TEE BEE’ELGFMENT OF A CQMPUTERHED MODEL FOR TEKCHING EfiGENEERING STATECS A Dissefloflon §or Hie Degree of p}! D. EICBiGAN STATE UNIVERSITY John William Johnson 19174 ' ‘I I This is to certify that the thesis entitled‘ . THE DEVELOPMENT OF A comerfikgm MODEL FOR TEACHING ENGINEERING STATICS presented by JOHN WILLIAM JOHNSON has been accepted towards fulfillment of the requirements for PHD' degree in HIGHER EDUCATION divé/ ; Major professor Date 42114127?” 0-7639 l girl/I that rapic anal) indiv ABSTRACT THE DEVELOPMENT OF A COMPUTERIZED MODEL FOR TEACHING ENGINEERING STATICS BY John William Johnson A.means was sought in this study of demonstrating that a digital computer could be employed to make a rapid, thorough, and efficient check of a student's analysis and synthesis in problem-solving courses with individualized feedback. It was the purpose of this study to develop a computer related instructional model capable of checking a student's analysis and synthesis of a broad range of statics problems on an individualized basis. The research was guided by the following objectives: (1) the devel- 0pment of the model would follow a systems analysis and design approach which would serve as a basis for the development of additional models for other problem-solving courses; (2) the student would be given the opportunity to address his attention solely to the analysis and synthesis portions of problem-solving by being relieved of the mechanics of calculations; (3) selection of the computer tranSpor evaluate The educatio in its c an IBM 3 would ac. core ava. be used, able for required and eithc Ext] model to and SYStE (3) the E (4) the F statics, The Statics F was Fermi for the S meending Called th JOHN WILLIAM JOHNSON computer and computer language would be based on maximum transportability of the model; (4) the model would be evaluated for preper operation and educational impact. The model was developed so it could be used by any educational institution with a relatively low investment in its computer hardware system. It was developed on an IBM 360-22 computer, but any digital computer which would accept Basic Fortran IV and had 16K(decimal) of core available for the model plus system overhead could be used, with minor modifications, if means were avail- able for overlaying subroutines. No terminals were required, any standard statics textbook could be used, and either the vector or scalar approach was permitted. Extreme caution was taken in the development of the model to be certain that: (1) data capture was simple and systematic; (2) very little keypunching was required; (3) the program could easily be debugged by the student; (4) the program would not interfere with the learning of statics. The model was a simulation of the solution of a statics problem. Any two- or three-dimensional problem was permitted if only one free-body diagram was required for the solution, except problems with friction at impending motion. The student entered pertinent data and called the appropriate subroutines to perform the required calculati gave the in the fc synthesis It V student \ and/or a students consiste in an en computer Populati Student: not Com: St Culty JOHN WILLIAM JOHNSON calculations. The computer then solved the problem and gave the student feedback (both positive and negative) in the form of diagnostics concerning his analysis and synthesis. It was assumed that the model could be used by a student with no previous experience with a keypunch and/or a computer. Therefore, two pOpulations of students were used in testing the model. One population consisted of eight sophomore engineering students, enrolled in an engineering statics class, who had completed a computer course in Fortran IV programming. The second population consisted of twenty-seven freshmen technology students, enrolled in a technology statics class, who had not completed a computer course in Fortran IV programming. Students' attitudes toward the computer related instructional model were measured by administering a pre-test and post-test to both populations. Engineering and technology students agreed that: (1) little keypunching was required; (2) no previous experience with a keypunch was necessary to effectively use the model; (3) previous experience with a computer was necessary or at least desirable; (4) the model did not interfere with the learning of statics. The engineering students experienced little diffi- culty using the model, considered the diagnostics to be valuable, themodel a valuable mieffect: tematic ll not agree items. When who taugh were cert from the probleuns Students, Addi model has fricthn two 01. NC experimer an exper: mEnt eff, Shou1d b. area of . groups a Should b achierm JOHN WILLIAM JOHNSON valuable, found the program easy to debug, considered the model a good teaching tool, considered the model a valuable asset to learning, and considered the model an effective means of helping them to become more sys- tematic in problem-solving. The technology students did not agree with the engineering students on any of these items. When the experiment ended, both faculty members who taught the technology class emphasized that they were certain their students would have gained much more from the model if they had been required to turn in all problems assigned as was required of the engineering students. Additional experiments should be conducted when the model has been further developed to include problems with friction at impending motion and problems which require two or more free-body diagrams for the solution. The experiments should be designed with a control group and an experimental group in such a way that the only treat- ment effect would be the use of the model. An instrument should be developed to measure achievement in the cognative area of learning which would be administered to both groups at the end of the experiments. The test results should be used to determine significant differences in achievement of the two groups. sis. 1’. compare using t if the from th JOHN WILLIAM JOHNSON A further study should include a cost-benefit analy- sis. A cost analysis of the use of the model should be compared with increases in student achievement when using the model. The analysis would be used to determine if the added cost when using the model could be justified from the added benefits the students would realize. THE DEVELOPMENT OF A COMPUTERIZED MODEL FOR TEACHING ENGINEERING STATICS BY John William Johnson A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Administration and Higher Education 1974 Chapter I II III Chapter I II III TABLE OF CONTENTS THE P ROBLEM O O O O O O C O O O O O O 0 Background . . . . . . . . . . . . . . . Statement of the Problem . . . . . . . . Purpose of the Study . . . . . . . . . . Importance of the Study . . . . . . . . Operational Definitions . . . . . . . . Overview . REVIEW OF RELATED LITERATURE . . . . . . Background of CAI . . . . . . . . . . . Problems Associated with CAI . . . . . . Recommendations to Improve CAI . . . . . NSF Experiment . . . . . . . . . . . . . DESIGN OF Section I: THE STUDY . . . . . . . . . . Development of the Model . . Problem Recognition . . . . . . . . Feasibility Study . . . . . . . . . Goals and Objectives . . . . . Information System's Blueprint Major Equipment Decision . . . ii Page 11 12 14 15 15 l7 19 21 24 24 25 26 26 27 3O Chapte Ki? Chapter IV TABLE OF CONTENTS (continued) Implementation Planning . . Traditional Information System Redesign - . . . . . . . . . . Scope and Objectives - - . . Analysis . . . . . . . . . . Specifications - - - . . . . Design . . . . . . . . . . . Implementation - . . . . . . Information System Mechanization Information System Modification - Section II: Student and Faculty Analysis Popu1ations o o o o o o o o o . o o Instrumentation - . - . . . . . . Data Collection - - . - - . . . . Analysis of the Data - - - - - - ANALYSIS OF THE DATA . . . . . . - . . summary 0 o o o o o o o o o o o o o 0 SUMMARY, CONCLUSIONS, RECOMMENDATIONS, IMPLICATIONS FOR FURTHER DEVELOPMENT AND RESEARCH . . . . . . . . . . . . Summary . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . Recommendations . . . . . . . . . . . iii Page 35 37 37 39 46 50 65 68 68 70 7O 70 72 72 77 89 94 94 97 Chapte TABLE OF CONTENTS (continued) Chapter Page Implications for Further Development and Research . . . . . . . . . . . . . . . 99 SELECTED BIBLIOGRAPHY . . . . . . . . . . . . . . . . . 102 GENEm REFERENCES 0 O O O O O O O O O O O O O O O I 0 lo 3 iv Table 10 ll l2 l3 14 15 16 17 Table 10 11 12 13 14 15 16 17 LIST OF TABLES Estimates of the Volumes of Data Required . . . Ratings of Computer Systems . . . . . . . . . . Data Structures Noted in the Function Flowchart for Stage I of the Analysis . . . . Tabulation of Data for Concentrated Forces, Designated: FORCES: Concentrated . . . . . . Tabulation of Data for Points and Lines, Designated: PTLINE . . . . . . . . . . . . . Tabulation of Data for Distributed Forces, Designated: DLOAD . . . . . . . . . . . . . Tabulation of Data for Moments and Couples, Designated: COUPLE . . . . . . . . . . . . . Summary of Data . . . . . . . . . . . . . . . . Tabulation of Data . . . . . . . . . . . . . . Arrays Used in the Information System . . . . . Student Introduction to Subroutines . . . . . . Core Required for Subroutines and Main Program Pre-test Results for Engineering Students in category 1 O O O O O I O O O O O O O O 0 0 Post-test Results for Engineering Students in category 1 O O O O O O O O O O O O O O O O Pre-test Results for Technology Students in Category 1 . . . . . . . . . . . . . . . . Post-test Results for Technology Students in Category 1 . . . . . . . . . . . . . . . . Results of Faculty Questionnaire . . . . . . . Page 30 36 44 51 51 52 52 53 55 64 68 69 79 80 81 82 88 Table 18 19 20 21 LIST OF TABLES (continued) Table Page 18 Summary of the Analysis of the Data for Engineering Students' Pre-test and POSt-tESt o o o o o o o o o o o o o o o o o o 90 19 Summary of the Analysis of the Data for Technology Students' Pre-test and Post- teSt O O O I O O O O O O O O I O O O O O O O 91 20 Summary of the Analysis of the Data for Engineering and Technology Students' Pre-teSt o o o o o o o o o o o o o o o o o o 92 21 Summary of the Analysis of the Data for Engineering and Technology Students' POSt-teSt o o o o o o o o o o o o o o o o o o 93 vi Figur LIST OF FIGURES Figure Page 1 Flowchart of the Solution to a Problem . . . . 7 2 Function Flowchart for Stage I of the AnaIYSis O O O O O O O O O O O O O O 0 O o O 44 3 Flowchart of All Possiblities When Subroutine MFGRR Solves the Matrix . . . . . 61 4 Operational Flowchart for Stage I of the Design . . . . . . . . . . . . . . . . . . . 63 vii Appendi LIST OF APPENDICES Appendix Page A Student and Faculty Instruments . . . . . . 