Kiewit: A Man-Machine Success Story
As a pervasive educational force on campus,Kiewit Computation Center in five years has taken onequal status with Baker Library and Hopkins Center
With a battery of primitive calculating machines, a staff of 15 to 20 persons worked in three shifts, 24 hours a day, six days a week, for 18 months to solve the complex mathematical problems associated with the construction of the first atomic bomb.
Today, any sophomore at Dartmouth College, starting from scratch, can in a single afternoon do all the mathematical computations that were done in a year and a half at Los Alamos.
And that student can do his work while 100 or more other students or professors are using the Kiewit Computation Center and the famed Dartmouth Time-Sharing System (DTSS).
When computer time-sharing came to Dartmouth at 4 a.m. on May 4, 1964, only three persons could use the system at the same time. In less than a decade, the DTSS has grown so rapidly—mainly to meet the everchanging needs and greater demands of its users—that in the final week of Winter Term 1972 a record of 133 simultaneous users was reached.
Of the 3250 undergraduates at Dartmouth in 1971, fully 2600 were considered frequent users of the giant computer housed in Kiewit Center. Well over half of the 600 graduate students and 400 faculty members logged time on the computer, with the heaviest users being candidates for the MBA at the Amos Tuck School of Business Administration.
But Dartmouth is not—never has been—content with just providing computer services for her own people. Along with the on-campus users in 19/1. some 9800 others—students, teachers, doctors, researchers, scientists, bankers, and businessmen—also signed onto the system via a unique and novel network connecting New Jersey, New York, five New England states, and a Canadian province to Hanover and the Kiewit Computation Center. And the number of off-campus users continues to grow each year while the percentage of Dartmouth undergraduates acquiring "hands-on" experience with man's newest tool creeps ever closer to the 100 per cent mark.
The Kiewit Computation Center, which was dedicated five years ago last December and named in honor of its chief benefactor, Peter D. Kiewit '22, has been the most pervasive influence in intellectual life at the College since the construction of Baker Library in 1928- 29. Hardly a student, professor or administrator can escape the computer's influence; hardly anyone really tries. For many students, as well as faculty, not having a computer at their constant disposal is unthinkable. The computer is at anyone's disposal 18 hours a day, seven days a week, 365 days a year; and yet a team of undergraduates called systems programmers are hard at work developing programs and equipment to permit Kiewit to run around-the-clock, not to set some kind of record but to meet the constant demand for more computer time.
The Dartmouth Time-Sharing System was designed principally by students for students under the direction of President John G. Kemeny, then Professor of Mathematics, and Thomas E. Kurtz, Professor of Mathematics and Director of Kiewit Center.
In 1959, Dartmouth acquired her own computer, an LGP-30, and a team of four undergraduates—including Dr. Robert F. Hargraves '61, Associate Director of Kiewit, and Dr. Stephen J. Garland '63, Assistant Professor of Mathematics—went to work producing small programs and a computer language for the new computer. It was their early work that convinced Kemeny and Kurtz that a viable and economically feasible computer system for students and faculty could be developed for the College. The LGP-30, however, "was too small for a large number of students and after five years the students literally and quite physically wore that computer out," Kemeny recalls.
Nine years ago, Professors Kemeny and Kurtz circulated a paper on campus calling for the creation of a massive computer center for students and faculty. In specifying a high-speed, time-shared facility, the two young mathematicians said knowledge and use of the machine must become a basic part of every student's education because of the long-range impact the computer will have on everyday living let alone education.
"An institution like Dartmouth College that trains many of the leaders of the nation in business, in government, and in research, has an obligation to make sure that our liberal arts graduates understand one of the most significant factors that will influence their lives," the two futurists said in 1963.
John S. McGeachie '65, director of data processing at Dartmouth, remembers: "I had never heard of a computer until the spring term of my sophomore year. My first involvement was in a course taught by Professor Kurtz in which you had the choice of doing problems by hand, by desk calculator, or by computer." McGeachie continues: "In my junior year, Professors Kemeny and Kurtz recruited a half-dozen students and we met periodically, to talk about a time-sharing system."
