The History of Computing at Princeton
In 1930, when it began
The history of computing at Princeton began as early as 1930 with the efforts of the great mathematicians Oswald Veblen, John von Neumann, and Alan Turing. Veblen sought to speed calculations for ballistics, von Neumann was fascinated by the mathematics of turbulent flows and sought to determine how best to shape the charge of a nuclear weapon, and Turing dreamed about using calculating machines to reduce complex problems into solvable algorithms.
Their efforts led inexorably to the building of the world’s third computer at Princeton’s Institute for Advanced Study and for Turing, the building of a machine, the Colossus, that helped to break Germany’s enigma coding.
Housed in what is today a day care center, the Princeton machine was also used for academic purposes, weather prediction, studies in evolution, the analysis of stellar transitions, and even the modeling of traffic on freeways. With separate units for input and output, an arithmetic organ, memory, and a processor, the IAS machine became the architectural parent of our present day computers.
Progress in the 1950's
In the late 1950’s the computer, dubbed the Maniac (Mathematical and Numeric Integrator and Calculator), was transferred to the University, but by this time, the University was interested in acquiring its own machines.
In 1952, renowned Princeton statistician John Tukey acquired a digital IBM machine, one of only 13-17 that IBM constructed. This was the University’s first computer. Tukey installed it in one of Forrestal’s greenhouses as part of a military weapons analysis group organizationally reporting to the Statistics Group. The offices for the effort were in the Theobold Smith house at Forrestal (on the western side of Route 1, the former Rockefeller Institute for Medical Research). The machine was a punched card programmed calculator that was used on a job-lot basis. It was served by a staff of four and led officially by Tukey, but he simply did not have the time for the effort.
A 1943 Princeton graduate, Forman Acton received his PhD in 1949 in Applied Mathematics in a new department at the Carnegie Institute of Technology. His first job was in the Mathematics Branch of the Bureau of Standards in their California office at UCLA where he gained experience with one of the new digital computers from IBM.
Tukey extended an invitation to Acton, who agreed to return to Princeton to lead the four-member group. Bell Labs fed them problems, all with top secret classification. One of the more interesting problems had to do with the warhead design for the first Nike missile, a fragmenting, not nuclear, missile. The group operated for approximately four years when Acton disbanded it.
Irv Rabinowitz, one of the programmers on the von Neumann machine, was the first person who spent the time to develop his own expertise in the computing area and essentially took over the Forrestal laboratory and made it run. The early equipment was quite massive and slow and required a great deal of cooling to handle the heat loads. This necessitated an inordinate amount of time designing the infrastructure. Most of the theoreticians had very little training in the use of computers and had to learn Fortran to be able to use them at all.
In 1959, the University acquired an IBM 650 and installed it in Gauss House, an unimpressive small yellow Victorian house on Nassau Street between Aaron Burr Hall and 185 Nassau St. (a few doors down from the present Thomas Sweet). The main users were the Statistical Techniques Research Group (Population Research) and the School of Engineering. The mainframe was situated on the back porch, the largest open area in the house. Data and instructions were stored in the form of magnetized spots on the surface of a drum four inches in diameter and 16 inches long, which rotated 12,500 times a minute. The drum memory could hold 20,000 digits at 2,000 separate "addresses."
Recalled Forman Acton: “I remember being worried about sitting in front of it in case the drum came apart.”
Recalled Ted Dolotta: “When I was a graduate student, they would hire me to run their programs because someone had to load the program deck followed by the data, all on punch cards, and it would grab a card every two minutes or so, compute, and spit out a result. Every two hours or so, you had to put in more cards. And so I used to make a buck and half an hour. I’d sleep on the floor, and when the card punch would stop spitting out cards, I’d wake up, load more cards, and then go back to sleep.”
A few months later, Project Matterhorn, now the Princeton Plasma Physics Laboratory, also acquired an IBM 650 that later came under the control of a University-wide committee, the forerunner of the Computer Center. In 1960, the University’s Institute for Defense Analysis (IDA), a government supported classified institution with its own building (now Von Neumann Hall), acquired its first computer.
Facing a flood in the IDA machine room and devastation to its CDC 1604, Jerry Mitchell telephoned the Princeton Fire Company urging them to come and pump out the room. When told that they only come in the event of fire, Mitchell responded “How big a fire do you need?”
In exchange for the space, the University was permitted use of the IDA computer for 20 hours a week. But because IDA’s work was classified, staff had to lock away all of their documents every day before faculty and students came in to use the 1604 in the evenings.