105 B User's Manual . . . . . . . . . . . . . . . llO viii many of t. Sali: use c and j perfc PaYrc Ancil simpl fUnct Can a “auth inst}: CHAPTER I THE PROBLEM Background The uses of the computer in support of education are many, and the field is rapidly expanding. Proliferation of terminology has been one result of this rapid growth. Salisbury has suggested three functional areas for the use of computers by educators: administrative; ancillary; and instructional. The administrative functions are those performed in direct support of the administrator such as payroll, record keeping, scheduling, and counseling. Ancillary applications are those in which the computer is simply a tool for problem—solving. The instructional functions serve the "learner" element of the system and can augment or replace the "materials," "monitor" and/or "author-teacher' elements. Salisbury further divides the instructional applications as follows: a. Computer-Administered Instruction (CAI): A man-machine interaction in which the teaching function is accomplished by a computer system without intervention of a human instructor. Both training material and instructional logic are stored in computer memory. (Also referred to pOpularly as computer—assisted instruction.) b. Computer-Supported Instruction (CSI): All computer applications in support of instruction in which the computer is used by a human in- structor to assist him in the accomplishment 1 Comp written 1 tutorial: and probl defines 6 Use moni tas} a p1 This var: unde Salisburj The and mat. stu« gen. rat‘ dri tut com The dial Con est EDd his \ . 1A1. ‘AS’I'EQmen net, 197 2 Te A1 a 3Ibi 2 of his instructional objectives; essentially all uses of the computer as a classroom training aid. Computer-Assisted Instruction (CAI) programs can be written in any of the following modes: drill and practice; tutorial; dialogue, conversational, or socratic; simulation; and problem-solving. The U. 8. Continental Army Command defines drill and practice as: Use of the computer to guide, control and monitor by repetition a specific task or set of tasks. The purpose of this mode is to develop a predetermined level of proficiency in a skill. This proficiency may be changing under a wide variety of constantly changing conditions or under a single set of consistent conditions.2 Salisbury has described the tutorial mode as: The tutorial mode is more complex than the drill and practice mode in that more instructional material is presented and more sophisticated student responses are often called for. It is generally used for presenting original instruction rather than supplemental, as in the case of drill and practice. More than any other, the tutorial mode exemplifies the augomation by computer of the programmed text. The dialogue, conversational, or socratic mode has been defined as: Conversational or socratic systems attempt to establish a two-way dialog between the student and the machine and allow the student to chart his own course through material made available 1Alan B. Salisbury, "Computers and Education: Toward Agreement on Terminology," Educational Technology, Septem- ber, 1971' pp. 35-40. 2Alan B. Salisbury, "An Overview of CAI," Educational Technology, October, 1971, p. 48. 3Ibid., p. 48. t< W} plays 2 options values informa the stu Th1 a mathe: algoritl the com; compute] CA] Gustave t0 teac} 1959, Dc of 1111, designed that CA] tiVely, and more traditic to him by the computer.4 When the simulation mode is used, the computer dis- plays an experiment of some real world situation with Options for varying parameters. The student then specifies values of the parameters, and the computer processes the information and presents the results of the simulation to the student. The problem-solving mode is used when the student has a mathematical problem to solve. The student writes an algorithm to solve the problem, stores the algorithm in the computer, and gives values of the variables. The computer then solves the problem and prints the solution. CAI began in 1958 with the pioneering experiments of Gustave J. Rath and Nancy Anderson in which they attempted to teach binary arithmetic on an IBM 650 computer. In 1959, Donald L. Bitzer and his colleagues at the University of Illinois began to develop PLATO, a CAI system especially designed to meet the needs of instruction. It was thought that CAI might offer instruction more cheaply, more effec- tively, more patiently, and do so in a less regimented and more individualized way than instruction presented by traditional approaches.5 4J. A. Howard, P. F. Ordung, and R. C. Wood, "On-line Computer Systems for Engineering Education—-State of the Art,” IEEE Transactions on Education, V01. E-l4, No. 4, November,I§7l, p. 210. 5Lawrence P. Grayson, "CAI: The Fifteen Million Dollar Experiment," Proceedings--Third Annual Frgntiers in Edu- cation Conference, IEEE Cat. No. 73 CHO 720-3E, p. 357. mpioamnm 0 '9 H“ The faj tance c process Ac rocess Diagnos Cause a minatio manner Systema. Sy; by tech] 4 Enthusiasm for CAI quickly developed and re- mained high throughout the 1960's. Hundreds of programs were written by hundreds of authors at numerous centers in dozens of computer languages on a large variety of computers. Programs were written to teach a variety of subjects such as physics, chemistry, mathematics, engineering, foreign languages, psychology. statistics, economics, and many others. Sixteen years after its inception, however, the promise of CAI as a powerful and acceptable educational method had not been fulfilled. The failure of these early efforts to recognize the impor- tance of linking computer applications to the learning process is well documented.7'8 According to Walker and Cotterman, "The learning process is, in large measure, the recognition of systems."9 Diagnosis, simulation, decision-making, problem-solving, cause and effect, if-then, input-output, and the deter- mination of how or why something would operate in a certain manner are all essential elements in the development of systematic problem—solving. Systematic problem-solving is frequently encountered by technical students and includes analysis, synthesis, 61bid., p. 357. 7E. J. Anastasio and D. L. Alderman, "Evaluation of the Educational Effectiveness of PLATO and TICCIT," Pro- ceedings--Third Annual Frontiers in Education ConferEHEe, IEEE Cat. No. 73 CHO 720-3E, p. 382. 8A. M. Mathis, T. Smith, and D. Hansen, "College Stu- dent's Attitudes Toward Computer-Assisted Instruction," JOurnal of Educational Psychology, Vol. 61, No. 1, February, I970, p.‘46. 9T. M. Walker and W. W. Cotterman, An Introduction to Computer Science and Algorithmic Processes, (Boston: AIlyn and Bacon, 1970), p. 451. and ca. one of and tel ics an< rigid i troubll I purpos free-b forces gram. forces to zer T are: In Steps libriU EXPQri lems 5 and calculations. Engineering mechanics is recognized as one of the core subject matter areas of most engineering and technology educational programs. The study of mechan- ics and particularly the application of problem-solving to rigid bodies at rest (statics) has proven to be extremely troublesome to students. In the study of statics, a rigid body is selected for purposes of analysis, and a diagram or sketch, called a free-body diagram, is then drawn. All of the external forces and moments acting on the body are shown on the dia- gram. Equilibrium exists when the sum of the external forces and the sum of the external moments are both equal to zero. The steps required to solve an equilibrium problem are: 1. read the problem 2. draw a free-body diagram 3. show all external forces and moments which act on the body 4. apply the principles of equilibrium 5. perform the necessary calculations Steps one through three comprise the analysis, and step four is the synthesis of the problem. From such a sketchy description of statics, it would seem that students would have little trouble solving equi- librium problems; however, students have traditionally experienced a great deal of difficulty working such prob- lems. By far the greatest problem has been the drawing of a corr and mo was dr from t equili studen soluti at a C matel; a stud know v most ( SYnth Wit: 6 a correct free-body diagram showing all the external forces and moments acting on the body. If the free-body diagram was drawn incorrectly or if the correct data was not used from the free-body diagram when applying the principles of equilibrium, no amount of perserverance on the part of the student during the calculations would yield the correct solution to the problem. According to Figure 1, there is only one way to arrive at a correct solution to a problem, but there are approxi- mately twenty ways to arrive at an incorrect solution. If a student arrives at an incorrect solution, how does he know where to start checking for his error? Although the most desirable approach is to first check the analysis of the problem, most students first check their calculations. In many problems, performing the calculations is the most time-consuming part of the solution to the problem. Check- ing the calculations for the problem is still more time- consuming and does not help the student to arrive at the correct answer if an error had been made in the analysis or synthesis of the problem. Students have become so oriented to finding the correct answer to a problem that they have lost sight of the importance of the analysis and synthesis portions of problem—solving. Having expertise in calculations is commendable, but it is almost useless without a thorough understanding of analysis and synthesis. The ability to perform calculations has always been an essential ingredient in problem-solving. However, when the calculations become lengthy and repetitive a great deal UUQHHOU CHOU Cfihm MUQQ ENwQ no» BUCK mm» mm; m0» [1 r _ _ _ F \ ZOHBDJOW mZOHBANqHDUwHANU WHWWIPFZCfiMU. mHMWKAWHd‘ZdN Ewanoum m 0» :ofiuoaom on» no uuono3on .H onsmflm OZ uoouuoocH mucouuou meowuoasoamuuoamo mucouuoo moamwocwumucwum mucouuou ouooumuoo mucouuou Emummwoummao no» no» mMAuooHHOU poomuomom oz Dzmwma OZ oz oz oz oz voouuoo . oaoo .cdum sumo poem no» no» no» no» men _ _ _ L _ _ _ _ ZOHBDAOm mZOHaflflbuqflu mHmWEBZMm . meMAectsz. and SCie of Purpo Wit COst-ben to d°cum \ 21S 22 A 24K neat. Ed . 10: 18 capabilities; poorly authored educational content; expensive and unreliable hardware; an educational bureaucracy resistant to the intrusion of computers in the classroom; the decentral- ized structure of the American educational system that leads to tens of thousands of school systems, each having to be 'individually sold on the idea.21 Alpert claims that the technology of the 1960's was not capable of making a significant and economically practical contribution to the nation's educational program. Present CAI systems entail total costs which range between two and eleven dollars per student-contact hour at a termi- nal.22'23 According to Zinn, the current state of instructional programming languages is characterized by proliferation and implicit assumptions. He presents a list of some forty-eight different CAI languages, and states that he expects less progress toward standards than in business and scientific programming, because of the great variety of purpose and process in instructional programming.24 With the increased interest in accountability and cost-benefit analysis, it has become increasingly important to document the benefits received when substantial amounts 21Stetten, p. 371. 22Alpert, p. 1586. 23Howard, p. 216. 24Karl L. Zinn, ”Instructional Programming Languages," Educational Technology, March, 1970, pp. 43-46. of money a have point Despi Assis lack upon deli‘ case: thei: Othe: CAI's ran use of co student (3 hours wri Manj be done Of instr Nev fi] “101 re< ac? \ 2E 26 What NC Edljcat‘ 2‘ 2 by 0 an A) 19 of money are spent on innovations. Anastasio and Alderman have pointed out: Despite substantial prior research in Computer- Assisted instruction, instructional systems typically lack detailed information regarding their impact upon the educational community. The development of delivery systems and course materials has, in most cases, proceeded without adequats attention to their educational effectiveness. 