The professors emphasized that the system—whatever its final form—should be designed in such a way that it was easy for students to use and sophisticated enough for hundreds of people to perform thousands of small jobs on the machine. While the six undergraduates worked on the computer system, Professors Kemeny and Kurtz worked on a computer language.
Michael Busch '66 and McGeachie designed the Executive Program which enabled the original GE-235 computer to talk to many people at the same time while also making internal decisions on whose program should be given priority. The professors developed the interactive computer language BASIC, a combination of English and high school algebra. Since its birth that early May morning in 1964 BASIC has traveled far beyond Hanover to become one of the most widely used computer languages in the world.
"What we did," McGeachie relates, "was to take off-the-shelf equipment, hook it together in a way never before thought of, write a language that was easy to learn by people and easy to understand by a computer, and call it time-sharing." In nine months, students and professors, supported in part by grants from the National Science Foundation totaling over $500,000, developed what has come to be known as the Dartmouth Time-Sharing System. It has been conservatively estimated that for a commercial computer company the same project would have cost between $2 million and $3 million and have taken several years to accomplish.
So much for history, other than to say that in less than a decade the system has grown to maturity from three simultaneous users to a potential 160; the old GE-235 has given way to two Honeywell-635 central processors, two Datanet-30 communications computers, and more than 300 teletypewriter and video display unit terminals located throughout the campus and up and down the eastern United States seaboard.
In creating DTSS, the system was designed as an educational tool first and a research instrument second. Originally, the computer was used primarily in mathematics. Students could feed into the machine a complex problem that required a great deal of arithmetic and the computer could come up with an answer in a fraction of a second. Then, as now, the computer can perform more than 500,000 calculations a second. By letting the machine do the so-called "grunt" work, the math student was free to concentrate on the logic behind the solution without worrying about making a mistake if he did the calculations by hand.
Today, however, mathematics is not the only beneficiary; the computer has found its way into virtually every academic department at Dartmouth and many of the administrative offices as well. At last count, there were more than 30 academic departments using the computer—everything from anthropology to zoology, including the classics, government, psychology, sociology, and urban studies.
With a College catalog offering more than 800 courses, it has been determined that more than 100 courses involve some use of the computer. And while a course—or an instructor—may not specify computer use, students are often ingenious enough to find a way in which to utilize the electronic brain on their own. Indeed, many of the courses which today depend heavily on computer power came to this position because students, using their own imagination, found .the computer could do things their professors never thought of. In a way, one development breeds another, and the effect at Dartmouth has been that of a gigantic computer snowball, building in size as well as quality, with use bounded only by the limitless minds of students and their teachers.
Computer power manifests itself every time one of those nearly 13,000 annual users sits down at a terminal and logs onto the DTSS system. A few examples:
• The graduate student at Tuck School simulates in one afternoon the activity of a stock portfolio over a five- year period on the New York Stock Exchange. He might acquire the same knowledge about the market and the portfolio by reading the financial pages for five years, but he has only two years in which to complete his MBA program.
• Students in sociology, anthropology, psychology, economics, history or geography have at their disposal Project IMPRESS, a bank of more than 50 data sets containing statistical information on past presidential elections, student-faculty-administrative attitudinal surveys and cross-cultural studies, to name but a few. Using IMPRESS, the student lets the computer do the "Mickey Mouse of statistical drudgery" while he concentrates his mental energies on examining relationships among variables within his own discipline. Instead of merely reading about research results in the social sciences, students using IMPRESS can actually challenge those findings on their own.