The greater demands of the 1960's
By 1961, the demands upon the IBM 650s become so extreme that they were running 91% of a 7x24 schedule. In the spring of 1962, in his junior year, Lee Varian was introduced to computing on the IBM 650. After one day, he and his roommate Alexander Jones become certified on the IBM 650 and gained keys to Gauss House, where the IBM 650 computer facility was located. It was discontinued on December 31, 1962 and returned to IBM. With the University’s computer requirements burgeoning, the University’s first Computer Center was created late in 1961 in a space reserved for it in the newly created Engineering Quadrangle. Edward J McCluskey, Jr, a professor of engineering with a specialization in fault tolerant computing, was named its director in 1962. Two of his graduate students become well known at the Computer Center: Ted Dolotta *61 and James Poage *63.
The first computer in the brand new E-Quad Computer Center was an IBM 7090 (32K words, 197K bytes of memory). The machine, operating by December 1962, was the workhorse for many undergraduate projects and graduate theses, primarily in the sciences and engineering, for the next seven years. The computer was later converted to a 7094 by upgrading its memory and by replacing some of its components with faster ones.
The Computer Center instituted “a daily consultation service” where members of the computer center were available in a conference room to provide individual help on programming problems. There was also a weekly seminar on novel uses of computers in various disciplines. Some early "clinic" humor included a sign posted in the clinic: “For several days, the random number generator has been generating random numbers. For assistance, please see Alex Jones.”
The 7090 had more than 50,000 transistors and 1 million magnetic cores, performed 229,000 additions per second and, with its auxiliary equipment, cost more than $3,500,000. The cost was met in part by a NSF matching grant of $700K and an IBM educational allowance of $1.9M. Princeton Alumni Weekly stated: “As the body of scientific knowledge explodes, so does it price.” Two FTEs were added when the 7090 arrived to bring the professional staff to 12.
In February, 1963, Roald Buhler, formerly a consultant on programming for the Social Science departments (Educational Testing Service), joined the computer center programming staff “writing standard programs which will be used in several departments and providing consulting services to users requiring help with their programs.”
In 1966, Ed McCluskey resigned his position in order to devote full time to his teaching and research in Electrical Engineering. Roald Buhler succeeded McCluskey as Director of the Computer Center.
A 1952 graduate of Oberlin College, Buhler did his graduate work at Rutgers and was associated for three years as a Research Computer Scientist with ETS in Princeton. Says Buhler: “Computers are gorgeous for statistics, but they are also extremely useful for anything that is simply described and very repetitious.”
During the 1960s, under Buhler’s direction and with support from the NSF, the University began the design and then the construction of the Computer Center building at 87 Prospect Ave. The administrative staff included Knawm Friedman (Operations Manager), Ted Dolotta, and the Systems Programming Group (Lee Varian, Jack Benoit). There were approximately 33 FTEs at the Computer Center.
In 1965, Buhler and Ken King at CUNY organized the first of several conferences of Ivy+ computing support personnel to discuss future hardware. The first meeting, held at Princeton, invited all main vendors of possible hardware to give presentations. After some reluctance, IBM sent their biggest guns. As a result, the Ivy League swung almost entirely behind IBM as the most credible prospect for interactive computing for the next ten years. Only Dartmouth differed. They stayed with Honeywell 6000 and Dartmouth Timesharing.
In 1966, planning began to transfer the Computer Center from the E-quad to a new $2,000,000 building on a four acre plot near Palmer Stadium with nearly twice the space occupied in the E-Quad. Princeton conceived of the new Computer Center as a University-wide scholarly resource, like the library, which members of the University community could use without charge. The new building was purposely located to be convenient to Peyton Hall (new home of Astrophysical Sciences) and to other buildings rising for the Department of Math and Physics. It was also intended to be near the E-Quad and the campus headquarters of the humanities and social science departments.
Quite the story teller, Buhler related tales from the late 1960s. In one, a red-faced user lost an hour of computing to daylight savings time and was mollified by the promise of an extra hour a month later.
And there was the letter from Robert Oppenheimer, then director of the IAS, requesting time on University systems in compensation for their donation of the Maniac.
In large part as a result of the efforts of Prof Frederick F. Stephan, Princeton was one of the first universities in the country to realize the unique needs of social scientists and subsequently humanists for special services in the areas of data preparation, data management, and data analysis. He also recognized the need for the development of a new information resource, the data archive, a collection of data in machine readable form for secondary analysis.