5 Other factors include: an uncertainty concerning CAI's range of application;26 many people thought that the use of computers in instruction dehumanized the teacher— student dialogue;27 authors are required to spend too many hours writing CAI programs,23p29 Recommendations to Improve CAI Many suggestions have been made regarding what must be done if CAI is to become an economically feasible means of instruction. According to Zinn: New techniques for preparation of curriculum files must be developed, techniques which are more powerful in the sense of fewer author hours required to write and revise materials which achieve the subject matter objectives intended. 25Anastasio, p. 382. 26Erik D. McWilliams, "The $15M CAI Experiment-- What NSF Expects," Proceedin s--Third Annual Frontiers in Education Conference, IEEE Cat. No. 73 CH5 726-3E, p. 357. 27Howard, p. 210. 28Herbert S. Diamond, "The Writing of a CAI Program by an Author New to Computers," Educational Technology, October, 1971, p. 42. 29Zinn, p. 44. Autl: div: in e Diar for each a course language that auti used.32 when dev twenty s material Psycholc ists, a] Sold 111 zi more as that be ComPUte waYs. used Inc Bt to Pro- \ 3 20 Authors cannot often afford the luxury of in- gividually shaping or tailoring each ling of text n each frame for each kind of student. Diamond has suggested that the same format be used for each lesson when developing and writing programs for a course.31 Alpert suggests the use of an easily learned language such as Tutor used for the PLATO IV system, and that authors receive royalties when their programs are used.32 Stetten suggests a split of strategy and content when developing courseware. His group has identified twenty strategies or logics of instruction. The educational material (courseware) is developed by teams of instructional psychologists, subject matter specialists, media special- ists, and programmers. The entire package would then be sold like textbooks.33 Zinn further advises that projects use the computer more as a learning tool than a presentation device, and that benefits are apt to be considerably greater when the computer does things which could not be achieved in other ways. Problem-solving and games and simulation should be used more.34 Boblick suggests that computer simulations be used to provide learning experiences which might not be available 30Zinn, p. 44. 31Diamond, p. 42. 32A1pert, p. 1589. 33Stetten, p. 372. to stude cost or factors Bit educatic estimate capital dent hou of less the TICC The det pre wit; com: in . has fro: the: ser‘ Mit: hav. TIC1 filing and. 21 to students because of factors such as safety, equipment cost or availability, prohibitive set-up time, or other factors of cost or convenience.35 Bitzer emphasizes that the cost of computer-based education must become far lower than it has been, and estimates that when PLATO IV is fully implemented that capital and operating costs will be fifty cents per stu- dent hour at the terminal.36 Stetten estimates a cost of less than one dollar per student contact hour with the TICCIT system.37 NSF Experiment The National Science Foundation in an effort to determine the current problems and opportunities presented by CAI is supporting a major experiment within the limited, but specific confines of the community college setting, and to a lesser extent in elementary schools. The University of Illinois has received $5 million of NSF funds, and $5 million from other sources, to complete the development and then test PLATO, which, in its present design, will serve up to 4096 terminals simultaneously. The Mitre Corporation and Brigham Young University jointly have received $4 million from NSF to develop and test TICCIT, a CAI system that will serve up to 128 ter- minals. The total experiment, which will last four- and-one-half years and will be completed in 1976, 35John M. Boblick, "The Use of Computer Simulations in the Teaching of High School Physics," Science Education, V01. 54 I NO. 1 ' Jan-Mar I 1970 ' pp. 77-§Io 36Donald L. Bitzer, Bruce Arne Sherwood, and Paul Tenczar, ”PLATO: Everyone's Answer,” Proceeding§--Third Annual Frontiers in Education Conference, IEEE Cat. No. 75—WO - s, p. 366. 37Stetten, p. 371. wil und con The PLATO IV' widely 5 large sc The dis the qua dis con jec F01 Th hardwar 22 will be evaluated by the Educational Testing Service under a $1 million grant from NSF. . . . Its oug- come may very well determine the future of CAI. 8 The two systems differ significantly in many respects. PLATO IV is designed as a computing utility to serve 4096 widely scattered student terminals simultaneously from a large scientific computer system. The heart of the student terminal is the plasma display panel, a flat sheet of glass upon which the computer can light up or turn off any of a quarter-million dots (in a 512 by 512 grid) to display text, graphs, and line drawings. The computer can select color photographs to be pro- jected on the back of the transparent panel. For technical reasons, this display device rep- resents a major advance over previous technology, including the cathode-ray tube. . . . Authors write their own materials in the TUTOR language which is powerful yet easy to learn. . . . When fully implemented it is estimated that capital and operating costs will be $0.50 per student hour at a terminal.39 The TICCIT system is designed to serve a single institution by using relatively inexpensive minicomputer hardware. Major innovations include: the use of audio and color TV displays in the student terminals to provide voice-accompanied multicolored alphanumeric and graphic displays (200 by 256 grid), as well as full-color movies; the use of a pair of minicomputers to provide the neces- sary computer power in a self—contained system of 128 terminals; the capability to deliver CAI and other socially 336rayson, p. 357. 393itzer, p. 360. relevan anew a high-qu control hardwar‘ than on« The Educatic 23 relevant computer services via cable television to homes; a new authoring system styled to support the production of high-quality CAI; a new and innovative use of "learner control" in CAI; a projected commercial cost including hardware, equipment maintenance, and CAI programs of less than one dollar per student contact hour.40 The PLATO and TICCIT systems will be evaluated by the Educational Testing Service. The scope of these demonstrations will make possible the collection of detailed information which reflects not only the cost and technical sophistication, but also the effects on achieve- ment and educational acceptance. Thus the NSF CAI project extends beyond a developmental exercise to a study of instructional technology's impact upon the educational institution, upon students, teachers, and administrators. The educational component of the PLATO and TICCIT evaluations will focus upon the consumers of educational innovations in order to determine the practical benefits and problems accompanying computer-based education.4 4oStetten, p. 371. 41Anastasia, p. 382. CHAPTER III DESIGN OF THE STUDY The Design of the Study is presented in two sections. Section I describes the development of the computer related instructional model and the testing required to make the model operational. Section II describes the collection and analysis of student and instructor attitudes concerning the model. Section I: Development of the Model A systems analysis and design approach was used in the development of the model to narrow the scope of the project. (The computer related instructional model will be referred to hereafter as the information system, which is consistent with the terminology used in systems analysis and design literature.) The approach also provides the potential for developing similar models for other problem— solving courses. The systems analysis and design approach was a modifi— cation of that proposed by Walker and Cotterman42 and was divided into the following parts: 42Walker, pp. 451-477. 24 It i 501ving i evidem:e into anal Computatil mentation Whic}, are! An analyS ti“ Win intereste 25 1. Problem recognition 2. Feasibility study a. goals and objectives b. information system's blueprint c. major equipment decision d. implementation planning 3. Traditional information system redesign a. scope and objectives b. analysis c. specifications d. design e. implementation 4. Information system mechanization 5. Information system modification Problem Recognition It is generally recognized that analytical problem— solving is the basis for engineering analysis. There is evidence to indicate that students gain broader insights into analysis and synthesis when freed from detailed computational methods. There is also need for experi- mentation with computer related instructional models which are directly addressable to the learning process. An analysis and comparison of the method and the informa- tion will provide valuable input to engineering educators interested in applying innovative instructional techniques. The parts: 9 print; n1 Thu determi: an info check c statics were: 26 Feasibility_Study The feasibility study was composed of the following parts: goals and objectives; information system's blue- print; major equipment decision; and implementation planning. Goals and Objectives The basic goal of the feasibility study was to determine the feasibility of designing and implementing an information system which would make a comprehensive check of a student's analysis and synthesis when solving statics problems. The objectives which guided the study were: 1. The development of the information system would follow a systems analysis and design approach which would serve as a basis for the develop- ment of additional systems for other problem- solving courses. 2. The student would be given the Opportunity to address his attention solely to the analysis and synthesis portions of problem—solving by being relieved of the mechanics of calculations. 3. Selection of the computer and computer language would be based on maximum transportability of the system. 