• President Kemeny, sitting at his personal console in Parkhurst, writes a simple program instructing the computer two blocks away to produce a table which projects the financial needs of the College for the next decade. Built into his program—his model of the College—are many of the complex variables such as student and faculty size, revenues raised, money spent, anticipated expenses, unexpected contingencies, and a host of other possibilities. These variables, reduced to algebraic equations and simple BASIC commands, will instruct the computer to perform the necessary calculations and produce the requested table of data. "Man working with a machine can do more in less time than either working alone," the President remarks. "Given the complexity of decision-making today, I don't see how an American executive can get by without a computer."
• At precisely the same time that the President is modeling the College on Kiewit's computer, dozens of students may also be playing computer games such as a simulated moon landing, a program which provides in five short minutes a thrill many students may never experience in a lifetime. Game playing is an important aspect in computer power. Not only does it provide a degree of relaxation and recreation, but many "games" are designed to introduce and teach students one or more specific scientific laws. The moon landing game, developed shortly after Neil Armstrong first stepped foot on the lunar surface, involves the same complex problems Armstrong faced: moon gravity, fuel supply, thrust, speed, direction, time, and other variables. The game isn't an easy one to play and one is as likely to have the computer comment, "Splat, you have just crash landed on the moon, try again," as it is to say, "Congratulations, you're the hottest thing since Charles Lindbergh."
• Dr. Stephen V. F. Waite heads a unique project in the classics. Editor of Calculi, an international publication for classicists interested in computer power, he has stored more than 160,000 lines of Greek and Latin text on the Kiewit computer. Included in his bank are the complete texts of the Iliad, Odyssey,Aeneid and the New Testament, all in either Greek or Latin and accessed either of the languages. And while he peruses any part of the 300,000 computer words of the Iliad, dozens of students are sitting at terminals practicing language drills in French, German Spanish, Italian, Latin, Greek, Russian and Esperanto.
The list of computer uses at Dartmouth is endless. Like space use is limitless. With more than 500 programs filed for public use in the computer memory system, the fertile minds of young men and women constantly design newer and more novel programs as though they are never quite satisfied they have exhausted the resources of their own creative human imagination.
The variety and preponderance of use often sounds like Babel, but development never goes unchecked for long. The first major study of computer applications in undergraduate curricula was carried out at Dartmouth under a federally funded program called COEXIST. Under this project, supplementary computing material was produced for use in college-level mathematics and physics courses.
Recently, the National Science Foundation awarded the College a $450,000 grant for the establishment of a center for Computer Oriented Materials for Undergraduate Teaching Teaching (COMPUTe) which will carry Project COEXIST a step further. Project COMPUTe will build on the knowledge acquired from its predecessor by pulling together the creative talents of active and sophisticated teachers who are engaged in the development of curricular applications of computing. This team, in turn, will work to produce high quality textbooks with computer problems and exercises keyed specifically to texts, first in the field of science and gradually in other disciplines.
With COMPUTe barely off the ground, Project CONDUIT is already on the drawing boards. A two-year experiment in the exchange of computer-oriented course materials, CONDUIT members include Dartmouth, the University of Texas, Oregon State University, the North Carolina Educational Computing Service, and the University of Iowa. Like Dartmouth, these schools are all centers of educational computing networks serving schools and colleges in their regions.
The object of COMPUTe is to reduce computer-oriented programs, while CONDUIT, as its name implies, . to disseminate nationally a master catalog of educational programs while at the same time facilitating the exchange of programs within the high- speed educational computer networks and outside them. The disciplines involved under CONDUIT are mathematics, physics, chemistry, biology, sociology, business-economics, and accounting.
And speaking of networks, Dartmouth operates the largest educational time-sharing network in the world. At last count, some 20 colleges and 30 high schools in seven eastern states and the Province of Quebec were linked to the Kiewit Center via ordinary telephone circuits and special switching equipment. With an estimated 100 terminals located off-campus, more than 65,000 hours were logged in 1971 on those terminals by students at the participating institutions. Another 2000 hours were recorded by other off-campus users in research and business.
