used veryeffectively in medical science, engineering, andphysical science presentations.

In evaluating the educational effectiveness ofthe PLATO system, the relatively small sample

sizes involved in the experimental studies precludeattaining results considered as definitive. Alpertand Bitzer do, however, present the followingconclusions:

1. That computer-based education is a plaustbleapproach to improved individualized instruction ina very wide array of courses or subject materialareas.

2. That the nature of educational testing andevaluation calls for and will be radically and sub-stantially affected by the availability of largecomputer-based education systems; a valid measureof effectiveness calls for a much larger sampling ofdata and a Larger period of comparison than hasheretofore been available (Alpert & Bitzer, 1969).

The PLATO program as it has evolved throughthree systems (PLATO I, II, and III) has been oneof the few that have included an effective researchand development program on all aspects of com-puterized instruction. This effort has involvedinnovations in computer hardware and software aswell as course materials and modes of application.

University of California at Santa BarbaraComputer Center

The Computer Center at The University ofCalifornia at Santa Barbara (UCSB) installed anIBM 360 Model 50 in 1966 to implement anadvanced instructional and research system. Dis-play consoles were placed in five departments foruse M a variety of courses including chemistry,calculus, electrical engineering, sociology, eco-nomics, and psychology.

The system uses a display console based on aTektronix Direct View Storage Tube (DVST) anda double keyboard separated into an operator orprogram control keyboard and an operational key-board with an alphanumeric capability. The unitprovides an excellent alphanumeric and graphicdisplay. The console, called Teleputer, wasmarketed commercially by Bolt, Beranek andNewman. The system is a time-shared, on-line,interactive computational application based on atechnique called on-line computation developedby Dr. Glen J. Culler and Dr. Burton D. Fried.

The system is an excellent example of thepotential of an interactive problem-oriented lan-guage used with a large-scale time-sharingcomputer. The system has evolved from the IBM.360 Model 50 through and IBM 360 Model 65 andnow is based on an IBM 360 Model 75. The 2075central processing unit has 500K bytes of corestorage and two million bytes of Large CapacityStorage (LCS), four 2311 disk drives, one 2314disk drive, two 1403 printers, two 2401 tape

4

13

drives, one 1402 card reader/punch, and 60teleputer consoles, 45 located on the UCSBcampus and 15 at other locations across theccuntry.

In January 1^70, the National Science Founda-tion announced an experimental program toexplore the use of the UCSB on-line system inundergraduate instruction in chemistry. Studentsat UCSB and nine other institutions will use thesystem for problem solving. An earlier experimentat UCSB showed that use of the system in teachingchemistry allowed the students to discover forthemselves properties of model chemical systems.Through use of the on-line system and relatedteaching techniques, it is hoped to greatly improveinstruction to the point that first and second yearstudents will understand concepts at a senior andgraduate student level.

The UCSB on-line system operates in a time-sharing mode and allows use of the computersystem for processing batched programs for re-search and other university use.

Naval Personnel and Training ResearchLaboratory, San Diego, California

Since 1967, the Computer-Assisted instructionResearch Department of the United States NavyTraining Research Laboratory has been developingand testing CAI course modules in basic elec-tronics. Course material has been developed inDirect Current Circuit Analysis, AlternatingCurrent, Inductance and Capacitauce, AC SeriesCircuits and Resonance, and AC Parallel Circuitsand Resonance. An IBM 1130-based 1500 Instruc-tional System with 15 instructional terminals isused to present the material. Each terminal isequipped with an audio unit, CRT with light pen,an image projector, and a keyboard. IBMCOURSEWRITER II is used for the CAI program-ming language. This system presents the elec-tronics course modules as tutorial sequencesinterspersed with problem solving and simulationexercises.

The project has developed and tested a varietyof instructional strategies. Dr. John Ford, Jr.,director of the project, describes two ways theterm instructional strategy is used to describe thetraining presented.

First, it refers to the sequence or flow of instruc-tion for a training objective. This is a more globalstrategy and encompasses many instructionalpresentations or "frames." In the second usage, theterm refers to techniques used within an instruc-tional frame. In this latter sense, we are explainingmany methods of bringing student responses understimulus control and of various types of feedback(Ford, 1970).

14

In evaluating the results of one of the CAImodules, it was found that those students trainedin using CAI methods scored significantly higherwhen tested on the CAI material than did thecontrol group. A time saving of from 33 to 44percent was also realized by use of the CAImodules. Student reactions to the CAI trainingwas also very favorable. Ford (1970) summarizedsome of the additional results of the project asfollows:

In addition to evaluating the quality of CAI inachieving training objectives, a better under-standing has been gained of the type of learningenvironment which is afforded in this CAI systemand of the system characteristics needed to supporta research and development project in CAI. A CAIsystem is an extremely powerful tool for discover-ing and analyzing learning problems of individualsand groups of trainees. The effects of a trainingmethod are made brutally obvious down to thefine grain step-by-step training sequence. First,student performance records are examined to de-termine whether the behavior was learned duringthe training sequence. Next, a check is made on theamount of practice which was provided. Then, testperformance is examined to determine the longerterm effects cf training. In this way, the causes ofineffective training can be identified and related tospecific training events. Also, the effects of differ-ent instructional strategies can be assessed muchmore accurately than before. Self-pacing hasproved to be a very effective technique foroptimizing training achievement . . . .

There is a growing emphasis today on develop-ing CAI capabilities which can provide significantamounts of education or training for a given pro-gram. Although there are problems, the potentialfor achieving this goal looks good. Some of thepioneering work in CAI focused upon research inlearning and instruction. We are beginning torealize or perhaps re-discover, how powerful a toola CAI system can be for such research.

Other Navy Programs

In addition to the project at the Training Re-search Laboratory at San Diego, the Navy is devel-oping a Computer-Managed Instructional system inthe Naval Air Technical Training Command atMemphis, Tennessee. In this system, the computeradministers the instructional process, but does nottake part in the actual instruction. The UnitedStates Naval Academy also uses an IBM 1500 In-structional System to help teach credit courses inRussian, thermodynamics, physics, and fluid dy-namics. The system also serves as a research tool tohelp shape the Academy's approach to computer-assisted instruction.

These projects demonstrate the interest in theapplication of the computer to instruction on thepart of the training establishment of the Navy.

Stanford University Programin Computer-Assisted Instruction

The Institute for Mathematical Studies in theSocial Sciences began a research and developmentprogram in computer-assisted instruction in Janu-ary 1963, under the direction of Dr. PatrickSup pes. Two basic approaches to computer-assisted instruction are followed by the Institute:drill and practice, and tutorial. The tutorialapproach to CAI uses the computer as a "teacher"to present new concepts as well 'as to determinesubsequent student work with the concepts. Incontrast, drill and practice systems supplementclassroom instruction by improving the skills andconcepts introduced by the classroom teacher(Suppes, Loftus , & Jerman, 1969). The institutehas developed drill and practice programs inmathematics, reading, and spelling for elementaryschoOl age children. Tutorial programs in elemen-tary mathematics,, reading, logic/algebra, and twoyears of college-level Russian have also beendeveloped by the Institute.

The tutorial Russian program provides com-prehension of spoken and written Russian as wellas training in grammar. and syntax. The Russianlanguage program uses a teletype terminal with aCyrillic keyboard and computer-controlled audiotapes recorded by a Russian language lecturer. Thecomputer system, which provides all regular class-room instruction, has been very successful.Computer-based instruction in languages appearsto have great potential at the secondary andcollege level.

The computer complex used by the instituteutilizes a Digital Equipment Corporation PDP-1,PDP-8, and a PDP-10 with 192K words of Ampexcore storage, two IBM 2314 disk files, two 2401-4tape drives (1600 bytes per inch), and a 2803 tapecontrol unit. The system is currently driving up to150 teletype terminals. Philco Ford "READ,"CRT consoles are used for curriculum develop-ment and course revision.

An important facet in the development of theStanford CAI project is the data analysis andresearch activity relating to curriculum develop-ment and the development and testing ofinstructional strategies expressed as mathematicalmodels. The use of CAI as a tool in educatidnalresearch has been clearly shown to be a majorcontribution.

Before the advent of computers, it was extremelydifficult to collect systematic data on how childrensucceed in the process of learning a given subject

A computer can provide daily information

6

15

about how students are performing on each part ofthe curriculum as it is presented, making it possibleto evaluate not only individual pages, but alsoindividual exercises (Suppes, 1966).