4. The system would be evaluated for proper Opera— tion and educational impact. The general c for the 2 steps we: along wit One the info: items we: courses \ ments on with any taught w: SYSteUl we become t1 Scientif: gnages tn System We be aVail. of COmpu‘ frOm Ver: Signed f< be avail: Many Sta1 becauSe c 27 Information System's Blueprint The blueprint of the information system documented the general design envisioned and provided a general outline for the systems analysis and design. General processing steps were determined, and data structures were specified along with an estimate of the volume of each. One of the primary purposes of the study was to make the information system as transportable as possible. Six items were considered. Since textbooks used in statics courses vary from campus to campus and even between depart- ments on the same campus, the system was designed for use with any standard statics textbook. Statics classes are taught with either the vector or scalar approach, so the system was designed for use with either. Fortran IV has become the most universally used computer language for scientific calculations, but a large number of CAI lan- guages have been develOped for use with a terminal. The system was designed using a computer language which would be available at many educational institutions. The sizes of computer installations in educational institutions vary from very small to extremely large. The system was de— signed for use on a relatively small computer so it would be available to a large number of educational institutions. Many statics courses are taught primarily in two dimensions because of the great amount of time required for calcula- tions when using three dimensions. Since the principles are the same for either case, the system was designed for use with system re so a new pemittin practical of perfor system wa statics p The to read a external rectangul designate the axes , RECESsaI-y finally t OD Comput Many of the Or the direc 28 use with either two- or three-dimensional problems. The system required no calculations on the part of the student, so a new dimension could be added to statics courses by permitting the student to analyze and synthesize many practical three-dimensional problems without the necessity of performing lengthy and repetitive calculations. The system was also designed for use with a wide variety of statics problems. The person using the information system was required to read a problem, draw a free-body diagram showing all external forces and moments acting on the body, show a rectangular coordinate system on the free-body diagram, designate the location of the origin and the direction of the axes, take the data from the free-body diagram necessary to arrive at a solution to the problem, and finally tabulate the data in suitable form to be punched on computer cards. Many textbooks have problems showing the location of the origin of the rectangular coordinate system and the directions of the axes, so the system was designed leaving these items to the discretion of the student. The information system was developed in three stages. The first stage was to develop an information system which would perform the necessary functions to solve the desired types of problems. The second stage consisted of modifying the system so that it could easily be debugged by students. The third stage consisted of modifying the system so it would mak ysis and through t Stage I The following couples, lations the resu Point ar forces : DOments equatig knOWn C. D.‘ lines the da and ar tYpe < 29 would make a comprehensive check of the student's anal- ysis and synthesis of the problem. Each stage was taken through the complete analysis and design process. Stage I The information system was required to perform the following functions: read data for forces, moments or couples, points, and lines; perform the necessary calcu- lations to transform the data into vector form and store the results; calculate the moment of a force about a point and store the result; calculate the sum of the forces and store the result; calculate the sum of the moments and store the result; fill a matrix with the equations of equilibrium; solve the matrix for the un- known quantities; present the solution. Data for forces, moments or couples, points, and lines could be given in various ways. Table 1 shows what the data was for, the types of data that were permitted, and an estimate of the elements of data required for each type of data. Stage II Several write statements were included to show the student just where in the program the execution was being performed in case the program was prematurely terminated. The data was read in and then written out for the student to check whether the data had been entered correctly. Data for Concentre Distribut Moments c Points Lines Stage II] Seve check the The Campus w System I COm Lan LOc SYStem 1 Con jLat 10c SYstem J Ccnw liar L0< 30 TABLE 1 ESTIMATES OF THE VOLUMES OF DATA REQUIRED Elements of data required Data for Types for each type Concentrated forces 6 7 Distributed forces 4 9 Moments or couples 3 3 Points 1 3 Lines 2 6 Stage III Several diagnostics were built into the system to check the student's analysis and synthesis of the problem. Major Equipment Decision The computer facilities available at the Fort Wayne Campus were very good and included the following: System I: Computer--IBM 360-22 Language--Basic Fortran IV Location--Fort Wayne, Indiana (In-house) System II: Computer--CDC 6500 Language--Fortran IV Location--Purdue University, West Lafayette, Indiana System III: Computer--CDC 6600 Language--Fortran IV Location--Indiana University, Bloomington, Indiana System IV Comp Lang Loce System V Com Lar LO< System ‘ Co 31 System IV: Computer--CDC 6500 (The PLATO system) Language--Tutor Location--University of Illinois, Champaign, Illinois System V: Computer--CDC 6500 Language-~PLANIT Location--Purdue University, West Lafayette, Indiana System VI: Computer--CDC 6600 Language--PLANIT Location--Indiana University, Bloomington, Indiana System VII Computer--CDC 6600 Language-~APL Location--Indiana University, Bloomington, Indiana The investigator has had experience writing programs in all of the computer languages of the seven systems except the tutor language for the PLATO system. Basic Fortran IV and Fortran IV have been used when writing many programs requiring scientific calculations, programs to process student records, and programs to plot output. CAI simulation programs concerning heat loss calculations, which required a mathematical model, were written in APL. Several CAI simulation programs, which did not require a mathematical model, concerning trouble-shooting of refrig— eration systems were written in PLANIT. In the programs, the stude finding t eration 5 union app trouble—s an exciti technicia The the advaz When deve universa. PEOple i sllbl’outi many Cal cards ca into thc COmDute Cards a A} but th. Main p r011tir jLEI-Ogre be en temi debut prob 32 the student was permitted to practice his reasoning in finding the cause of a given malfunction in a refrig- eration system. Instructors in the local steam fitters union apprentice program were very impressed with the trouble-shooting programs. They felt the programs added an exciting new dimension to the training of refrigeration technicians. The following are the investigators views concerning the advantages and disadvantages of the computer languages when developing an information system. Fortran IV is the universal language for scientific calculations, and most peOple in the scientific field are familiar with it. subroutines can be called from the main program to perform many calculations. When a program is being develooed, cards can be punched on a keypunch and the program fed into the computer. The program is debugged from the computer print-out, and corrections are made by repunching cards and feeding the program back through the computer. APL is a completely different language from Fortran IV, but the programs are similar to Fortran programs since a main program and subroutines can be written with the sub- routines called from the main program as needed. When the program is being developed, however, all instructions must be entered at a terminal, and the program debugged from the terminal. Although the APL program can be punched in, debugged, and tested for prOper operation immediately, a problem can exist if it is difficult to gain access to a terminal. FLA writing This can can be p computer debugged is a dis Tut very eaSj the PLATl an expen System 13, graphics The the deVe of this PLATO Sy termirlal and aCce in the e 33 PLANIT is another completely different language, and writing a program consists of writing a series of frames. This can be very awkward and time-consuming. The cards can be punched on a keypunch, the program fed into the computer without a terminal, and then the program can be debugged from the terminal or by repunching cards. This is a distinct advantage if terminal access is a problem. Tutor, used with the PLATO system, is claimed to be very easy to learn. However, the greatest advantage to the PLATO system is its graphics capabilities. It requires an expensive special terminal, however, and the information system being developed may not require the outstanding graphics capabilities. The use of PLATO, APL, or PLANIT posed a problem for the development and operation of the information system of this study. Only one terminal was available for the PLATO system, and it was used almost continuously. Two terminals were available for the APL and PLANIT systems, and access to them was limited. Also, the APL system was in the experimental stage, and it would have been impossible to use it for the development of an information system such as the one proposed in this study. The selection of the computer system to use for the information system was based on five items: capabilities; accessability; transportability; hard c0py; the cost to the department for the use of the system. The questions considers CapeI l . 2. Acce Tra: 34 considered under each item were: Capabilities: l. 2. Did the computer have the required core? Did the computer have the capability to perform all of the transformations of information required? Accessability at the Fort Wayne Campus: 1. Could the student gain access to the system without wasting a lot of time on such things as trying to gain access to a keypunch, waiting his turn for a terminal, and turn around time on the computer? Transportability: 1. Would the computer system be the size that most educational institutions would have available? 2. Would the computer language be fairly universal? 3. WOuld the computer language be such that most students and faculty members would be familiar with it? Hard Copy: 1. WOuld the output of the system be such that the student could take it with him for future refer— ence? 2. If the output were a hard c0py, would it be in a neat, systematic form? Cost 1. 2 . Tabl system on received info rmati for use 0 Ava implement, Lem 35 Cost to the department at the Fort Wayne Campus: 1. WOuld the department be charged anything for the use of the system? 2. What charges would be made to the department? Table 2 shows how the investigator rated each computer system on the various items. The IBM 360-22 computer received the highest rating, so it was decided that the information system would be written in Basic Fortran IV for use on the IBM 360-22 computer. Implementation Planning A variety of major activities were necessary for implementation of the system which are listed by stages. Stage I . detailed design of each subsystem l 2. algorithm development and testing 3. installation of the information system 4 . review and evaluation of the system 1. modification of algorithms and testing Stage III l. detailed design of subsystems 2. algorithm develOpment and testing 3. review and evaluation of the system 4. preparation of written procedures for faculty and students 36 qmfl “mom mm» mm» uoom pom comm oco HH> 8H24Am nwmm no» new moon com comm UGO H> 9H24Am Hflmm new we» Moon @000 comm ooo > Aoamqmcuouse noon >Hm> oz oz Moog >um> ucmaamoxm comm UQU >H >H cmuuuom coco mow no» @006 ucoaamoxm comm U00 HHH >H cmnuuom coco new mom coco ucoaaooxm comm coo HH >H cmuunom owmmm ucmaamoxm oz mm» nemaaooxm poom suo> «Nuomm zmH H MBHAHmdemommz¢ma omomdmu Mmoo MBHAHmHmmmooe mmHBHAHmmmdo moH0m "~ng @GHUDOHQQW G053 uwflUflHdnHmmom HHAN HO UHfl£O30Hh .m OHfiOflh Amco«umamu oumuzc mcowuooom xcmuH mczocxc: H+xcmuH coflusaom ucaum Amcoflumamm ouwu3c mcomumovm xcmuH assassin HSEMHH Amcowuoaou ouwuzv coauoaom macauopom xcmHH ucwum mczocxcc ~+xcouH Amcowumamm mamas. moofiuoocm xcmun masocxcc ~+xcoHH vummma O" 0V 2| 0A xcmuH eanMH wflmmmH mcowumoom ucmocomoo aduoocwa oz mummmH xcmmuxamua oceoaoonz mzomuz mummma ozmomq ncowumoom ucmocomoo hammocwq Y NHMMMH OA xcmuH 0v HummmH .. . 3:1... . 1.2%....» u .— -..... $54.... 62 1. No rows or columns in the matrix. 2. The zero matrix. 3. The number of equations is less than the rank of the matrix. 4. More equations than unknowns. 5. All sums are zero for the active forces and moments. Therefore, the body was in equilibrium without any reactive forces or moments. 