In addition to the educational timesharing network, Dartmouth also managed a Regional Computer Consortium under federal funding which ceased in June 1970. In its final year of federal support, however, the consortium had 13 participating colleges in New England and the service provided first-hand computing experience for 2300 students and faculty. Despite the loss of government funds, the number of colleges using DTSS during 1970-71 grew from 13 to 29 and the total number of students served increased to more than 3000.
But even the consortium isn't the end of outside service for Dartmouth. A third organization, the New England Regional Computing Project (NER-ComP), is a non-profit membership of New England colleges and universities which share computing resources and knowledge. Dartmouth exports DTSS service to nine institutions in the Boston area alone over three computer lines.
In the past year, many developments have transpired for Dartmouth and her Kiewit Computation Center, but two can be considered particular standouts.
The most noteworthy event was the installation of an exact duplicate of the Dartmouth Time-Sharing System at the United States Naval Academy at Annapolis, Md. On Feburary 17, 1971, the Naval Academy system was dedicated, making it the first educational institution in .the nation to duplicate in toto the Dartmouth system. When the Academy's system came "on-line" it provided for 4200 midshipmen and faculty the same high-speed, time- shared services that had been enjoyed for the past nine years at Dartmouth and other regional schools. Some 1400 entering midshipmen each year are required to take an introductory course to learn about computers and to write and execute their own computer programs—in BASIC, most likely, but then again, like the Dartmouth system, students might also utilize one of the other six computer languages compatible with DTSS.
And finally, the new Cancer Research and Treatment Center at the Mary Hitchcock Memorial Hospital in Hanover is making final arrangements for implementing a cancer registry for both New Hampshire and Vermont. When completed this summer, the registry will contain the personal and medical records—some 87 different entries per person—of about 50,000 cancer victims in the two states. The registry presently contains information on 9000 cancer patients, all or most of whom have been treated at the Hitchcock-Dartmouth Medical Center.
The cancer registry will be completely computerized on the Kiewit system for use by therapists studying survival statistics and registry officials keeping tabs on patients who are required to have periodic check-ups with their physicians.
A second and equally important function of the Dartmouth computer for the Cancer Center is the production of individual radiation treatment plans for patients. Using sophisticated electronic equipment, therapists can provide the computer with essential data about a patient—information relating to body contours, radiation field sizes, and location of tumors and healthy organs—and the computer can perform the thousands of calculations required to produce a series of acceptable radiation treatment plans. Up to this point, the patient has not yet received any treatment. The initial process of man working with computer is simulation, and through a question and answer format, man and machine eventually arrive at the best possible plan of treatment. The final decision of which plan is best for the patient, however, is left to the therapist, not the computer. As in other instances at the undergraduate level, students have been responsible for developing the many dosimetry programs as well as the cancer registry which are now stored in the Kiewit system.
Kiewit officials estimate that one student systems programmer is worth $100,000 a year at commercial rates for the work he has done in keeping the DTSS system healthy, happy, and alive. At the same time, an official of the Cancer Center noted that one Dartmouth student, now in his second year at Tuck School, has accomplished more in six months of part-time work in getting the improved cancer registry and systems programming ready for full implementation than could be accomplished in the commercial world in two years.
It is this kind of capability and dedication on the part of students and faculty alike that has made Dartmouth such a pioneer in computing and that will keep the Dartmouth Time-Sharing System the envy of the educational world.
A not unusual "no vacancy" situation in Kiewit's public teletype room.
Prof. Thomas E. Kurtz (left), director of Kiewit Computation Center, plays hostto Vice Admiral James S. Calvert of the U. S. Naval Academy, which now has inoperation a duplicate of the Dartmouth Time-Sharing System.
Part of the "hardware" that teams with human brainpower at the College.
A study in the spring of 1970 showeduse of DTSS by 871 upperclassmen, indicated on the chart by their majors.
Another Kiewit study disclosed use by136 faculty and staff members during aperiod of three months last year.