Evaluation of the effectiveness of the StanfordCAI program has shown significant gains in stu-dent achievement in the three years the system hasbeen used in an operating school environment. Acomprehensive evaluation of the mathematics andRussian language programs points out the diffi-culties in providing a "hard data" evaluation ofCAI programs (Suppes & Morningstar, 1969).

Atkinson and Wilson (1968) also comment onthe inherent complexities of evaluation of CAIprograms:

Evaluation of a computer assisted instructionprogram is only partially an evaluation of thesystem and equipment. Primarily, it is an evalua-tion of the instructional program and as such isbasically an evaluation of the program designerwho is the real teacher in a computer-assistedinstruction system. The evaluation question thenbecomes, "to what extent did the curriculumdesigner provide the computer with an appropriateset of instructional materials and an adequatedecision structure for branching among them?"

This same article also points out that measuringeffectiveness of CM is also difficult because of a"current lack of a sound theoretical basis for des-cribing levels of learning and achievement."

The Stanford CM project has developed aunique approach to the course design in both theelementary mathematics and reading programs.Th,.: course context is divided into "strands"which are defined as basic component skills. Thestudent's instruction is individualized as he prog-resses through the various strands based on hisperformance. The mathematics curriculum has 15strands, and the reading program has six strands.Progress within a strand is based on performancecriteria in that a student goes on to a new exerciseonly after successful completion of the currentexercise. A reporting system provides the teacherwith the individual status of each student in eachstrand and a status report for the entire class. Thisreport assists the teacher in providing remedialwork and determining areas requiring additionalemphasis.

Another application of the drill and practicemathematics program is a project in the New Yorkschools called "Dial a Drill." Using a touch-tonetelephone receiver as a terminal, 2,400 studentcan receive automated arithmetic drills. Thestudents receive three 5-minute drills per weekthat are automatically adjusted by the computer

to the level at which the student can performsuccessfully. The project is demonstrating thepotential of CAI in providing instruction in basicskills using a very low-cost terminal.

This brief summary, of the Stanford CAIproject shows the progress that has been madefrom its modest beginning to a 150-terminalsystem serving schools located in six states and theDistrict of Columbia.

The University of Pittsburgh LearningResearch and Development Center

The Learning Research and DevelopmentCenter was established at the University ofPittsburgh in April 1964. The general purpose ofthe Center is the

. . . . scientific study of the problems of learningand instruction with particular attention to thenature of the educational and psychological envi-ronment required to maximize the potential of theindividual learner. The broad goals of the Centerare the following: (1) To contribute to the growthof those aspects of behavioral science which arerelevant to educational practice. (2) To developprototype models of instructional procedures andappropriate hardware and software through the in-volvement of scientists, engineers, and educators.(3) To develop and evaluate experimental instruc-tional systems, including materials, proceduralcomponents and personnel training (Yeager &Glases, 1968).

The center is currently conducting exploratoryresearch projects in stimulus control, learningstrategies, memory, and psycholinguistics. Thisresearch is centered in the Basic Learning StudiesProgram. The CAI program is concerned with the"analysis of learning processes as students work inpartially automated environments, the design ofexperimental student stations, the study anddevelopment of effective CAI courses and theevaluation of the potential role of CAI in a rangeof educational applications (Yeager & Glaser,1968)."

The center has experimented with the use oftouch-sensitive display, audio and visual accessdevices, and CRT displays. Much effort has alsogone into research efforts concerned with the useof the computer in the management of the educa-tional process. Ferguson (1970) points out theservices that the computer can provide in this area:(a) collection and storage of data, (b) informationretrieval, (c) testing, (d) instructional planning,and (e) instruction.

A pilot project at the Oaldeaf School insuburban Pittsburgh involves an educational data

7

1.6

system for both classroom management and re-search purposes. The system wies an IBM 360Model 50 computer located at the UniversityComputer Center, an optical scanner, a cardpunch/card reader, and a typewriter console at theschool. A typewriter terminal at the LearningCenter also provides access to the research data forthe project staff. The system allows the classroomteacher on-line (over telephone lines) rapid accessto student instructional management data. Thefeedback of information on the instructionaloperation at the school is being used to improvestudent performance. The computer provides adata storage and analysis capability that would notbe feasible by any other means.

Results of this project and studies ofComputer-Assisted Testing (CAT) and Computer-Assisted Instruction (CAI) have resulted in anotherproject involving a small local computer at aschool. The proposed terminal configurationincludes teletypes and cathode ray tubes. Theproject will test the hypothesis that "a small com-puter can support educational management,computer testing, and some CAI lessons(Ferguson, 1970)."

Another of the center's major programs isIndividualization of Education. One of theprincipal projects in this area has been the develop-ment of a system called Individually PrescribedInstruction (IPI). The system is designed toprovide instruction tailored to a student's individ-ual abilities and needs. The system was firstoperated in the Oakleaf School in 1964 and is nowbeing used by 160 schools in 32 states. The systemprovides instructional materials in elementarymathematics, reading, science, handwriting, andspelling. The system was originally nonazitomated,but a computer-based management informationsystem was added to facilitate collection of datafor system improvement and analysis. The WIsystem is based on an instructional model de-scribed as a sequence of operations as follows:

(1) The goals of learning are specified in terms ofobservable student behavior and the conditionsunder which this behavior is to be manifested. (2)When the learner begins a particular course ofinstruction, his initial capabilities those relevantto the forthcoming instruction are assessed. (3)Educational alternatives suited to the student'sinitial capabilities are presented to him. Thestudent selects or is assigned one of thesealternatives. (4) The student's performance ismonitored and continuously assessed as he learns.(5) Initruction proceeds as a function of therelationship between measures of student perform-ance, available instructional alternatives, and

criteria of competence. (6) As instructionproceeds, data are generated for monitoring andimproving the instructional system (Cooley &Glaser; 1969).

Exploratory research and development is alsobeing conducted in the use of computers insecondary schools. The program proposes a varietyof hands-on experience for both teacher andstudents. The experiment will explore the applica-bility of a general "computer- activated learning"model, develop a supportive structure for theapplication, and develop specifications for a con-trolled dissemination program.

The work of the center has been summarized inwhat the faculty considers the

. . .themes of investigation which are of specialsignificance to research and development relevantin education. These themes are the following: (1)The importance of contact with scientists, engi-neers and subject matter scholars and the mutualinterplay between basic and developmental activ-ities so that a body of technology can be madeavailable for the design of educational practices.(2) The importance of adapting educational envi-ronments to the requirements and requests of theindividual learner. (3) The significance of evalua-tion not only for the assessment of studentprogress in the course of his education, but also forthe monitoring of educational procedures andmaterials in order to redesign and improve them,and to put their respective contributions to theeducational process into the proper perspective(Yeager & Glaser; 1968).

Computer Terminal and VisualDisplay Devices

Another important aspect of this survey wascollection of data on the type of readable (charac-ter display) terminals available for use incomputer-based instruction. This was of immedi-ate concern since character terminals would be thefirst used in the system, and since most characterterminals are easily interfaced. The primary infor-mation which was sought were the parameters ofcost, display speed, ease of interfacing, noise, andlegibility. While information is available in allother parameters, no information was found onlegibility, although observed terminals were found.to differ visibly in this respect.

The following section summarizes characteris-tics of each family of terminals although it mayidentify the class by a particular brand name ofterminal. This is not an endorsement of anyparticular terminal over other terminals, but isused to improve identification of the class bymentioning more commonly seen devices.

8

17

Since the visual information given to a studentnormally contains a greater amount of informationthan any other sensory input, it is important thatthis input be quite readable and thus not easily tirethe student. Equally important, since a great dealof information will be presented over the terminaldevice, its operation must not overly tax thecomputer. Since this part of the system will alsobe a part of almost every student terminal, it mustalso be reasonably priced with respect to itsperformance.

There are many different ideas about whatdevice constitutes a perfect terminal; in mostcases, the differences of opinion are legitimate.There was, however, no evidence of a carefulexperimental evaluation of a number of thedifferent devices along common criteria. Rather, ineach case, some particular constraint (e.g., cost,environment, innovation) dictated the particulardevice to be used.