In 1965, Judith Rowe joined the Office of Survey Research and Statistical Studies (located in the basement of Green Hall). Carl Helm was Assistant Director. Judith’s husband Peter Rowe was the IBM Salesman. Her hire began a period when many women were being hired as programmers. Shirley Gilbert, Judith’s first programmer, wrote a program called Survey. Subsequently Nita Rome, our second programmer, used to walk from Green Hall on Washington Road to the computer center at 87 Prospect carrying their cards. When driving, programmers parked on the roof above the machine room with a note in their windshield “Carrying Cards.”
Under Judith’s direction, Princeton became a leading preserver and center for analysis of the US Census.
"The sorter that’s downstairs on the first floor of 87 Prospect… that was our sorter," said Judith. "When I came to the computing center, we brought it. Knawm Friedman acknowledged that he removed the brushes and internals to discourage any further use."
With direct charging, every job card now had to identify the department. At the end of the year, there was an adding up of all of the sponsored research use of the machine and everything else (faculty and student use). Having done that, the percentage of sponsored research was thrown into the overhead rate for the next year. Princeton was the last significant university to go to some sort of specific charging. Auditors were complaining that the University was using more computing time than the Air Force.
By 1969, the University’s primary source of computing was the IBM 360/91 installed in the computer center at 87 Prospect Avenue. Work is submitted to the 360/91 as decks of punched cards. Remote Job Entry (RJE) stations for submitting work to the 360/91 were located in the E-Quad, Jadwin Physics, Plasma Physics, Guggenheim Lab, Geophysical Fluid Dynamics Lab (GFDL) New South, and a number of small stations throughout the state. The load was averaging more than 1,000 jobs per day; 75% experience turnaround of less than 30 minutes.
In July, 1970, James Poage ’55 became the third director of the Computer Center. Management included Lee Varian, Hale Trotter, Knawm Friedman, and Judith Rowe. Howard Strauss took over the management of the clinic, which had been reporting to Hale Trotter.
In 1970, Joyce Bell became the supervisor of the keypunch operators. “I was petrified,” said Joyce, “but I figured that I already knew how to punch those keys… how hard could it be to supervise?”
In 1971, ASDP [Administrative Systems and Data Processing] merged with the Computer Center operation. ASDP had their own mainframe, an IBM 360/50, until June of 1971. Beginning in June 1971 the ASDP load was transferred to the IBM 360/91 and in September 1971 the IBM 360/50 was sold and replaced with a RJE station to the 360/91.
Also in 1971, ASDP drew up a document to outline its long range plans over the next five years. Work began on a new payroll system, a new personnel and benefits reporting system, a new bookkeeping and financial analysis system, a new student record system serving West College and the Office of the Controller, as well as maintenance and refinement of the existing budget system.
In 1972, a survey run by the Computer Center with the cooperation of the Bressler Commission on the Future of the University tried to determine the penetration of computing into the academic programs. Questionnaires were sent to all 778 faculty. Responses were received from 271. Roughly 70% of the faculty using computers responded. 5-7% of undergraduate courses either required the use of the computer or had assignments where use of the computer was optional.
By 1972, there were 52 FTEs at the Computer Center site, and 20 part time students. ASDP had 13 FTEs performing system analysis and programming functions.
In 1974, Poage predicted “that the work done today by students, faculty, and some administrators including the use of the library for some functions will be done via interactive terminals connected to computer systems. We are already behind the field in our interactive and graphics support for students and research and it is precisely in those areas that technology and marketing strategy will bring the greatest changes.”
With these trends in mind, Poage established the Interactive Computing Graphics Laboratory in 1974 with a number of time-sharing terminals and supporting equipment. A laboratory area was set aside in the Engineering School, room E-432, for ICGL. In the first term, ICGL was used by 150 registered users from several courses, most notably Stat 101.
In 1975, Princeton purchased an IBM 370/158 to provide time-sharing services with the VM operating system.
The 1980s - the beginnings of microcomputing and networking
These efforts notwithstanding, in the early 1980s, the faculty staged a revolt over their desire for access to microcomputers and networking. The result was a faculty report on computing and the hiring of Ira Fuchs, then head of computing at CUNY and the co-founder of BITNET, a store-and-forward network that eventually helped to open up the USSR and China to telecommunications.
The 1990's - the decade of University-wide connectivity
Under Ira’s leadership from 1985-1999, Princeton greatly expanded the computing staff and built a world-class telecommunications infrastructure that interconnected every building and office within the University by 1990.
The New Millennium
Betty Leydon became the Vice President for Information Technology and CIO in 2001. Under her leadership, Princeton is taking advantage of that infrastructure to enable the effective use of information technology in teaching, learning, and administration.