6. Inconsistent set of equations. Figure 4 shows the operational flowchart for stage I of the design, and Table 10 lists the arrays used. Stage II All items noted in the specifications for Stage II to help make the system easily debugged by the student ‘were included with Stage III. Stage III Five additional subroutines were necessary to make a check of the students' analysis and synthesis. Subroutines FILKME, FILKMC, AND FILKMM were written so that answers to certain parts of the problem could be entered by the faculty members. Each subroutine contained the answers to five problems. For each problem, data was entered for chapter number, problem number, number of active forces, number of active moments, number of reactive forces, number of reactive moments, and a CALL ZFORC CALL ZPTLN CALL ZSMOM L ZARR [READ CHAPTER AND PROBLEM NUMBERS] CALL PTLINMALL COUPLFH [CALL FORCES CALL DLOAMALL COUPLE] 3 [CALL MOMPT ] L ICALL EQUA I CALL MFGRR [CALL DUMP IT] 1 [CALL FINISH] Figure 4. Operational Flowchart for Stage I of the Design 64 H HRH omumnmndaa Hoonom mom me viH HMOHIUwMOEdz xmuumz m h nxa ummmucwuomumeoz xwuume mo mcEsHOU AOOH m mxa Homoucauomumsdz xfiuume mo mzom zomH n bxa HomoucwnofiumEdz xwuume mo mcESHOO OzomH m mxH ummmucflIOfiuoeoz anume mo mzom zzomH n th ummmucwnownossz xwuuoe mo mCEdHOU 02 m mxH Hummucwuowumeoz xmuume mo m3om mz n hxa Hometowuoesz sumo mcwooom m mucoeoe v Nxm umooucHIOfiumEoz one mOUHOm mo mumnesz 4232 mm hxm nomoucwuomumsoz muo3mcc 2x m mxH owumeocmcmad mmemz zHummmznz w>wuflmomuma oz nosmwm + +m~o.H A h +m~w.~ o m NH oz Renown + +mbm.H c m ooo.m N N Ha oz Honmfim + +oms.H o m +m~o.~ N m ca oz Hannah + +m~a.~ o w +mhm.~ o e m oz Hocmwm + +omh.a o m +om~.~ o w m oz Hannah + +mbm.~ a m +m~w.m H w 5 oz Hocmwm + +mha.~ o m +~mv.~ o m m oz nonmflm + +vom.H o m +mnm.~ o m m oz Hmnmwm + +Hm~.~ H h +mwm.~ m n v oz Honnwm + +Hmm.a o m +HNm.N o h m oz Honnfim + +mnm.~ H h +mmm.~ w v N oz Hannah + +mmo.~ o m +~Nm.N o m a moocououmao Home smoE.:fl coo: «z «A com: «z «m huomoumu unmowmwcowm mocmsu amounumom unouloum Emmalamom QZ¢ Bmflalflmm .mBZmDDBm UZHMQmZHUZm MOE flfidd HEB ho mHmwndzd HEB m0 MMdZSDm 0H Nflmda 91 o>wummozuz o>wuwmomuma oz umanm - +mhh.~ m mH +qoa.~ o n «H oz umanm + +om~.~ N «H +mmm.~ H HH HH no» uoamHm . -vv¢.m 6H s +th.~ o h oH mo» umanm . ooo.m a HH +Hme.~ H «H a no» ouasvm-Hnu . -mmm.m HH 4 +mmH.~ o mH a oz umanm - +~mw.~ HH mH +man.~ e m n no» mumsvmano . -Nmm.m mH m +MH¢.~ H «N 6 mm» muunvmano - nom~.m SH 5 +mom.~ H m~ m no» umanm . +Hmu.~ m mH +mHm.~ H mm 4 no» mumavm-Hao . uqu.m «H NH +mmm.~ H AH m mm» mumsvmquu . -Hv~.m ow m +ovm.~ a pH N no» mumsvmano u -mNH.m 6H m +mmm.~ H on H noocoummwwv amok some SH com: .2 «m :80: «z «m whomoumo unmonHcmHm «mango unoulumom umoulmum BmQBIBmOm 92¢ Emmalmmm .mfizmQDBm NUOAOZEUNB mom dadn was ho mHmHA42< mma ho flmdzzam mH mflmflfi 92 m>Hummoznz o>HuHmomumo oz noonz ozHuoooHocm .+oon.~ o p +m~o.~ o m «H oz noonz zooHoozooa +mom.~ H HH ooo.m ~ ~ HH oz nonon mzHuoooHocm +Hoh.~ o n +m~o.~ n m oH oz woanm mcHuoochcm +Hmv.~ H «H +mhm.~ o v m oz nocon zooHoozooa +moH.~ o «H +om~.~ o o a oz nonon maHuoocHoam +msh.~ o m +m~o.~ H o 5 oz nonon zooHoozooe +nHo.~ H ow +~mo.~ o m 6 oz uoonz ooHuoooHoom +oom.~ H m~ +omm.~ o a n no» noonz zooHoozooa +mHn.~ H HH +nom.~ m m o oz nosmHz ocHuoochcm +mmo.~ H AH +H~m.~ o o m oz nonon zooHoanooa +oom.~ 5 pH +omo.~ m o a oz nonon ooHuoooHoom +omm.~ H ou +-m.~ o m H nUUSUHUMMHo puma o>HUHnom cum: «2 on cum: «2 «m muomouou unmonHcmHm umoz amoHocnooa msHuoocHucm Emmalmmm .mazmabam HUOAOZmUNB 92¢ UZHKHHZHUZH ON mAmHummozuz o>HuHmomumo oz nozmHz ooHooooHoom +mnh.~ a mH +m~o.H H p «H oz nozmHz ocHuoooHoom +oo~.~ N oH +mam.H o a HH mo» nonon ooHuoooHocm -voo.m oH a +omp.H o m oH oz nozmHz ooHuoooHoom ooo.m a HH +m~H.~ o o a moz oozmHz moHuooonom -mmm.m «H o +omn.H o m o no» nozmHz ooHooooHoom +~mm.~ HH mH +m~m.~ H m a no» nozmHz ooHooooHoom -Nmm.m mH o +ohH.~ o m o no» uoonz ozHuoooHoom -om~.m SH 5 +oom.H o m m oz noanz ooHuoooHoom +Hmp.~ m mH +Hm~.~ H a o no» noanz ooHooocHoom -ooH.m HH NH +H~m.H o o n no» noanz ocHuoooHoom -Ho~.m om m +m>m.~ H A N mo» noanz ooHuoozHoom -mHH.m oH m +mmo.~ o o H moocouomuwc umoa 0>HuHmom com: .2 am com: «2 «m muomoumo uoooHoHoon oooz zooHosnoma mcHHomchcm somauamoo .mazmozam zooqozmoma oza oszmmzHozm HN mflmdfi mom dfidd MZE ho mHmNA¢24 NEE ho HMGZZDm CHAPTER V SUMMARY, CONCLUSIONS, RECOMMENDATIONS, AND IMPLICATIONS FOR FURTHER DEVELOPMENT AND RESEARCH Chapter V includes a summary of the results of the student pre-test and post-test, conclusions concerning the objectives of the study, recommendations for mod- ifications to improve the model, and implications for further development and research. Summary The primary purpose of the study was accomplished when the computer related instructional model was operational. The model was capable of making a comprehensive check of a student's analysis and synthesis of a broad range of statics problems on an individualized basis. A secondary purpose of the study was to evaluate the model for educational impact. Attitudinal information was gathered by administering a pre-test and a post-test to the engineering and technology student populations. The engineering students were positive toward the model in all categories of the pre-test except the one concerning keypunching. These students became more positive toward the model in all categories of the 94 95 post-test, but none of the changes were statistically significant. The technology students were positive toward the model in all categories of the pre-test. These students became less poSitive or negative in all categories on the post-test, and all changes were statistically signif- icant except those concerning previous experience being required on a keypunch, too much keypunching being required when using the model, and the value of the card for the keypunch drum. No significant differences were found between engineering and technology students on the pre—test except the category concerning the model interfering with the learning of statics. Significant differences were found between engineer- ing and technology students on the post-test in all categories except those concerning the model interfered with the learning of statics, previous experience with a computer was necessary, too much keypunching was re- quired when using the model, and the value of the card for the keypunch drum. Conclusions The development of the model, as outlined in the Design of the Study, follows a systems analysis and design approach which will serve as a valuable basis for the development of additional models for other problem- solving courses. 96 The model is written in Basic Fortran IV, requires less than 16K of core including systems overhead, can be used with any standard statics textbook, and requires no terminals or telephone lines. These features insure transportability to other computer systems. The model provides an opportunity for students to focus solely on the analysis and synthesis portions of problem-solving by being relieved of the mechanics of calculations. The output of the model is a hard copy in a neat, systematic form which checks the student's analysis and synthesis. Students may perform the analysis and synthesis of many three-dimensional problems which were heretofore impossible or very difficult due to lengthy calculations. The model visually brings out important points concerning Simultaneous equations with as many as six equations and six unknowns, and provides feedback to the student on an individualized basis. The model does not interfere with the learning of statics. Very little keypunching is required, so no previous experience with a keypunch is necessary to effectively use the model. However, some previous exposure to a computer is necessary or at least desirable. The main computer program is a pat program capable of working five problems with each computer run. It consists of call statements for canned subroutines, thus giving the students experience in calling such subrou- tines and keeping them involved with the computer. 97 Approximately 2.5 minutes are required on the IBM 360-22 computer for each computer run. Engineering students experienced little difficulty using the model, considered the diagnostics to be valuable, and found the program easy to debug. They also considered the model a good teaching tool, a valuable asset to learning, and an effective means of helping them to become more systematic in problem-solving. Technology students did not agree with the engineering students on any of these items. The design could have been strengthened through the employment of control and experimental groups. Recommendations The student user's manual should be reduced in length. Plow charts of subroutines and the relationships between subroutines and arrays should not be included. An inde- pendent study package should be designed to complement the manual which would include the use of a slide projector and tape recorder to provide an explanation of the model. It would be preferable to have the package consist of two or three short presentations to be used as the class progresses through the course. The package would offer the advantage of reducing the class time required to introduce the model, and the student could study the material at his convenience, repeating the material as many times as desired. 98 Instructors using the model should require the students to work the problems and hand them in to be checked. The technology class was selected on very short notice since it was the only one available for the study, and the technology faculty members were introduced to the model during the beginning of a new semester. It is recommended that when the model is used again, the faculty member or members have a much longer lead time to become familiar with and experiment with the model before attemp- ting to use it with a class. They should also be thoroughly familiar with the strategy for using the model. The pat main program should be made a subroutine and put on the disk, thus reducing the compiling time significantly. It is estimated that this could reduce the time required to work five problems from 2.5 minutes to a little less than a minute on the IBM 360-22 computer. 'Also, a few of the arrays could be reduced in size or eliminated, thus cutting down on the total core required. When an error in analysis or synthesis is detected in a problem, a diagnostic message is given and the program is terminated. The model will be modified so that when such an error occurs the particular problem is terminated, but the program will continue to the remaining problems included in the program. 99 The sheet for tabulating the data will be revised so that the actual width and columns correspond in length to those on a computer card. Then when a student punches a card, he can lay it down on the data sheet to make certain he punched data in the desired columns. Implications for Further Development and Research. The potential for a computer related instructional y model, such as the one in this study, is almost unlimited. Any subject area which involves lengthy and/or repetitive calculations is a possible candidate. Furthermore, many computer programs have been written over the years for use by students which relieve the student of the calcu- lations, but do not have the means for checking the student's analysis and synthesis with appropriate diagnostics. A thorough analysis of many of these programs would probably disclose that means could be incorporated to perform this added function, thus making them much more valuable as a teaching tool. Plans are already in progress to explore the possi— bility Of modifying the model so its usefulness will be greatly expanded. Problems with friction at impending motion will be considered first. Then, the possibility of including problems which require the free-body diagram to be broken into two or three additional free-bodies will be considered. The incorporation of these two items would increase the power of the model considerably. 