The devices are discussed in the context ofwhere and how they were used in an educationalsituation, with an attempt to pinpoint why theywere chosen for the situation. Those devices notobserved in a CAI application are also described,and probable CAI application is mentioned.

Teletype ASR-33 Terminals

The teletype terminal is undoubtedly the mostcommonly used terminal device in the country.Constituting the basic terminal in the CM systemssold by RCA and Hewlett-Packard and somesystems sold by IBM, the teletype terminal is prob-ably the least expensive device currently availablefor sale. Its primary attributes as stated by mostusers are: (a) low cost; (b) hard copy output andkeyboard; (c) ease of interfacing with computers;(a) ease of remote use over telephone lines; and (e)availability of service.

Similarly, a great number of complaints werefound criticizing the device, including: (a) highnoise level; (b) variable reliability; (c) slow displayof information; and (d) lack of graphic capability.

To understand why these terminals werechosen, one needs. only to look at their intendeduse. In the Stanford CAI program, the teletypewas selected for use in the elementary schoolsystem for administering mathematics and spellingdrill and practice (Suppes et al., 1969). In theapplication, the amount of information to beviewed at one time was limited to several lines,such as a simple math problem or a single line ofwords to be spelled. In the mathematics drill, the

student was not reading or considering compli-cated text material, so the background noisegenerated by several teletype terminals was notconsidered too great a distraction. Since studentswere wearing headphones in the spelling drill, thenoise in that instance was somewhat attenuated.On the other hand, a silent CRT display mighthave been desirable, since only small chunks ofinstructional material were displayed at a time.However, in this case, cost was the overridingfactor, since there were 150 terminals in thesystem This is the same reason RCA markets theA512-33 with their Instructional 70 system forsimilar drill and practice uses.

Similarly, the HP2000A is marketed byHewlett-Packard with teletype terminals or theirplug-to-plug equivalent. It was stated by severalmanufacturers that a CRT display could be madecost-competitive with teletype displays if suffi-cient sales. volume .could be reached to justify ahighly automated assembly line. Such a marketdoes not currently exist, however.

Dot Matrix CRT Displays

The dot matrix CRT display was found in onlyone CM system as a student terminalthe IBM1500 system. This particular CRT allows not onlycharacter display, but also has a limited graphicscapability. Lines which may be drawn as a trail ofdots on the screen are adequate for low-resolutiondisplay of data. The screen image is actuallygenerated by a raster scan technique, but the rasteris one of point-addressable dots. Although the dis-play is legible, it is more difficult to read thanmost digitized television displays, and is prone to aslight flicker and jitter (vertical vibration).

The 1502 CRT includes a light pen devicewhich allows the student to touch the screen witha probe. The CAI program can establish pro-grammed meaning to locations on the screen andthus determine if the student touched one of themeaningful locations.

The 1502 CRT was introduced as the terminaldevice in the 1500 system in order to obtain feed-back from its users which IBM could use toconsider the design of a next generation CAIsystem. At this time, no such improved machinehas been produced, and the 1502 CRT is neithermanufactured for nor supported by the 360 seriesof IBM computers.

The attributes of the 1502 CRT as stated byIBM are: (a) very rapid display of information; (b)graphics capability; (c) light pen; (d) high perform-ance per cost in ,1500 context; and (e) silent

9

18

operation. The disadvantages are: (a) no hardcopy; (b) video disk unit required to refresh dis-play so no telephone lines can be used betweencomputer and terminal; and (c) less than perfectvisual display could be tiring if much text weredisplayed.

This display device was chosen by IBM ratherthan by the system purchasers, since it was to bean IBM research and development device. There-fore, the applications did not always reflectappropriate ,Jse of this CRT in the way that otherapplications directed the purchase of teletypes.One exemplary use of the system at FortMonmouth and at the Navy Electronics School inSan Diego is the experimental teaching of a blockof basic electronics on the 1500 system. In thiscontext, the CRT is used as a text and questiondisplay device and as a light pen or keyboardresponse device. There is nothing in this type ofusage which definitely requires CRT or light penuse, but the devices are being used experimentally,since they are included in the current 1500 systemconfiguration.

In a statistics teaching program at Florida StateUniversity, the graphics ability of the scope is usedto display statistical distributions generated in theCAI package as examples of the statistical distribu-tion types being explained to the students.Implementation of APL as the computer languageallows graphical display of results of computa-tions, but does not support light pen use.

IBM 2250 Vector Display Scope

The IBM 2250 vector display is marketed byIBM for use with the 360 series of computers as agraphics display device. Its display is a set ofvectors which can connect any two points on thescreen and which are refreshed from core memoryby a small controller associated with the 2250device. The 2250 offers extensive graphicscapability, especially in situations where the dis-played information must change frequently. Whileits use as a student terminal is untested, it offersthe advantages of: (a) graphical versatility; (b)rapid display change; and (c) large screen size. Itsdisadvantages, in many ways stemming from itsadvantages, are considered to be: (a) high cost; (b)excessive flicker and image instability; and (c)poor character display.

The primary observation of the 2250 was in itsuse in the POGO (Program-Oriented GraphicsOperation) graphics system at the Rand Corpora-tion. In this use, as an author graphics device, theuser could generate line and figure drawings from a

Rand tablet (a two-dimensional touch-sensing in-put device) and manipulate their size and orienta-tion on the 2250 screen. In this use, the rapiddisplay changes on the screen were quite useful aswas the vector-generated display. The POGO sys-tern offers much potential as a device for gener-ating automated graphical displays which could bephotographed and displayed on mouon picturefilm. Similarly, it could be used to constructschematics and hydraulic diagrams, or logic,PERT, or flow charts, all of which require the typeof graphics offered in the 2250.

The use of the 2250 as a student terminalwould have to be considered quite carefully as thecost of the terminal and computer supportrequired is high in comparison to the supportrequired in teletype terminals. There are severalother types of vector displays available, notablyfrom the ADAGE Corporation, but these devicesare also quite expensive and require computermainframe support of some type. Technology isthus available for graphic display of information,but the cost involved is an overriding factor inapplications not absolutely requiring the dynamicterminal.

Digitized Television Display (CRT Terminal)

Digitized television is a basic technique andcomes in many variants depending on the exactuse of the device of which it is a part. A digitizedTV display could appear as an array of characterssimilar to the dot matrix display, as a dot graphicdisplay, as a bilevel (black and white areas only)picture, or as a normal TV picture with gray tonesas well as black and white areas. The technique ofusing raster lines to compose characters is the basisof 20 to 25 terminal devices recently introducedfor use with computers. In these devices, typically,the set of characters to be displayed is transferredto and stored in the terminal in MOS (metal-oxidesemiconductor) memory devices from which thedisplay is hardware refreshed. These devices are allplug compatible with teletype units so they can beused interchangeably with ASR-33 units. Their dis-play change speeds can generally be adjusted fromteletype speeds, up to 100 times teletype speeds,and some terminals offer storage of several screensfull of characters within the terminal. Prices rangefrom approximately $1,500 for 32 by 12 characterdisplays up to $5,000 for 60 by 24 by 4-pagedisplays.

The primary advantages for this type of displayaxe: (a) no refreshing required from computer; (b)teletype compatibility and telephone lineconnection; (c) silent operation; and (d) highlylegible display with very little jitter and no flicker.

10

These advantages are again weighted against thetypical disadvantages of: (a) higher cost than tele-type terminals; (b) no hard copy; and (c) nographic capability. No CAI uses of this type ofterminal .were observed. Several facilities usingteletypes expressed a desire (but not the necessity)to use this type of terminal to replace existingteletype units; they did not do so because of thehigher cost. Although the basic terminal can notdisplay any different type of information thanthat displayed by means of a teletype terminal, ithas the advantage of silent and faster displaygeneration. These devices should probably becompared in an educational stituation todetermine when the advantages of speed andsilence would justify the additional cost of theCRT terminal.

Digitized Television Driver

This unit is really a variant of the digitized tele-vision display. Rather than building a separatecharacter generator, memory, and keyboard, atelevision monitor or home TV set is used as thedisplay device. The digital driver simply connectsto the monitor or sweep circuit of the home TVand acts as an interface to the storage device orcomputer which is generating the raster. Typically,a video disk is used to furnish the raster; the dis-play driver may contain a character generator tomake the display look like the display of thedigital terminal. These devices are sometimes use-ful in applications where a large number ofidentical displays are needed. The advantages are:(a) low cost when multiple, identical displays areneeded; and (b) silent display. However, the dis-advantages are: (a) high cost of video disk; (b)displays must be located near driver; (c) TV tubesmust be augmented by keyboards for use asterminals; and (d) non-character display rastersrequire a great quantity of storage.

Although none of these systems were observed,a system based on this technology is beingdeveloped by the Mitre Corporation. Mitre basesestimates of the system's low operating cost on theeconomy of size (10,000 terminal system) and onlow cost of the computers used as simple storage/retrieval devices for the television rasters. Thesystem limitations which have been noted are notconsidered serious since the system was designedas a presentation &Nice for use in elementaryschool systems where the information is primarilytext display to the students.

Direct View Storage Tubes (DVST)

The direct view storage tube is an outgrowth ofoscilloscope and storage device technology which

has only recently been in use in CAI systems. Theobserved usage of the device was in a computa-tional assistance system programmed as back-ground processing on a 360-75 at the University ofCalifornia at 'Santa Barbara (UCSB). The DVSTwas used 'as the output device in the system andwas interfaced to the 360-75 by hardware built atUCSB. 'The DVST, built by Tektronix Corpora-tion, has a 4 by 4 inch viewing surface upon whicheither characters or any graphics information maybe displayed. While the screen does have address-able points, the display is drawn with continuousvectors. This device provided thc highest qualitydisplay observed both for graphics and forlegibility.

The DVST requires no computer refresh sinceimage persistence is obtained by. the physicalcharacteristics of the DVST. This same characteris-tic results in an image which is totally free offlicker and jitter. The only drawback is that it isbest viewed in a dimly illuminated room.Character generators at UCSB were located in theinterface unit and shared among 32 terminals, butthe character display bay be generated as a dotmatrix or line vectors, and may be generated eitherby computer software or terminal hardware asdesired. At that time, the price of the basic tubewas about $600, with UCSB figuring the price ofeach student terminal plus interfacing and128-character keyboard at about $1,200.Tektronix has since introduced a larger 16 by 21cm DVST at a cost \-1f $2,500. This DVST could beused either as a large student terminal or as anauthor graphics terminal with POGO type soft-ware. The advantages of the DVST are: (a) lowcost in 4 by 4 inch size; (b) flicker and jitter freedisplay; (c) no refresh required; (d) highly legibledisplay; (e) graphics capability; and (f) silent dis-play. The disadvantages of the DVST are: (a) nohard copy; (b) dim lighting or screen hoodrequired for viewing; and (c) must be augmentedby keyboard and interface to computer.

In the context of the UCSB system, the DVSTis an excellent output device since it is a studentterminal with graphics and does not require aspecial purpose driving system such as the IBM1502 scope on the 1500 system. Since studentterminals are all located within a reasonableproximity of the main computer, no special linesare required, and the terminals are compatiblewith voice grade telephone lines. The UCSBsystem drives 16 terminals located at variouslocations around the United States and a total ofabout 45 terminals at UCSB, all of the DVST type.Since the terminals are silent, many can be located

11

20

in the same room without the problain of nciseinterference from nearby terminals.

Computer-Controlled Optical Projectois

A computer-controlled optical projector con-sists of some type of film media (such as strip,slide, or microfiche), an indexing mechanism, anda projection screen. The types observed were astrip and slide projection device and an indexibleprojector. There have been many indexible filmdevices, but probably the first under computercontrol was the projector available as part of theIBM 1500 system. The programmer could pro-grammatically select individual frames to displayto the student and ask questions or present text onthe adjacent 1'502 CRT terminal. No light pentype device was available for the projector, but itwas possible to augment the presentation withvisual displays. The major advantages in this typeof device are: (a) presentation of photographicinformation; (b) presentation of color coded infor-mation; and (c) local storage of photographicmaterial. The disadvantages are: (a) difficulty ofpreparation of course material; (b) difficulty ofalteration of course material; and (c) difficulty ofchanging material at terminals.

These devices were observed only in experi-mental use at the Navy Electronics School in thecourse in basic electronics. The slides augmentedthe programmed instruction being presented onthe 1500 system.

Addressable Projection Display

The addressable projection device with lightpen is essentially a frame-selectable, cartridge-loading projector, with a probe that the studentcan point at program-addressable places on theprojected display. This means that photographs,photographed wiring, hydraulic diagrams, orphotographed devices could be displayed to a stu-dent, and the student could input to the computereither questions or answers about specific places inthe device. While this device was not observed inoperation in a CAI system, it has obvious ad-vantages over CRT displays for certain applica-tions: (a) information may be graphics, characters,or photographs of actual devices; and (b) studentinputs in two-dimensional language (i.e., a place onthe screen) rather than with verbal response.Weighed against these advantages, there are theconsiderations that: (a) film must be loaded atdevice rather than at computer; (b) no hard copy;and (c) must be augmented by some other termi-nal device with keyboard.

This is not a terminal by itself but is, rather, adifferent type of input/output device and shouldbe considered as an addition to rather than a sub-stitution for the normal terminal. The device ismarketed by IBM and. is compatible to their 360series of computers.

Owens-Corning PhotosensitiveGlass Terminal

This terminal is a new device based on new glasstechnology. Its display medium is a glass platewhich is sensitive to ultraviolet and infrared light.Upon exposure to ultraviolet light, the glassdarkens and may be mace transparent again byexposure to infrared light. The device isfunctionally similar to the direct view storage tubein that either graphic or alphameric informationmay be displayed on the tube, and the imageremains until the entire screen is erased byexposure to infrared light. Unlike the direct viewstorage tube, nothing can be displayed withoutputting it permanently on the tube. To showsomething temporarily, it must be added to thedisplay. To erase the temporary addition, theentire screen must be erased, and all but thetemporary addition must be redrawn.

The advantages of this device are primarily theadvantages of the DVST except for the fact that ahard-copy device is built into the terminal, and theprice is about $19,000. The advantages enumer-ated are: (a) hard copy output; (b) silentoperation; (c) rapid change of contents of screen;(d) graphic capability; (e) no refresh required fromcomputer; (j) highly legible display; and (g)flicker- and jitter-free image.

The terminal was very recently introduced andhas not been observed in any CAI system. Whilethe device offers many potential uses as a terminalin a system requiring graphic capability, the cost isquite high for consideration as a student terminal.No information was initially released regarding theinterfacing of the device with telephone lines.

Portable Printing Terminals

The portable printing terminals are a group ofterminals which are very similar to teletype unitsin function. They consist of a keyboard usuallycompatible, with the ASCII coded teletype-character set and a printing device, usually a wirematrix. rather than an impact printer. The wematrix printing unit allows a reduction in weightand noise of the unit and results in a package thatweighs typically under 30 pounds. An accostic

12

21

coupler is built into the unit which allows it to beused with any common telephone.

The major advantage of this unit is portabilitywhich allows it to be taken wherever the terminalis needed. The unit can be used as a nonportableunit, and the acoustic coupler avoids the cost of adataphone while allowing the terminal to be usedat an individual's desk through normal telephones.The only disadvantage of the device in thisapplication (cowered to a teletype) is that it ismore expensive than teletype and usually requiresheat-sensitive paper as compared to the very low-quality, low-cost paper required by the teletype.This terminal's advantages are: (a) portability; (b)compatibility with telephone lines; (c) low noiseoperation; and (d) hard copy. Its disadvantagesare: (a) higher cost relative to teletype; and (b)special paper required.

These units were all recently introduced andoffer minimal advantages in the typical CM class-room consisting of fixed terminals. The portabilityof the device is useful where a student must betaken to some large, possibly immobile trainingdevice. Perhaps the application of the portableterminal as an author terminal, which can go tothe author as opposed to the author going to theterminal, offers the most effective use of theauthor.

Portable CRT Terminals

The portable CRT is a recent developmentanalogous to the portable printing terminal. Itoffers the same advantages to a fixed CRT as theportable printer offers to the teletype user. Theadditional advantages over the fixed CRT are: (a)portability; and (b) compatibility to telephonelines.

The disadvantage is higher cost than a fixed.CRT, part of which is the acoustic coupler cost.For the same reason that portable printing deviceshave not found their way into CAI applications,the portable CRTs have also not been used. Thesame potential advantages are present as for port-able printers, but they have not yet beenexploited.