100 The model was developed to be used as a supplement to a regular statics class. Possibilities are being explored to extend its use to independent study courses, freshman engineering design courses, courses for the professional development of graduate engineers, and engineering review or refresher courses intended to prepare graduate engineers for professional tests required for licensing. Other areas being considered for instructional models are dynamics, thermodynamics, strength of materials, engineering design, and cost-analysis for design-courses. An experiment should be conducted with the model when it has been further developed as noted in the recommendations above. A large population of students should be selected for the study, all of whom would be enrolled in the same statics class with one instructor. Control and experimental groups should be randomly selected for the experiment in which only the experimental group would use the model. An instrument should be developed to measure achievement in the cognative area of learning which would be administered to both groups at the end of the experiment. The test results should be used to determine significant differences in achievement of the two groups. A further study should include a cost-benefit analysis. A cost analysis of the use of the model should be compared 101 with increases in student achievement when using the model. The analysis could be used to determine if the added cost when using the model could be justified from the added benefits the students would realize. SELECTED BIBLIOGRAPHY Alpert, D., and Bitzer, D. L. "Advances in Computer-Based Education.” Science, (March, 1970), 360, 1586. Anastasio, E. J., and Alderman, D. L. “Evaluation of the Educational Effectiveness of PLATO and TICCIT.” Proceedings: Third Annual Frontiers in Education Conference. New York: IEEE Cat. No. 73 CHO 725-3E, (April,'I§73), 382. Bitzer, Donald L.; Sherwood, Bruce A.; and Tenczar, Paul. "PLATO: Everyone's Answer." Proceedings-~Third Annual Frontiers in Education EOnference. New York: _IE__EE Cat. No. 73"“c'ao""'7§6-"§z, (April, I973), 360-370. Boblick, John M. ”The Use of Computer Simulations in the Teaching of High School Physics.” §gience Education, VOl. 54, No. l, (Jan-Mar, 1970), 77-81. ‘ Brown, B. R. ”Experimentation with Computer-Assisted Instruction in Technical Education.” Semi-annual Pro ress Re rt, Project No.gg§C-5-85-U7Z. University Parfi, Pa.: TEe Pennsylvania State University, 1966. Bueschel, Richard T. "Time-Sharing--A Pragmatic Approach in the School." EducationalTechnologyp (March, 1970), 21-23. Diamond, Herbert S. ”The writing of a CAI Program by an Author New to Computers.” Educational Technology, (October, 1971), 42. Edwards, E. M. ”APL: A Natural Language for Engineering Education, Part II: The First Programming Language Suitable for Engineering Undergraduates." IEEE Trans— actions on Education, (November, 1971), 179. Goldstine, H. H. The Computer from Paggal to von Nuemann. Princeton: Princeton University Press, I Grayson, Lawrence P. ”CAI: The Fifteen Million Dollar Experiment." Proceedings--Third Annual Frontiers in Education Conference. New York: IEEE Cat. No. CH6 725-3E, TAprII, 1973), 357. 102 103 Howard, J. A.; Ordung, P. F.; and Wood, R. C. ”On-line Computer Systems for Engineering Education-~State of the Art.” IEEE Transactions on Education, VOl. E-l4, No. 4, (NOvember, 1971), 210,216. Mathis, A. M.; Smith, T.: and Hansen D. "College Students' Attitudes Toward Computer-Assisted Instruction.” Journal of Educational Psychology, VOl. 61, No. 1, TFebruary,1§75), 43-51. McWilliams, Erik D. ”The Fifteen Million CAI Experiment-- What NSF Expects." Proceedings--Third Annual Frontiers in Education Conference. New YOrk: IEEE Cat. No. Rudberg, D. A. "APL: A Natural Language for Engineering Education." IEEE Transactions on Education, (November, 1971), 183-184T' Salisbury, Alan B. ”An Overview of CAI." Educational Technology, (OCtOber, 1971), 48. Salisbury, Alan B. ”Computers and Education: Toward Agreement on Terminology." Educational Technology, Stetten, Kenneth J. "Toward a Market Success for CAI (An Overview of the TICCIT Program)." Proceedings-- Third Annual Frontiers in Education Conference. New YorE: IEEE Cat. No. 73 CHO 725-3E, (ApriI, 1973), 371-372. Walker, T. M., and Cotterman, W. W. An Introduction to Computer Sgience and Algorithmic Processes. ficston: Allyn andBacon,197 . Zinn, Karl L. "Instructional Programming Languages." Educational Technology, (March, 1970), 43-46. GENERAL REFERENCES Boblick, John M. “Computer-Based Simulation of an Acid-Base Titration." School Science and Mathematics, VOl. 71, (January, 1971), 49-54. Boocock, Sarane S. ”Using Simulation Games in College Courses." Simulation and Games, VOl. 1, No. 1, (March, 1976), 67—77. Burke, J. Bruce: O'Neill, Julia; and welsch, Kay. "A Humanized Model of a Computer Managed Instructional System.” Educational Technology, (November, 1972), 31-36 0 104 Deep, Donald. "The Computer Can Help Individualize Instruction." The Elementagy School Journal, V01 0 70 ' NO. 7 ' (April ' 1.9—7 ' - o Gaunt, Roger, N. "Benchmarks May be More Important Than Trademarks in Selecting Computers." Colle e and University Business, (June, 1971), 68—70. Gaunt, Roger N. ”Computer Choices: Ignore it, Hire it, Buy it, or Get Together to Use it.” College and University Business, (August, 1971), 2fil27, Gaunt, Roger N. "Everything You Always Wanted to Know About Computers, But Didn't Know Whom to Ask." College and University Business, (October, 1971). “-3 6 o Gaunt, Roger N. "Selection of Computer System Must Start with Understanding of Hardware-Software Capabilities." College and Universitnyusiness, (April, 1971), 74-78. Hansen, Duncan N. and Harvey, William L. "Impact of CAI on Classroom Teachers." Educational Technology, VOl. X, No. 2, (February, 1970), 46-48. Jacobson, Milton D. and MacDougall, Mary Ann. "Computer Management of Information and Structure in Computer- Supported Instructional Materials." Educational Technology, VOl. X, No. 3, (March, 1970), 39-427 Jamison, D. and Suppes P. ”Estimated Costs of Computer ASsisted Instruction for Compensatory Education in Urban Areas." Educational Technology, (September, 1970), 49-57. Kerr, Eugene G.; Ting, T. C.; walden, William E. ”An Instructional System for Computer Assisted Instruction on a General Purpose Computer." Educational Technology, VOl. x, No. 3, (March, 1970), 28930. Kooi, Beverly Y., and Geddes, Cleone. "The Teacher's Role in Computer Assisted Instructional Management." Educational Technology, VOl. X, No. 2, (February, ”76), 42-45. Proceedings of_a Conferencg_on_Computers in the Under- fiéraHUate Curricula. The University of Iowa, Iowa City, Iowa, (June, 1970). Swartz, G. Boyd. "Using a Small Computer in the College Mathematics Curriculum." Educational Technology, VOl. x, No. 3, (March, 1970),31-32. APPENDIX A STUDENT AND FACULTY INSTRUMENTS 9. 10. ll. 12. 13. 14. 15. 16. 17. 105 POST-TEST While using the COMPUTER PROGRAM I felt challenged to do my best work. I was concerned that I might not be understanding the material in the statics course because of the COMPUTER PROGRAM. I was not concerned about missing a problem because I knew I would receive diagnostics describing errors in my analysis. I tried to get the COMPUTER PROGRAM run rather than trying to learn statics. I knew whether my analysis was correct or not before I was told. I guessed at the method of analysis when using the COMPUTER PROGRAM. As a result of having used the COMPUTER PROGRAM, I am interested in trying to find out more about statics. I was more involved in understanding the COMPUTER PROGRAM than in understanding statics. The COMPUTER PROGRAM made the learning too mechanical. I felt as if I had a private tutor while using the COMPUTER PROGRAM. I was aware of efforts to suit the diagnostics specifically to me. I found it difficult to concentrate on statics because of the COMPUTER PROGRAM. Diagnostics were given in the COMPUTER PROGRAM which I felt were not relevant to the material. The COMPUTER PROGRAM is an inefficient use of the student ' s time . While using the COMPUTER PROGRAM I had a great deal of trouble keypunching. The COMPUTER PROGRAM made it possible for me to learn quickly. I felt frustrated while using the COMPUTER PROGRAM. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 300 31. 32. 33. 34. 106 The COMPUTER PROGRAM approach was inflexible. Even otherwise interesting material would be boring when using a COMPUTER PROGRAM. In view of the effort I put into it, I am satisfied with what I learned while using the COMPUTER PROGRAM. In view of the amount I learned, I feel that the use of the COMPUTER PROGRAM is superior to traditional instruction. With a course such as I am taking while using the COMPUTER PROGRAM, I would prefer the COMPUTER PROGRAM to traditional instruction. I am not in favor of the COMPUTER PROGRAM because it is just another step towards depersonalized instruc- tion. Previous keypunching experience is necessary in order to perform easily while using the COMPUTER PROGRAM. The COMPUTER PROGRAM was boring. The use of the COMPUTER PROGRAM made me more syste- matic in problem-solving. Previous experience with a computer is necessary if a student is to benefit from the COMPUTER PROGRAM. I had a great deal of trouble finding my programming errors while using the COMPUTER PROGRAM. I found the diagnostics given in the COMPUTER PROGRAM to be very poor. I found it very confusing shuffling call cards for subroutines in the main program. I found there was too much keypunching required while using the COMPUTER PROGRAM. I feel the punched card for the keypunch drum was very helpful. I found it difficult to organize the data when pre- paring to punch data cards. I found it difficult to understand how to use the COMPUTER PROGRAM. 107 PRE-TEST 1. While using the COMPUTER PROGRAM I will feel chal- lenged to do my best work. 2. I am concerned that I might not be understanding the material in the statics course because of the COMPUTER PROGRAM. 3. I am not concerned about missing a problem because I know I will receive diagnostics describing errors in my analysis. 4. I feel I will try to get the COMPUTER PROGRAM run rather than try to learn statics. 5. I will know whether my analysis is correct or not before I am told. 6. I will guess at the method of analysis when using the COMPUTER PROGRAM. 7. As a result of having used the COMPUTER PROGRAM, I will be interested in trying to find out more about statics. 8. I will be more involved in understanding the COMPUTER PROGRAM than in understanding statics. 9. The COMPUTER PROGRAM will make the learning too mech- anical. 10. I will feel as if I have a private tutor while using the COMPUTER PROGRAM. 11. I will be aware of efforts to suit the diagnostics specifically to me. 12. I will find it difficult to concentrate on statics because of the COMPUTER PROGRAM. 13. Diagnostics will be given in the COMPUTER PROGRAM which I will feel are not relevant to the material. 14. The COMPUTER PROGRAM will be an inefficient use of the student's time. 15. While using the COMPUTER PROGRAM I will have a great deal of trouble keypunching. 16. The COMPUTER PROGRAM will make it possible for me to learn quickly. 