Color CRT Displays

The color CRT offers a quite recent and inevit-able extension to CRT technology. In addition tonormal alphameric display in black and white,current color CRTs have the ability to displaycharacters in (typically) eight colors, or as black

characters on a colored background. The additionof color to the display is seen as an advantagewhen the terminal is used to input data or tomonitor operations which require quick operatorattention. It is difficult to estimate the value ofcolor display in a CAI application. Since the dis-play tube used in these color displays is a dot-matrix shadow-mask tube of the type used inhome color receivers, the display suffers from thesame lack of resolution characterized by small-screen color television.

The advantage of th. color CRT over the blackand white CRT is simply the ability to displaycolor characters. The disadvantages are: (a) highercost; (b) lower resolution in display; (c) loss ofdirect teletype compatibility; and (d) unknownvalue in CAI applications.

This device was not observed in any CAIapplications and does not appear to have beenused in any to this time. These devices wereapparently introduced as data entry terminals andthus do not feature very large screens and arelimited in the number of characters they displaysimultaneously. Refreshing information is storedin the terminal, so no computer-generated refresh-ing is squired. The keyboard is usually teletypecompatible, but the use of the color capabilityrequires that additional information be sent to theterminal to specify the color of the displayed text.

Computer-Controlled Speech

In certain applications it is desirabl3 to augmenta verbal or visual presentatioi with audio infor-mation. The devices discussed in this section areused to generate the audio presentation which ismade to the student. Some of these devices do notinterface to the computer as plug-in units butinstead require direct computer time to producesound ficm stored information. Other devices aremerely switched on or off by the student or by thecomputer. All are concerned with audio present-ation of information.

Computer- Controlled Sequential Tape

The basic design of this device is a tape recorderwhich can be started and stopped by a computerprogram. The computer program presents a sectionof tape to the student, and then asks him aquestion either orally or on a teletype unit. Thecomputer then corrects his response if needed andpossibly repeats the question or adds additionalmaterial until the student responds correctly. Inany case, the student goes to the next sequentialunit of sound on the tape. Thus the computer is

13

22

used to receive responses, correct mistakes, keepan item score and student score internally, andcontrol the rate of the presentation. Use of audiobrings in an additional sense (hearing), commandsattention (the student must listen at presentationrate and must be alert), and allows lecture-likematerial to be presented at any time of the day;illness or temporary interruptions of the course donot cause the student to miss any of the coursematerial. Every student answers every questionpresented in the course. The student is immedi-ately corrected if wrong.

The validity of this type of presentation hasbeen demonstrated in a two-year Russian languagecourse offered for credit at Stanford University byMrs. Elise Belenky. Comparison of studentperformance in lecture courses as compared to thecomputer-administered lecture course shows a 40percent reduction in dropout rates (based on30-student classes) and far superior performanceon paper-and-pencil tests administered to bothgroups. The improvement is due in no small partto the quality of the material which has beenrefined and developed over a three-year period byMrs. Belenky. The permanence of the presentationonce developed ensures that good material ispreserved while bad material can be easilychanged. The nature of this system makes itexcellent for languages, but it also applies to anyinformation which can be presented verbally in alecture. There is also a possibility of using anytype of film projector, also computer-controlled,to augment the audio material and a veryinexpensive device to respond to the computer, atouch-tone phone. If the system could becombined with a random access audio device, anentire lecture could be presented over a telephoneby letting students simply call into the computeron telephones.

Computer-Controlled Random AccessAudio Device

This typ4 of device consists of a band of shorttapes. disks, or drums which can be easily playedin a random order under computer control. Thistype of equipment, which typically handlesphrases or commonly used audio messages, is oftenused for temporary recordings such as stockquotations, weather messages, and airline flightinformation. A device of this type might be usefulin correcting student responses made on thesequential-type presentation which has beendiscussed. At the present time, this deviceapparently is not used by any educationalinstitution. In actual use, this device has charac-teristics similar to digitized speech.

Digitized Computer-Composed Speech

Digitized speech has been used by severaluniversity and research organizations and, experi-mentally, by commercial computer corporations.In operation, a computer accesses randomly froma vocabulary of words or phrases stored incomputer disk or drum memory. These words arepresented as audio information to desired studentterminals. Speech is actually stored by taking in-coming speech waveforms and sampling theiramplitudes at a rate of approximately 30 kc. Theresulting amplitudes are compressed to bit strings,and the bit strings are stored on computer disks.By catenating these strings, phrases or sentencescan be constructed. Because of the current state ofthe art with regard to speech quality, digitizedspeech is probably best used for short responsesneeded frequently or in a particularly randomorder.

A system teaching phonetics and spelling tograde school children in the Brentwood School ofEast Palo Alto, California, makes exclusive use ofdigitized speech to present material and informstudents when they have submitted correctanswers through their keyboards. Students wearheadphones to isolate their sound ftom etherstudents and to isolate them from the noise of theteletype terminals they use. Compared to theperformance of students taught in conventionalmanners, superior performance has been shown forstudents using the system for 10 to 15 minutes aday (Suppes, 1966).

Digital/Mechanical Time-Compressed Speech

Time-compressed speech is normal humanspeech which has had selective portions of eachword and pause removed and the remainingportions catenated together and smoothed. Thesimplest form of compn ssor is an electro-mechanical device which physically chops andthen packs information onto a magnetic tape. Thisdevice deletes fixed length segments at varyingspeeds to achieve compression to greater then 50percent of original length. Although frequency, ortone, does not vary in the process, the rate ofpresenting the material increases. This bothreduces storage requirements for information anddecreases the required presentation time of theremaining material.

An alternate approach for time compression isto digitize the speech with a computer and thenmechanically discard repetitive subsections of thespeech (as described in the preceding paragraph) ordelete "redundant" information according to some

14

23

algorithm. The sound information is most mani-pulative when in digitized form, but it requiresconsiderably more equipment to process indigitized form. Compression is more or less "free"in any environment which is comprehensiveenough to perform digitization of speech since theelimination of substrings of speech is relativelytrivial once strings art digitized. Compression ofcomputer storage requirements can reduceconsiderably the cost of storage equipment for aparticular size vocabulary.

Very little work has been done to determinethe effectiveness of digitized speech in training,but more general studies have led to the con-clusions that it enhances attention by the listenerand requires less time to impart a given level ofretention. It might be particularly useful for end-of-course reviews or film narration to reduce timeand production cost of such materials.

Touch Tone Telephone Entry

Touch-tone entry constitutes the cheapestpossible type of data entry when the data are to betransmitted over the public telephone network.The touch-tone user first calls the computer bypunching in the number to establish the line. Fromthis point, any pushed buttons transmit data tothe computer which was called. The computermay also respond to the caller with three tonesgenerated by the receiving unit (Western Electric403A MODEM), or it may respond with verbalinformation using any of the techniques whichhave been described.

Touch-tone entry could be used as a very in-expensive way to administer tests and direct study,dispense information, review courses, or processincoming student data. It could also be used withspecial texts to teach material to remotely locatedstudents, or even to administer language courses ashas been indicated. The limited number ofresponses (buttons) on the pushbutton panelsuggests that some uses are more appropriate forfull teletype keyboards, but the 10 to 12 buttonpanel is quite adequate for many applications. TheIBM 2721 portable audio terminal is an inex-pensive extension to any telephone which allowsfull keyboard input where it is needed.

CAI Languages

After completion of the field trips andcollection of literature, all documentation on CAIlanguages was examined in an attempt to deter-mine the advar_tages and disadvantages of

languages being used for CAL This documentationwas supplemented by a limited review of manu-facturers' literature and professional journalarticles describing languages." It was decided tocategorize the languages by the purpose for whichthey were originally intended and then comparethem with examples from typical members of eachcategory.

The fundamental purpose of a CAI language isto communicate with a student through amechanical medium. If the language is incapable ofcommunicating the desired information, the infor-mation will either be distorted or will be preser redinadequately for learning to occur.

There are many potential types of informationwhich can be displayed; some can be easilyhandled internally by computer, and some can beeasily accessed at the direction of a computer.Upon looking at many of the CA! languages avail-able today, one would probably assume that theonly displayable information is typed text and theonly reasonably analyzable answers are in suchforms as "A," "B," "TRUE," "NO," or a number.From a knowledge of other languages, it appearsthat minor variations of this information might beobtained, but at the, expense of hundreds of hoursof author effort for every hour of communicationtime. It is perhaps reasonable to spend one-hundred hours planning an hour of communica-tion, but once the message has been determined, itshould not be necessary to spend a noticeablefraction of the preparation time actually codingthe message.