17. I will feel frustrated while using the COMPUTER PROGRAM. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 108 The COMPUTER PROGRAM approach will be inflexible. Even otherwise interesting material will be boring when using a COMPUTER PROGRAM. In view of the effort I will put into it, I will be satisfied with what I will learn while using the COMPUTER PROGRAM. In view of the amount I will learn, I feel that the use of the COMPUTER PROGRAM is superior to traditional instruction. With a course such as I am taking while using the COMPUTER PROGRAM, I would prefer the COMPUTER PROGRAM to traditional instruction. I am not in favor of the COMPUTER PROGRAM because it is just another step towards depersonalized instruc- tion. Previous keypunching experience is necessary in order to perform easily while using the COMPUTER PROGRAM. The COMPUTER PROGRAM will be boring. The use of the COMPUTER PROGRAM will make me more systematic in problem-solving. Previous experience with a computer is necessary if a student is to benefit from the COMPUTER PROGRAM. I will have a great deal of trouble finding my pro- gramming errors while using the COMPUTER PROGRAM. I will find the diagnostics given in the COMPUTER PROGRAM to be very poor. I will find it very confusing shuffling call cards for subroutines in the main program. I will find there will be too much keypunching required while using the COMPUTER PROGRAM. I feel the punched card for the keypunch drum will be very helpful. I will find it difficult to organize the data when preparing to punch data cards. I will find it difficult to understand how to use the COMPUTER PROGRAM. 109 FACULTY QUESTIONNAIRE l. The COMPUTER PROGRAM interferred with my teaching of statics. 2. The COMPUTER PROGRAM helped the students to become more systematic in setting-up problems. 3. I will definitely consider using the COMPUTER PROGRAM when I next teach statics. 4. It is good experience for the students to be exposed to the use of a computer with a COMPUTER PROGRAM such as this one. 5. It is too difficult for the students to understand the COMPUTER PROGRAM and how to use it. 6. The COMPUTER PROGRAM is a valuable teaching technique. 7. The diagnostics in the COMPUTER PROGRAM were very good. 8. My students were able to analyze problems which were heretofore impossible or very difficult. 9. I noticed a great deal of resistance by the students concerning the use of the COMPUTER PROGRAM. 10. I feel that the COMPUTER PROGRAM helped the students to better understand statics. COMMENTS: APPENDIX B USER'S MANUAL USER’S Objectives . . . . . . . . . Assumptions .'. . . . . . . Equations of Equilibrium . . Data . . . . . . . . . . . . Tabulation of Data . . . . . How to Enter Data . . . . . Flowcharts of Subroutines . Main Program . . . . . . . . beekoooo on 000’... 110 MANUAL Page . . . . . . . . . . . . . lll . . . . . . . . . . . . . 111 . . . . . . . . . . . . . 112 . . . . . . . . . . . . . 113 . . . . . . . . . . . . . 113 . . . . . . . . . . . . . 129 . . . . . . . . . . . . . 131 . . . . . . . . . . . . . 145 . . . . . . . . . . . . . 145 Computer Print-out of Main Program . . . . . . . . . . . 147 Computer Print-out of the Deck . . . . . . . . . . . . . 148 10. 11. 12. 111 OBJECTIVES Check a student's analysis and synthesis of a problem with suitable diagnostic messages. Require no calculations on the part of the student. Be sure the computer does not interfere with the learning of statics. Afford the students an opportunity to analyze and synthesize many practical problems which heretofore were impossible or very difficult due to lengthy calculations. Give instant feedback to the student on an indivi- dualized basis. Require a very simple main program consisting of call statements for canned subroutines. Make the program so that it can easily be debugged by the student. Require very little keypunching on the part of the student. Be usable by a person with little or no previous experience with a computer or keypunch. Give the student experience in calling canned subroutines. Visually bring out points concerning simultaneous equations with several variables which are ordinarily only discussed in class or in a textbook. Keep students involved with a computer. ASSUMPTIONS The analysis and synthesis are the most important parts of problem-solving. Many of the errors in problem-solving occur during the analysis of the problem. Given enough time, most students can make accurate calculations. 112 EQUATIONS OF EQUILIBRIUM Vector form: 2 i=5 2 i=5 Scalar form: 2 ano 2 Mx=0 .2 Fy=° Z N550 z rz-o .2 Mzao For this program, the student must distinguish between active and reactive forces and moments. Vector form: zFactive+z reactivego E Mact ive+ ZMreactiveflo Scalar form: ZFxA +ZFXR =0 zMxA +§MXR =0 ZFYA +2FYR =0 EMYA +§MyR no ZFZ +2F2R =0 2M2 +242 =0 A A R 113 DATA 1. Forces: A. Concentrated B. Distributed II. Points: III. Lines: IV. MOments: TABULATION OF DATA I. Forces: A. concentrated Name NO. F1 F2 F3 F4 F5 F6 F7 1 F l m n x y z 2 F 81 Y1 21 x2 Y2 22 3 F x8 y8 28 x y z 4 Ff Fx FY Fz x y z 5 F 1 x y 6 Fx FY Fz x y 2 A4 I4 F8.0 F8.0 F8.0 F8.0 F8.0 F8.0 F8.0 1-4 5-8 9-16 17-24 25—32 33-40 41.48 49-56 57-64 You have a free choice of the coordinate axes. (Except for NO 5.) 2 Y Y x x 114 2 No. 1 [Fl Given: Direction c‘giness: l,m,n 1H4x,y,z) Magnitude: Fl Point: x,y,z Y x Example: Name=FlO IFI =100.lb x33. 1 =-o.716f y=0. m = 0.275 z=4. n = 0.642 DATA CARD: Name NO. F1 F2 F3 F4 F5 56 F7 [_Fio I lllOO. J-O.7lGJO.275 [.642 jra. I I 4. I NOTE: Fl-F7 must have a decimal point NO has no decimal point. If a value is zero, leave the columns blank such as y=0 (F6 above) No. 2 Given: y [E] Magnitude: IE] (XZ'YZ'ZZ’ Two points: From(x1,y1,zl) . to (x2,y2.22) x /4/ //<;4XI'Y1'Z1’ Example: Ngme= F4 x2=-5. [Fl a 500.11)f y2=6. Y1 = -1. 21 a 8. DATA CARD: Name NO . F1 F2 F3 F4 F5 F6 F7 [ F4 J zlsoo. [2. [-1. [8. [-5. I6. [ l 115 /‘Y37/ 521.2 E1222? 2 x32 Magnitude: IFI "T Slope: x8,y8,zs ' Point: x,y,z :11. X ‘leIz) Example: Ngme=Fll x=3. IFI =3oo.1bf y=-l. =2. z=4. ys =3. Name NO. F1 F3 F4 F5 F6 F7 [r11 [ 3[3oo. [2.F2 [3. [-4. [3. [-1. [4. [ NO. 4 [Fl Given: (x,y,z) Magnitude: IFI Y Point: x,y,z Parallel to F1: Fxl'Fyl'le X, " IFI=Fx1i+FY1j+lek Example : Name=F [Fl =250.1bf x=0 Fx1 =3. y=0 z-O Name No. F1 F2 F3 F4 F5 F6 F7 [ F [ 4[250. 3. [2. -1o. [ [ J I] NO. 5 * 116 (For two-dimensional only) Y _- IFI ”(XIY) O‘ 91116.4: Magnitude: lF' Point: x,y Smallest angle F makes with the x-axis: 6‘ (always positive) x Directions in the x and y directions: l,m Exampl : Name=F3 lF|=1000.lbf 0:. =30 x =2. 1 =‘1. y 3-40 m =-l. DATA CARD: Name NO. F1 F2 F3 F4 F5 F6 F7 [_F3 [ s[1ooo. 30. [-1. [-1. [2. [—4. [ NO. 5(Continued) Given: y [Fl Same as above W‘er) 0‘ x Example: Name=F7 lib-100.11:f ¢*' =35 x =4. 1 =-10 y =3. m =1. DATA CARD: Name NO. F1 F2 F3 F4 F5 F6 F7 LpF7 [ 5[1oo. [35. [-1. [1. [4. [3. [ 7[ 117 NO. 6 Given: z (x,y,z) FOrce F: Fx,Fy,Fz Point: x,y,z y x Fein+ij+sz Example: Name=FO x=2.5 Ex :3; 2231'“ F: =12. DATA CARD: Name no. F1 F2 F3 F4 F5 F6 F7 [F0 [ 6[ [3. [-7. [12. 2.5 [-l.8 [8. ] I. Forces: B. Distributed LIMITATION on coordinate axes: Y NO. 1 x For a given problem, the z axes must be the same for z distributed forces, forces, NO. 2 moments, points and lines. Y x w(x) is the distributed load: w(x)=F(x)=I=Ax2-l-Bx-:-C-+-Dx'S In the computer: Func(x)=A*(x-R)**2+B*(x-R)+C+D*(x-R)**.5 118 LIMITATION on the origin of the axes: The origin must be at the left end of the member. LIMITATION on the coefficients of func(x). (A,B,C,D) All coefficients must be zero except one for each distributed load. LIMITATION on R, R1, and R2 shown below: R, R1, and R2 must always be positive or zero as measured from the origin to the right. Data Cards: Name NO. A B C D E G R R1 R2 [[11] I] l1l1] A4 I4 F7.0 F7.0 F7.0 F7.0 F7.0 F7.0 F7.0 F7.0 F7.0 1-4 5-8 9-16 17-24 25-32 33-40 41-48 49-56 57-64 65-72 73-80 NOte: R is the distance the curve has been translated to the right from standard position. 119 EXAMPLES OF DLOAD CASE I: A > 0 y Given: R1 2 R __— Name=Dl R2 > R NO=1 =2 . B=O C=0 D=0 C=0 Z x G=O 31 R=10. / R2 R1311 o R2=15. z / DATA CARD: Name NO. A B C D E G R R1 R2 E1 1 2. [ [10 [11. [15.1 CASE II: A) 0 2 Given: Rl < R Name=D2 NO=2 R2 S R A=1 . B=0 C=0 ‘ D=O R1 y E-O R2 GBO R=8 . / R R134 0 DATA CARD: Name NO. A B C D E G R R1 R2 £32 I 2[1. [ j [8. NOTE: A CANNOT BE LESS THAN ZERO. [4. l6. j 120 CASE III : B > 0 2 R1 2 R z=B (y-R) Given: R2 > R 1 Name=D3 \\\ NO=2 A80 \ B-.. _ _ C=0 R y D=O ./ R1 G=0 R=4 . / R2 41-6. xr R2=10 . B <0 y Given : R1 < R y=B (x-R) Name=D4 NO=1 R2 S R A-O 38.1 o C=O x D=0 R1 E=0 R2 G=0 R=12 . z / R R1=3 . I R2=10 a Name NO . A B C D E G R R1 R2 121 CASE V: C > 0 y Given : R1 _>_ R y-C Name=D5 R2)>R NO=l 7 A80 B=O C=25 . R x D-O E=0 “1 G=0 R=10 . 32 81-10 . z r R2-20. DATA CARD: Name NO . A B R1 R2 [05 I 1I I [25. L I I I10.J10. I20. I NOTE: C CANNOT BE LESS THAN ZERO. CASE VI: 2 .5 Given: D>O z=D (y-R) . -\ Name-D6 R1 2R NO=2 - A=O R2 >R BIO C=0 Raf/ Y 333' G=0 / R1 R=10 . l/r R2 Rl=11. , R2=l7. x DATA CARD: Name NO . A B C D E G R R1 R2 I06 I 2] I i J2. L I [10. I11. [17. I CASE VII: D < 0 z z=D (y-R) ' 5 Given: m < R ' N D7 7’ ame= R2 R N082 s y/ M B80 C80 D3- 2 o /_&_/ y E=0 ‘ G=O R2 R=10 . R / Rl=5 . x R2=8 . r DATA CARD: 123 IV. Moments: (Moments or couples) Name NO. F1 F2 F3 F4 F5 F6 F7 1 m 1 m n 2 CK c2 c2 A4 I4 F8.0 F8.0 F8.0 F8.0 F8.0 F8.0 F8.0 1-4 5-8 9-16 17-24 25-32 33-40 41-48 49-56 57-64 NO. 1 z ’/////:r M 9.1122: Magnitude: [CI ///I y Direction cosines: l,m,n Example: x Name=C3 [Ola-300. 1=-.716 m=.275 n=.642 DATA CARD: Name _NO._ F1 F2 F3 F4 F5 F6 _ F7 IC3 I 1I300. I-.716I.275 [.642 I I I I NO. 2‘ z I I IE Given: 5: Cx,Cy,Cz Y x Example: Name=Cl Cx=3. Cy=4. Cz=-5. DATA CARD: Name NO. 2 F3 F4 F5 F6 F7 Ic1 I 2I £113.]? 124 II. Points: III. Lines: Name NO. F1 F2 F3 F4 F5 F6 F7 1 l m n x y z 2 x1 Y1 21 x2 Y2 22 3 x y 2 A4 I4 F8.0 F8.0 F8.0 F8.0 F8.0 F8.0 F8.0 1-4 5-8 9-16 l7-24 25-32 33-40 41-48 49-56 57-64 Maine) 2 (errz) Given: Direction cosines: l,m,n Point: x,y,z ‘\ y x Example: Name=Ll x=2. NO =1 y=4. 1 --.716 z=-3. m =.27S n =.642 DATA CARD: Name NO. F1 F2 F3 F4 F5 F6 F7 I-.716I.27S ILl I» III 125 NO. 2 (Line) (x2,y2,22) Given: 2 Two points: From(x1,y1,zl) Y Example: Name=L4 xZI-S. NO =2 X1 =5. y2=30 Y1 =30 22--lo 21 =0. DATA CARD: Name Ro._ F1 i F2 i F3 i F4 F5 F6 _ F7» IL4 I 2| [5. I3. I I-5. I3. I-1.I NO. 3 (Point) rh‘rer) Given: Y /// Point: x,y,z Example: Name=Pl NO =3 x =5. y =2. 