This is not true with some of these languages.Some languages recognize that certain types ofmaterial require special devices under programcontrol and, in some cases, simple computationsupon the data input obtained from the student.Other languages recognize the need for modularityin order to minimize data transfer with the con-trolling system. Still other languages reflect thelimited capability of the systems which implementthem.

For purposes of this discussion, CAI languagesare classified into four basic categories. Thesecategories are based on the original purposes oftheir defnm languages rather than the types of.material presented when used as CAI languages.After discussion of the categories in general terms,a more detailed description is presented of thatone language which best typifies the use of alanguage from its category in CAI application.

The four categories have existed for some timeand were recognized in a paper by Dr. CharlesFrye (Frye, 1968). He notes the following fourcategories:

Category 1. General Purpose Problem Lan-guages

Category 2. Extended Existing Problem Lan-guages

Category 3. Interactive Terminal-Problem Lan-guages

Category 4. Author-Student Problem-OrientedLanguages

Each of these categories can be typified by thefollowing members:

1. PL-1, ALGOL, COBOL, FORTRAN, andLISP

2. SUPER-BASIC, FOIL, and CATO

3. APL, BASIC, QUICKTRAN, and JOSS

4. CO URS EWRITER , HONEYWELL AU-THOR LANGUAGE (HAL), PLANIT,DIALOG, and TUTOR

In an attempt to clarify the categories, each isdiscussed along with a description of one majorlanguage within the category.

Category 1. General Purpose Problem Lan-guages. The characteristic of most languages in thiscategory is that they are oriented toward mathe-matical problem coding with emphasis upon accessto typical peripheral devices in computer systems.These languages are most frequently furnished"free" with computers in that their developmentcost is included in the price of the system. BothFORTRAN and COBOL software are oftenrequired for consideration of the system in agovernment purchase; consequently, they are themost commonly found languages. IBM offers. PL-1for its 360 computers; several other corporationsoffer ALGOL, the Burroughs version being themost versatile implementation. Since FORTRANis the language most commonly offered on bothlarge and small computers, it may be considered astypical of this set of languages.

While FORTRAN is ill- suited to CAI applica-tions, it is possible to use it. The basic datastructures of FORTRAN do not include "strings"(i.e., groups of characters which may be mani-pulated by name). FORTRAN implementers havelong recognized this flaw, and most working

15

24

versions of FORTRAN allow for arrays of data tobe initialized as strings (i.e., CAI messages) or tobe set to strings by FORTRAN read statementswhich ignore the non-numeric nature of the databeing read into numeric-labeled storage areas. Sub-routines may be coded so that through numericoperations the types of operations which areusually done with strings are simulated, providingthe program is run on a particular machine withknown. binary representations of character strings.In particular implementations which allowFOTATIRAN programs to call assembler languagesubroutines, it is possible to extend the basicFORTRAN by defining CM-oriented functions.FORTRAN implementations on machines whichsupport da'ta communications terminals frequentlyallow these terminals to be used as input/outputdevices; an interactive situation may thus be esta-blished between the. FORTRAN program and thestudent user at the terminal. The following set ofFORTRAN statements may be used to provide acomment to a student sitting at a terminal on a32-bit word, 8-bit byte machine where the termi-nal is defined as device number 7.

INTEGER NOUT (5) PTHE EARTH IS FLAT?

33 FORMAT (61)°,5A4)

WRITE (7,33) NOUT

The message printed to the student would be"THE EARTH IS FLAT." The first statement isused to define the storage area NOUT and to giveit the initial value which is the printed message.The second statement is used to define a techniqueused to arrange and display the message to thestudent. The third statement causes the displayaction to be executed and specifies the device, thedata arrangement, and the actual data to beshown. This is more complicated than the equi-valent statement in COURSEWRITER (i.e., TYTHE EARTH IS FLAT) for several reasons:

1. FORTRAN does not assume that a studentis the input/output recipient.

2. FORTRAN users often need to explicitlyposition their output on a page.

3. FORTRAN does not include strings, so thedata have to be disguised as an array ofintegers.

If the author wished to do computation in theprogram using a student response, he would find itquite easy to do, since computation was theoriginal purpose of FORTRAN. For example:

16

2i

REAL ANSWER, AREA, R, SAY /'GOOD'/,NONOrBADV

33 FORMAT ('b', 1A4)

13 FORMAT (F3.1)

R = 3.0

40R=R+ 1.0

50 AREA = 3.14 * R * * 2

WRITE (7,13) R

READ (7,13) ANSWER

IF (ANSWER. EQ. AREA) WRITE (7,33) SAY

IF (ANSWER. NE. AREA) WRITE (7,33) NONO

GO TO 40

In this example, the author is picking values ofradius and is asking the student to calculate thearea of a circle. The program computes the correctanswer (statement 50) and then compares thecorrect answer with the student answer receivedfrom the terminal. Only one statement is requiredto do the computation (statement 50), whereas inCOURSEWRITER several staatements would berequired. If it were desired to keep a record ofcorrect and incorrect answers, a statement such as:

WRITE 78(,43) AREA, ANSWER

43 FORMAT (2F3.1)

could be added to write the responses on a filedeclared as device 78 (perhaps a tape or disk file)which would store both answers each time throughthe loop.

By the examples given, it should be obviousthat some CAI could be done in FORTRAN;however, the things frequently done, reading andwriting strings, are a harassment to the author. Inaddition, the author must be quite familiar withFORTRAN to exploit its capabilities. It is oftennecessary for the FORTRAN program to beentered through a card reader and compiled in itsentirety; thus, an author cannot debug and checkhis program as it is generated but must build itcompletely. (Reference to unfurnished materialcauses syntax errors in FORTRAN.) This couldprove to be a source of problems since the coursewriter would likely be unfamiliar with the materialand with computers in general.

The most obvious advantages of FORTRANare: (a) availability; (h) ease in performing

arithmetic calculations; and (c) accessibility tomost computer peripherals.

The many disadvantages for CAI applicationsare: (a) difficulty of expressing string data; (b)difficulty in analyzing student responses; (c)normal environment of FORTRAN is batchprocessing so a great deal of core storage may berequired for each student; (d) programs arenormally entered in batch mode rather than fromterminals;. (e) author must be a skilled pro-grammer; and (f) courses programmed are difficultto examine and understand since so much datamust be disguised, to fit into allowable FORTRANdata types.

Category 2. Extended Existing Problem Lan-guages. When a category 1 language is extended tomake it easier to use, a category 2 language results.Development of such a language can be showntaking FORTRAN as its base. Suppose the fol-lowing construct were added to the FORTRANbasic data types:

STRING X, Y,

and the assignment as well as the write statementwere to be modified to refer implicitly to astudent terminal. Then the FORTRAN write state-ment could be added as follows:

WRITE 'THE EARTH IS FLAT'

or at least

X = 'THE EARTH IS FLAT'

WRITE X

Similarly, a routine called READNUMBER (X)might be supplied which could accept a studentresponse and treat it as a number, storing it inpropel' form as a real number X. A function mightalso be supplied called KEYWORD (A,B) whichwould be interpreted as true whenever the studentresponse stored as B contained an occurrence ofthe author-defined string A. The compiler mighteven be rewritten so that authors could enter eachline of the program from a terminal and have amistake corrected immediately. The implementa-tion could even be modified so that each authorstudent running a FORTRAN program-could sharecommonly used program routines, reducing therequired storage. Languages such as CATO andFOIL are examples of languages based uponFORTRAN which use these types of extensions toincrease the basic capabilities of FORTRAN.

There are a number of advantages which arisewhen a category 1 language is extended to includeCAI constructs. A major advantage is the reduced

17

28

cost of extending an existing compiler as com-pared with the cost of pt.ogramming a completenew language. The cost is frequently prohibitive,or sometimes capable people are not available atany price to do the job. The fact that a compilerfor the language will'probably be furnished with anext machine generation ensures that this startingbase will be available if the current machine has tobe replaced later.

Another major advantage is that knowledge ofthe language is generally more common when apopular language is extended. This can result inthe easy addition of analysis capability in the lan-guage by adaptation of commonly found sub-routine libraries already programmed in thestarting language. Such a package is obtained fromFORTRAN in the Scientific Subroutine Package.This package of matrix handling and statisticalroutines would be as available in the extension asit was in the type I language in which it wasoriginally written.

The primary disadvantages to this technique arethose 'associated with the operating environmentof category 1 languages: (a) frequently large corestorage requirements; and (b) programs normallyentered in batch mode rather than from a terminalin a conversational manner.

Category 3. Interactive Terminal-Problem Lan-guages. Category 3 languages are defined as inter-active terminal languages. The better of theselanguages attempt to reduce the time required toget a program into operation, as they are usedfrequently in an environment where expensivescientist time and expensive on-line computer timeare both being consumed.

A typical and frequently found example of thistype of language is BASIC. The BASIC language issimilar in many ways to FORTRAN, but the basicstructures have been modified along the followinglines:

1. Input/output is easier in BASIC than inFORTRAN.

2. Individual lines of code in BASIC do notinteract as much as individual statements inFORTRAN or ALGOL.

3. Strings of characters can be output inBASIC without explicit storage.

4. Computer peripherals are usually notsupported by BASIC (although this varies fromimplementation to implementation).

Since the exact features available in BASIC varywith implementation, BASIC is discussed here in

the context of Hewlett-Packard Time-SharingBASIC. The normal method of using BASIC is tobuild a BASIC program in conversational modewith each line checked as it is entered. When theauthor finishes his program, he may execute it bytyping in a RUN command. The program isexecuted (usually interpretively) until completion,unless it is terminated by error or by the author'spressing of the BREAK key in his console.

An example of a BASIC statement to print themessages shown earlier would be:

10 PRINT "THE EARTH IS FLAT'

Since BASIC knows the user is at a terminal, it canassume the terminal is the output device, and thatthe author knows the message will be positionedagainst the left margin; the need for printed stringswas recognized in BASIC. (BASIC was developedmany years after FORTRAN and thus did nothave to retain compatibility with earlier baddecisions.)

BASIC was also developed as a computationlanguage, so it is not difficult to obtain resultsrmilar to those in the FORTRAN example given.

The following BASIC sequence can accomplish thesame goal:

10 FOR R = 3.0 TO R STEP 1.0

20LET AFtEA = 3.14 * R f 2

30 PRINT R

40 INPUT ANSWER

50IF ANSWER = AREA THEN 80

60 PRINT "WRONG"

70 GO TO 90

80 PRINT "GOOD"

90 NEXT R

It should be noted that this algorithm stated inBASIC is two lines shorter than the FORTRANalgorithm. It is also easier to read than theFORTRAN routine which does the same thing.The readability factor should be considered whenmaterial is to be modified by persons other thanthe original author.

While the format is not ideal for constructingdisplay frames, it entails fewer cryptic coding stepsthan FORTRAN, and it does perform a drill-and-practice type teaching lesson. If a mistake hadbeen made in entering the program, the authorwould have been notified immediately, and theerror could have been corrected immediately. Tocheck the performance of the material, it can be

18

27

"run" immediately, so that the author knowsquickly whether logical errors have been made.These are definite benefits to the unsophisticatedprogrammer.

The major advantages of BASIC are: (a) avail-ability; (b) conversational implementation; (c) easeof performing arithmetic computations; and (d)ease of performing input/output operations.

The major disadvantages of BASIC are: (a)difficulty of analyzing student responses; and (b)necessity for the author to be reasonably skilledalthough errors are easier to detect.

Category 4. Author-Student Problem-OrientedLanguages. Category 4 languages are those mostoften billed by manufacturers as "CAI languages"for their systems. Their primary purpose is theconstruction of frames of information for presen-tation to the student. hi this type of languages, theauthor is primarily furnishing data which are to behandled in orthodox, standard manners by a pro-gram handler and which assume a whole set ofconventions for presenting data to students.Typical assumptions in these languages are:

1. Text must be presented for reading by thestudent.

2. Questions should be multiple-choice, true orfalse, or based on keywords submitted bythe student.

3. Arithmetic to be done is trivial.

4. Computer peripheral support is not availablein CAI systems.

Whether these assumptions are valid is deter-mined by the hardware selected and the nature ofthe course material. If the hardware in the systemdoes not include projectors and sound devices,then constructs to drive them are superfluous. Ifsuccessful instruction depends on input and dis-play of analog voltages, and the language offers theprogrammer no access to such devices, then thelanguage is inadequate. Of the various languagesfitting category 4, it is safe to say that none ofthese languages contain all the requirements of areference language for CAI algorithms.

There is a reasonable variance among the lan-guages in this category. PLANIT and HAL areprobably representative of the extremes of flexi-bility. HAL is an assembler format language whichcontains virtually no arithmetic capability andmust be entered on cards into the computer.PLANIT is a highly conversational language withon-line computation ability and an almost

free-form-decision-frame structure to controlbranching. In this range of flexibility, COURSE-WRITER is more versatile than HAL, less versatilethan PLANIT, and probably the most commonlyused language due to the IBM 1500 systeminfluence. Because of these characteristics,COURSEWRITER is explained in this section.

COURSEWRITER is an assembler format lan-guage in that each author instruction consists of atwo-letter mnemonic and an 'operand list. Thesemnemonics are used to designate frames of infor-mation to display to the student. For example, thestatement:

RD THE EARTH IS FLAT

will display the frame "THE EARTH IS FLAT" tothe student and wait until he signals, by typingand response, to continue. If the student shouldsign off at this point, he will restart here the nexttime he signs on the machine.

Certain mnemonics are classified as major andothers as minor operators. Implicit branching isperformed on the basis of these operations. Theexecution of a major operation such as QU, thequestion operator, results in . all minor operationswhich follow QU being executed. Thus, in thesequence

1 2 QU IS THE EARTH FLAT? (MAJOR)1 2AD1/C5 (MINOR)1 CA YES (MAJOR)1 TY VERY GOOD (MINOR)

2 WA NO (MAJOR)2 TY SORRY, YOU ARE INCORRECT (MINOR)

1 2. QU IS THE SEA BLUE?

the statements labeled "1" would be executed ifthe student responds "YES" to the first question,and the statements labeled "2" would be executedif the student responds "NO." Notice that in anycase the next question is presented. If the studenthad signed off instead of answering the question,then he would restart the lesson with this questionwhen he next signed on the system. Themnemonics CA, WA signal that their fields, if theymatch the student answer, are interpreted by thesystem as (CA) correct or (WA) wrong answer andthus should be recorded accordingly. Note theinstruction "Al) 1/C5." This instruction adds 1 tothe counter C5. Similar constructs exist forsubtract, multiply, and divide. Thus, to achieve thesimple arithmetic example shown in the

19

28

FORTRAN example (statement 50), the sequencewould be replaced by the following COURSE-WRITER sequence:

FORTRAN: AREA = 3.14 R * 2

COURSEWItITER:

SB AREA/AREAAD 3.14/AREAMP AREA/RMP AREA/R

COMMENTSET AREA TO 0ADD IN 3.14 AS A FACTORMULTIPLY BY R FOR 3.14 * RMULTIPLY BY R FOR 3.14*R**2

Thus, the simplicity of expressing text material iseasily overshadowed by the complexity of per-forming arithmetic in an assembler-like manner.

Since the 1500 system included audio andvisual devices, certain mnemonics for positioningthose devices were included in COURSEWRITER.The "FP" operator positions a peripheralto a selected frame. The addition of amodifier (e.g., "FP(p)") causes the selected frameto be displayed to the student.

Explicit branching is allowed by use of the"BR" branching operation. The branch may bemade to any labeled point in the program, i.e., toany frame. It may also be made to implicit returnpoints, such as the last restart point on the lastwrong answer return point. The branching may beunconditional or conditional upon numeric .testsof counters, switches, or combinations of numericand switch tests. In this way, the author can"remember" student performance over severalframes and control branching on this basis.

Certain other operations exist in COURSE-WRITER which allow the author to recognize asloppy, misspelled, or nearly correct studentanswer, but most of these are coded as functionswhich modify the student response before it iscompared to the answers given by the author.