2 =_3 0 DATA CARD: Name NO. F1 F2 F3. F4 F5 F6 F7 IPl I 3I I I I I5. [2. I-3. I 126 The preceding material has been condensed for your convenience. All of the data has been placed on one sheet. You will be given a loose sheet of this form for use when solving problems. Another special form has been developed to assist you in tabulating the data. Loose copies of this sheet will be provided to you as you need them. Copies of the forms follow. You will be provided with a special card for the keypunch drum which will cause the keypunch to tabulate to the desired column when you strike the skip key. 127 SUMMARY OF DATA FORCES: CONCENTRATED NAME NO F1 F2 F3 F4 F5 F6 F7 1 F 1 m n x y z 2 F x1 Y1 21 x2 y2 z; 3 F x8 y8 28 x y z 4 F Fx F_ FZ x y z 5 F l m x y 6 Fx Fy z x y z PTLINE: NAME NO F1 F2 F3 F4 F5 F6 F7 1 1 m. n x z 2 x1 y1 21 x2 y2 z; 3 x y z COUPLE: NAME NO F1 F2 F3 F4 F5 F6 F7 1 C 1 m n 2 Cx Czfi Cz DLOAD: NAME NOI A B kc c (E G R _R1 R2 I 11 l l I l FOR DLOAD ONLY: . _ y NO=1 N082 128 TABULATION OF DATA NAME NO F1 F2 F3 F4 F5 F6 F7 A B C D E G R R1 R2 1 45 89 1617 2425 3233 4041 4849 5657 6465 7273 80 DECIMAL POINT a-I 129 HOW TO ENTER DATA I. Forces: A. COncentrated: Data is read into the computer by calling SUBROUTINE FORCES The subroutine continues reading forces (one force/card) until it reaches a BLANK CARD. LIMITATION: Only SIX FORCES may be read at a time. This includes both concentrated and distributed forces. B. Distributed: Data is read into the computer by calling SUBROUTINE DLOAD The subroutine continues reading forces (one force/card) until it reaches a BLANK CARD. II. Points: III. Lines: Point and line data is read into the computer by calling SUBROUTINE PTLINE The subroutine continues reading data (one point or line/card) until it reaches a BLANK CARD. LIMITATION: Only five points and/or lines can be read at a time. 130 IV. Moments:v Data is read into the computer by calling SUBROUTINE COUPLE The subroutine continues reading data (one moment/ card) until it reaches a BLANK CARD. LIMITATION: Only six moments can be read at at time. The following are flow diagrams of the subroutines and diagrams showing the relationships between subroutines and arrays . SUBROUTINE ZFORC BLANKS OUT ARRAYSA ANAM AND FNAM I ZEROES OUT ARRAYI FORC FORCES (NF) EQUAL TO ZERO SUBROUTINE ZPTLN BLANKS OUT ARRAY *PTNAM ISETS COUNTER FORI ‘ ZEROES *OUT TREI _ARRAY PTLN SETS COUNTER FOR POINTS AND LINES (NP) EQUAL T0 ZERO 131 SUBROUTINE ZSMOM BLANKS OUT MY SMNAM ZEROES OUT _ARRAY SMOM' MOMENTS (NM) EQUAL TO ZERO SUBROUTINE ZARR [SETS COUNTER FORI EROES OUT THE ARRAY z [ARR, SF, AND SM 132 SUBROUTINE FORCES [NRITES: SUBROUTINE FORCESJI [wRITEs: p—————DJREADS A CARD] YES N080 ? DATA CARDS FOR FORCES] NO WRITES OUT DATA] JUST READ YES o-[WRITES THE ARRAY FORC] WRITES OUT ARRAys/ FNAM AND FORC ‘No ICALCULATES Fx'FY'Fz I T PILLS ARRAYS FNAM AND FORC J ujfwnITEs: THE MAX. NO. FORCES IS 6 i Gm. Em) OF/' 133 ARRAYS USED WITH SUBROUTINE FORCES —.I H.- A, .[___. w ’ I I 1 I 1 I [CalculateszI Fx I Fy I Fz I I I 134 SUBROUTINE DLOAD I WRITES: THE ARRAY FORC FOR DLOAD ONLY WRITES OUT ARRAYS FNAM AND FORC IWRITES: DATAI CARDS FOR DLOAD] I NN=N+1 8 [WRITESz SUBROUTINE DLOAD] WRITES: THE MAX. NO. OF FORCES IS 6 -H>4READS A DATA CARD] YES (CALL EXIT) NO=0 7 , GIVES DIAGNOSTIC NO ERRORS WRITES OUT DATA FROM I CARD JUST READ GALL EXI’D YES A,B,C,D,E,G, R,R1,R2,0K? NO NOTE: Subroutine DLOAD uses the same arrays as subroutine FORCES. CALCULATES EQUIVALENT FORCE AND ITS LOCATIONV FILLS ARRAYS I FNAM AND FORC 135 SUBROUTINE PTLINE IWRITES: SUBROUTINE PTLINE] I [NRITES: DATA FOR PTLINE] I A—JWRITES THE ARRAY PTLN] F————d~{READS A CARD] YES WRITES OUT ARRAYS PTNAM AND PTLN N030? NO [WRITES OUT DATA JUST REAQ] [NP=NP+1] I WRITES: THE MAx. NO. OFJ] POINTS AND LINES IS 5 I YES (CEALL EXIT:> NP‘)>5? NO CALCULATES VALUES TOI ~FILL ARRAY PTLN I FILLS ARRAYS PTNAM AND PTLN 136 ARRAYS USED WITH SUBROUTINE PTLINE PTNAM PTLN Name 1 m n x y z I I I l I I I Name I Calculates: l m n I I I ‘ Type NO F1 F2 F3 F4 F5 F6 F7 I , L _DATA CARD 1 1 1 l 137 SUBROUTINE COUPLE 8 [WRITES: SUBROUTINE COUPLE] [WRITESz DATA ICARDS FOR COUPLE] I ’fizeads a card J . ' IWRITESzrloigRgguglfigny YES WRITES ARRAYS SMNAM AND SMOM no @ [WRITES OUT DATA JUST READ] NO=O? WRITES: THE MAXIMUM NO. OF MOMENTS IS 6 I (CALL EXITZ) CALCULATES VALUES TO FILLI ARRAY SMOM IFILLS ARRAYS SMNAM AND SMOM] J 138 ARRAYS USED WITH SUBROUTINE COUPLE SMNAM SMOM ‘ __1 Name Mx 5% Mz LCalculates: Mx My MzJ . *‘ T f * Name NO F1 F2 . F3 - F4 F5 F6 DATA CARD SUBROUTINE SUMFOR 139 SUBROUTINE SUMMON [NRITEs: SUBROUTINE SUMFORJ' [[NRITES: SUBROUTINE SUMMONI SUMs THE FORCES IN FORC— AND FILLS ARRAY SF [NUML(1,I)=NO. OF FORCE§J [TELLS THE NO. OF FORCES] GIVES THE SUM OF THE FORCES: Fx 'FL ,Fz SUMS THE MOMENTS IN SMOM— AND FILLs ARRAY SM LNUML(112)=NO. OF MOMENTS] [TELLS THE NO. OF MOMENTg] MOMENTS: M xz,M'.y,M / GIVES THE SUM OF THE SUBROUTINE MOMPT NAME OF FORCE K GOES] _IN SMNAM _CALCULATES MOMENT OF ] FORCE K ABOUT POINT J { [FUTS MOMENT IN ARRAY SMOM] WRITES THE FORCE NAME YES AND MX'MY'MZ FOR EACH FORCE 140 ARRAYS USED WITH SUMFOR 30:18 SUMS , SUNS SM 1’! ARRAYS USED WITH SUBROUTINE SUMMOM SMOM M . M H M _ x y z SMF [ [j 141 SUBROUTINE SOLVE SUBROUTINE MFGRR [WRITES: SUBROUTINE SOLVEj' [SOLVES THE MATRIX ARR] e 1 RETURNS THE SOLUTION l l/TELLS NO. OF UNKNOWN FORCES AND MOMENTS TO THE MAIN PROGRAM IN * ARRAY ARR ‘NUML(2,1)=NO. OF UNKNOWN FORCES ‘ NUML(2 , 2)-NO. OF UNKNOWN MOMENTsJ [FILLS ARRAYS ANAM AND ARR] WRITES ARRAYS ANAM, ARR, AND SUMS WHICH GIVE THE EQUATIONS OF EQUILIBRIUM IN MATRIX FORM [zEROES OUT ARRAYS SF AND SMJ SUBROUTINE EQUA SUBROUTINE TYFSYS. WRITES: YOUR EQUATIONS [ WRITES: SUBROUTINE TYFSYS 1 OF EQUILIBRIUM ' WRITES THE TYPE OF FORCE WRITES OUT THE EQUATIONS SYSTEM AND THE EQUATIONS OF EQUILIBRIUM OF EQUILIBRIUM.YOU USED. WRITES IF YOU ARE CORRECT OR NOT. IF NOT CORRECT, IT TELLS WHAT IT SHOULD HAVE BEEN. ARRAYS USED WITH SUBROUTINE SOLVE 142 FNAM FORC SMNAM SMOM IEEEEI Fx FY Fz x y 2 'FN;;;1 Mk My Mz I F1 Fxl Fy1 F21 x y 2 M1 M1x Mly Mlz F2 sz Fyz F22 x y 2 M2 M2x M2? M22 F3 Fx3 Fy3 F23 x y 2 F1 Mle MFlY MFlz ‘ F2 Msz MFZY MF2z “___-A _ __EE_I MF3x MF3y MF3z SUMS |SFK|SFY|SFZISMXISMYlst] SF [SF(1)[SF(2)[SF(3)I SM [SM(1)[SM(2)ISM(3)] SF ‘ ’ * ‘ ‘ and ANAM [ F1 [ F2 1 F3 [ M1] M2] vS%L_. 82!? ’ ARR Fx1 F22 Fx3 SF(1) 1SFx=o Fy1 Fy2 Fy3 SF(2) SFY-o F21 F22 F23 SF(3) SFZIO MFIx Msz MF3x MIx M2x SM(1) SMx=o MFIy MF2y MF3y MJ.y sz SM(2) SMY=0 MFIz MF2z MF3z M1z M2 SM(3) ‘SM2801 143 SUBROUTINE CHECK SUBROUTINE DUMPIT LWRITES: SUBROUTINE CHECK] [WRITES: SUBROUTINE DUMPIT] CHECKS THE NUMBER OF ACTIVE WRITES OUT INFORMATION AND REACTIVE FORCES AND THE INSTRUCTOR CAN USE MOMENTS YOU HAD. TO SEE WHAT THE OUTPUT OF ‘ § * SUBROUTINE MFGRR WAS. WRITES THE NUMBER OF ACTIVE AND REACTIVE FORCES AND MOMENTS YOU HAD. SUBROUTINES FILKMC, "FTEKMETFTLKMH"‘ NO comm YES FILLS ARRAYS To CHECK 1 ANALYSIS AND SYNTHESIS [WRITES THAT YOU ARE CORRECT] WRITES WHAT THEY SHOULD HAVE BEEN. WRITES: CHECK YOUR FORCES AND MOMENTS. 1 (CALL EXIT:) 144 SUBROUTINE FINISH _ NO . 2mm YES [WRITES: SUBROUTINE FINISH] I [WRITES TYPE OF ERROR _I [CHECKS SOLUTION FROM MFGRR] LINEARLY DEPENDENT EQUATIONS ? NO TELLS'WHAT THEY ARE YES Y (CALL 8x19 MORE UNKNOWNS A _THAN EQ. » [WRITES SOME VARIABLES LfiITEs THE RELATIONSHIPS I [WRITES THE SOLUTION Ja—J IN TERMS OF FREE VARIABLES 145 MAIN PROGRAM The main program consists of call statements for subroutines. All cards for the program will be given to you. A print-out of the main program, which is to be used for the solution of five problems, follows. DECK A deck of computer cards will be given to you which includes all cards needed to use the computer program except the calls for subroutines and the data cards. The cards in the deck will have nothing typed on them except a number. They will be numbered sequentially from 2-84 so that you may easily put them back in order in case you drop the deck. A computer print-out of the deck follows which shows you what is punched on each card. The portion of the main program which consists of call statements for subroutines is to be placed between cards 26 and 27 and the data cards between cards 83 and 84 of the deck. 146 There are six possible error messages which will be given as ERROR __. The following is an explanation of what each error means: 1. 2. 3. NO rows or columns in the matrix. The zero matrix. The number of equations is less than the rank of the matrix. > More equations than unknowns. All sums are zero for the active forces and moments. Therefore, the body was in equilibrium without any reactive forces or moments. Inconsistent set of equations. 25 147 COMPUTER PRINT-OUT OF MAIN PROGRAM COMMON 5 COMMON PTLN!5.6),PTNAN(5).F(7I95VSTEMI6.6).NRI6I'NCITI COMMON ARRIvaIoFNAMIOIvSMNAMIOIoSUMSI6Io$CHCIIoANAMIOI COMMON IRDHELOJIICOLHLII DIMENSION lROH(6)oICOL(7IoS(42) EQUIVALENCE (ARRlleIpSIIIIvIIROHM(IIoIROHIIIIo(ICOLMIIIuICOL(III EDOUBLE PRECISION DUMPIIJFORCES.SUQFQRLPILINELMDMPToSUMMCNLZSMDM DOUBLE PRECISION lPTLNolFORCQSULVEQZARRoCOUPLEoOLOAOoMFGRRgEQUA DOUBLE PRECISION IYPSYSOCHECKOFILKHEOFILKHCoFILKMMgFINISH REAOIIOIIBLANKOISUMSIIIlI'IDOIOSCH FORMAI‘A406A3OAII REAOIIoZSII(SYSTEMIIOJIOJ'IOCIOI'IOOI FORHAIIQLCAQII REAO‘II5)ZFORCOSOLVEOCOUPLCOOLOAOOMFGRRerUA REAOIIOSI FORCES'SUMFOR'PTLC~ECHOHPTQSU”MDMQZS"OMQZPTLVOZARR READ!l.S)TYPSYSJCHECKgflLKMEnFJLKHC.FlLKNMgFlNlSH.DUMPlT FORMATIIOAOI OO 50 KKK'Ios - A aF I CALL LOADERIZSMOMvNMgBLANK, CALL LOADERIZPTLNONPOOLANKI CALL LOADERIZARRI REAOIIOIOIKOH IO FORMAII8X9I206X9I3I 66 ll LAfiBLIEL3g6OIKgfl FORMATIIHloSXo'CHAPTER',16,5X9'PROBLEN'915) CALL LOADER(FORCESoNFI 000022 000021 00002‘ 000025 OODO;§_ 21 CALL LOAQ§R(DL0A00NF) 23 31 61 63 CALL LOADERISUMFORgNFI CALL LOADERIPTLINEUNPI _CALL LOADERLMUMPTQAIIINFLNH’ CALL LOAOER(COUPLE¢NHI CALL LOADER(SUMMOH9NM,NFI ___Ah___CALL_LDAQEBILEDRCLNEEHLANKl SI 63 71 CALL LOADER(ZSHOM.NM.BLANKI CALL LOADERIFORCESnNFI ‘R DLNFL, CALL LOADERIMOMPToIoIoNFoNMI CALL LOADER‘COUPLE NM) 91 CA LL LO ADERIFILKHEI 9Q 83 SO CALL LOADERISOLVEVNFDNH NOR, NOCo BLANK. CALL LOADER(EQUA0NOTI CALL LOADERIIXPSISDKDMONORI CALL LOADERICHECKoKoMpMMpNOII CALL LOAOERIMFGRR,ARR.6,79607O1.05-70IRANKoIROHMuICOLMoSI CALL.LOAOERIOUMPIIQNOKJNOCLIRANK) CALL LOADERIFINISHDNORONOC’IRANK’ CONTINUE CALL EXIT END 000027 000028 148 COMPUTER PRINT-OUT OF THE DECK - ~---~.. II 005: OUIPUIvzcoo .w , _ _ “ _ __ _- ,_,_ ,_ 000002 - II JCH FORCES 00000i _II CPIIUN LINK.“ _, 000004 9FASL EDRIFAIN.NOCI 000005 [I ExCC FCRIRAN V -7000006. corch EIECI¢.¢I.srn~IA.31.sr¢31.er31.vas.7I.NUFLI2.2I 000007 COPPCN ptLNIE.LI. PIAANIII. F17).SYSIEP16. 6). NRIAI.NCI7I __ 000000- COVPLW A0010.71. r'AHICIoSVNAPIOIISUV§16)4‘CHIII. ANA“(6) 000009 DIMENSION INLAILI.ICLLI7I.: I421.IQUUvIUI.ICULrI?I_ 000010__ EDUIVALLNCE (Aflkllo 11.511)1.11wCUFI11.InDuIIII.IICCLMIII. ICCLIIII 000011 , LOUdLE PRECISION ECRCES.SL~EC".PILINE.M0vpt.suw~0M. zsrcw , 000012 _ DOUBLE poECISICN zpan.zEcRC.soLvE.7ARR.CCUpLE.nLOAD.rr0nR.E0UA 000013 DDUULE PNECISICN vasvs.ChCCK.EILNFE.FILAFC.FILNFF.EINI$N "000014. PSACI1.1)HLAI.K.ISUFSI13.181.63.5CH 000014 ___1__ FURNAIIA4.EA3.AII 000010 PEADII..>III$I$IEVII. J1. 4:1. AI.1-1.0I 000017 25 FORVAII316A41) a ____y , 00001! _ pEADI1.5I2EDRC.SDLVE.CCUPLE. DLCAD. PFGRRUECUA 000019 READII.51 FORCFSISLVFUR. PILINF. POPPY. sunnor.zsncv. zFILN. 219R 000020 PEADI1.5I7Ypsvs.CHECN.EILxNE.EILxrc. FILKMF.FIN|$H 000021 5 _FORV_AT(1CAGI 000022__ CALL LDACERIzEDRC.NE.ULAAkI 000023 y, CALL LOADERIISVOPIN”.FLAAKI.- * _ _- g ---H,- _“ -___ _000024 - CALL LDADCRIszLN.NP.ULANxI "000025 .”. CALL LOADERIZARfi) _ _ p _- --_______ _ - M__ __000020”_ CALL exII 000027 END 000020__ I- 000029 _-IKCLUCE IJYAQXIA_ _ _ _- __ _-_ ."____.___- -“.m0000030._ INCLUDE 1415501 000031 INCLUDE IJISSCNQ.h-h_ _- _- I. a -_ - __ ___“ ”H*______ M000032 [\CLUDE IJISLOC . W00033 __INCLU0E_IIIE59N 000034 INCLUDE IJIARxR 000039 _ PrASE COUPLEoO -h - ___,____ __..._.L. _ -_.UL ”_"___u_______“_~__w___ 000030_ INCLUDE CCUPLE ”000037 _ PtASE SOLVE.CDUPLE,__ ,,.__ _ _"_ __ _-_ ._____u___,_-___“*_ _ -000039 _ INCLUDE SOLVE 000039 _PPASE IEDRC. COUPLE 000040 :INCLUDE ZFCRC 000041 _ PrasE ZARR.COUPLE .h---h “ _ , .-.*n L.-M.-__h.qu-_-..___._. M000042 INCLUDE ZARR 000043" _ PFASE ISM0~.COUELE ”w _______ __ __ _ - _ * .__*-___,___.__-_____ #000045_ [\CLUDE 2520» #000045- PrASE zeILN. CCUBLE _ _ _ 000040 INCLUDE zPILN 000047 - PtASE suvvov. CFUPL£._ 7‘,,_a_d -_‘ ,.-.L-L.__.,_-m,4_.-mr---_ --_.L._-°°0°*§p" INCLUDE surro» 000049 _ PPASE FOPPI.COUPLE _” ‘m‘.~_r _-_ ,__.‘-_ _LA-I_____._,--__I__h._-u 000050~_ INCLUDE POPPY ”“000051 _hPrASE PILlNEo CQUFLE___ ____ 000052 INCLUDE IILINE 000053 PrASE SUPFURICOUPLE‘- -___ I". -7. H.-.A._.L.._L.u__u..__u m_ _ 000054_“ INCLUDE suwsoa 000055 PPASE FORCEs.CDUPLE_m_ _"U__