The major advantages of a COURSEWRITERtype language are: (a) ease of displaying writtentext; (b) ease of entry to the system; (c) ability toaddress terminal peripherals such as projectors andtapes; and (d) ability to be interpretively executed(low core requirement).

The primary disadvantages are: (a) emphasis onverbal information as opposed to arithmeticoperations; (b) pattern of frame is pre-determinedby language rather than at option of writer; and(c) low readability since branching and flow iscontained in intrinsic attributes of mnemonicsrather than explicitly visible in the program code.

It should be mentioned that PLANIT doesallow a separation of frame' data and branchinglogic and is more readable as far as program logic isconcerned. One unfortunate characteristic ofPLANIT is that its complexity almost requires aninterpretive implementation; the known examplesof PLANIT have been encoded in type 1 languagesas interpretive systems in a lime-sharing operatingenvironment. PLANIT does represent considerableinnovation as compared to COURSEWRITER andshould be examined as an example of a high-levelauthor language (Feingold, 1968).

HI. CONCLUSIONS AND RECOMMENDATIONS

Terminal Devices

Many useful and sophisticated terminal devicesare currently available which would be suitable foruse in a computer-based technical training system.Most of these devices can be easily attached toexisting computer hardware. Other devices areavailable which may be useful for collection anddisplay of audio, voltage, temperature, pressure,and electrical waveform, but they are not directlyinterfaceable to all computer mainframes. Foreconomical use of central processors, and for eaR1of interfacing, the current best method is to inter-face nonstandard devices or large collections ofinput/output devices through small processorscalled mini-, (midi-, maxi-)' computers. Interfacingin this raaniier enlarges the class of suitablemainframe processors to include most majormanufacturers and, in many cases, eliminates theneed for special hardware design to perform theinterconnection.

Languages

Nonstandard terminals sometimes requirespecial software to make them useful in instruc-tional contexts. A course writer entering materialinto an instructional system must be able to easilyaccess special devices in his instructional program,particularly while he is collecting data (verbal orotherwise) to be

toto the student. It is

unreasonable to expect current software tosupport input/output devices not yet available oncomputers. However, it is not unreasonable toexpect languages- for future instructional systems

20

29

to interface these new devices so that they areavailable to course writers; such facilities should beconsidered in the initial design of such languages.

Another important feature of a course writinglanguage is the ability to support the author inon-line data entry. The language should allow theauthor to devise nonstandard presentationsequences which can use the unique featuresavailable in an advanced computer-based instruc-tion system. By allowing the author to devise neweducational strategy and easily implement it, oneof the greatest assets of the computer, the ease ofchanging data, can be of genuine use to theinstructor. This very important attribute is notsupported in any of the CAI languages which havebeen described in documentation.

Suggestions for Research

As an initial step in the development of anadvanced instructional system, the following stepsshould be undertaken:

1, Investigation of available mini-computers todetermine the best unit for interfacing non-standard terminals and peripherals to various main-frame hardware.

2. Thorough investigation of terminal charac-teristics to determine the cost-effectiveness of thevarious types of terminal devices.

3. Immediate research to design and determinevalidity of instructional techniques and languageswhich would become available upon acquisition ofa computer-based instructional system.

Step 1 will be undertaken as part of the con-tinuation of this CAI survey project; it shouldprovide specifications for a terminal controllerwhich would make the research in step 2 possible.With the available hardware to perform step 2, itshould also be possible to interconnect with aprototype system which would make step 3 possi-ble. The limiting fcactor in the time sequenceindicated would probably be the time required toget from the specifications of step 1 to the avail-able equipment of step 2 (i.e., procurement lagtime) and from the conclusions of step 2 to theavailability Of hardware for step 3. At step 3, timedelay will depend primarily upon software devel-opment time (a function of the system selectedand available software for the particular com-puter). It would be possible to perform steps 2 and3 simultaneously if adequate funding andflexibility in hardware selection permit the use ofadequately programmed and interfaceablemainframes currently available.

REFERENCES

Alpert, D. & Bitzer, D. Advances in computer-based education. Report X-10. A ProgressReport on the PLATO Program, University ofIllinois, Computer-Based Education ResearchLaboratory, July 1969.

Atkinson, R.C.,& Wilson, H.A. Computer-assistedinstruction. Science, 4 October 1968, 162,73-77.

Cooley, W.W. & Glaser, R. The computer and in-dividualized instruction. Science, 1969, 166.

Feigold, S.L. PLANIT - a language for CAL Data-mation, September 1968, 41-47.

Ferguson, R.L. Computer-assistance for individ-ualizing instruction. Computers and Auto-mation, March 1970, 27-29.

Feurzeig, W. The Socrates system, a computersystem to aid in teaching complex concepts.Status Report No. 3. Cambridge, Mass.: Bolt,Beranek and Newman, Inc., 1964.

Ford, J.D. Developing and testing CAI coursemodules in basic electronics. Proceedings ofconference on application of computers totraining Washington: National Security In-dustrial Assoc., February 1970.

Frye, C.H. CAI languages: Capabilities and applica-tions. Datamation, September 1968, 34-37.

Hansen, AN., O'Neil, H.F., Brown, B.R., Rivers,L.C., & King, A.D. Annual Progress ReportJanuary 1, 1966 through December 31, 1969.Report No. 8. Computer-Assisted InstructionCenter, Florida State University, February1970.

Seidel, R.J. Project impact: Computer-administered instruction concepts and initialdevelopment. Technical Report 69-3. TheGeorge Washington University, HumanResources Research Organization, March 1969.

Suppes, P. & Morningstar, Mona. Computer-assisted instruction. Science, 17 October 1969,166,343 -350.

Suppes, P. The uses of computers in education.Scientific American, September 1966, 215 (3).

Suppes, P., Loftus, E..F., 7 Jerman, M. Problemsolving on a computer based teletype. TechnicalReport No. 141. Stanford University, Institutefor Mathematical Studies in the Social Sciences,March 1969.

Yeager, J.L. & Glaser, R. The Learning Researchand Development Center at the University ofPittsburgh. The Journal of Research andDevelopment in Education, Summer 1968,31-44.

21

30

UnclassifiedSecurity Classification

DOCUMENT CONTROL DATA - R & D(Security classification of title, body of abstract and indexing annotation must be entered when the overall report is classified)

1. ORIGINATING ACTIVITY (Corporate author)

Technical Training DivisionLowry Air Force Base, Colorado 80230

20. REPORT SECURITY CLASSIFICATION

2b. GROUP

3. REPORT TITLE

APPLICATION OF COMPUTERS IN EDUCATIONAL AND TRAINING SYSTEMS: A SURVEY OF COMPUTER-ASSISTED INSTRUCTIONAL CENTERS

4. DESCRIPTIVE NOTES (Type of report and inclusive dates)

5. AUTHORISI (First name, middle initial, last name)

Gerald S. WalkerEdward M. Gardner

6. REPORT DATE

December 197078. TOTAL NO. OF PAGES

21 I

lb. NO. OF REFS14

8a. CONTRACT OR GRANT NO.

b. PROJECT NO. 1121

c. Task No. 112102

d. Work Unit 112102 002

9a. ORIGINATOR'S REPORT NUMBEFUS;

AFHRL-TR-70-24

95. OTHER REPORT NOIs1 (Any other numbers that may be assignedthis report)

10. DISTRIBUTION STATEMENT

Approved for public release; distribution unlimited.

11. SUPPLEMENTARY NOTES 12. SPONSORING MILITARY ACTIVITY

Technical Training DivisionLowry Air Force Base, Colorado 80230

13. ABSTRACT

This paper presents an overview of representative centers active in the research, development, and application ofthe computer in education and training. Field visits were made to military and civilian agencies in order to obtain across-section of various types of applications and research projects. The information presented is based on the fieldvisits and the most recent reports available on the activities of the various centers. Computer terminals used incomputer-assisted instruction are discussed, and computer-controlled audio presentation is described. Programminglanguages which have been used in computer-assisted instruction are categorized, and a typical language from eachcategory is examined.

-..t NOV 65

31

UnclassifiedSecurity Classification

UnclassifiedSecurity Classification .

1 .KEY WORDS

LINK LINK B LINK CROLE WT ROLE WT ROLE WT

computer-assisted instructionCAItrainingtechnical training

instructioncomputer-based oncomputer-managed instructioncomputer terminal

UnclassifiedSecurity Classification

32