A Place of Many Firsts

In the fall of 1961 hoping to improve the hopeless grades in both my major and overall, I signed up for Ms 400 - FORTRAN computer programming. It was only a one-semester-hour course, but I had been told by a number of people that it was almost a guaranteed "A" grade. The course was taught by Fred Japp, the university Computer Center manager. I did very well, and Fred asked if I would be interested in a job a the Computer Center. I had been working full time all the way through my college years and did not have as much study time as most students. I attributed my poor grades primarily to this fact. My work, night auditor at various motels in town, was not that challenging or inspiring. Since I was in my senior year I was eager to get a job which would provide useful work experience and allow more time for study. Half-time in the Computing Center sounded like a perfect job to me, so I told Fred that I was definitely interested. He asked me to fill out an application for him and that he would let me know.

After a couple of weeks, Fred called and asked how soon I could start work. I have forgotten the details, but I gave notice at the Holiday Inn. I also put in application with a foundation for a student loan to help fill in the financial gap between the full time and part time jobs. I began work and was trained as an operator for the Computer Center's IBM 709, a machine that took up a very large room with huge boxes filled with vacuum tubes. The primary box was the Central Processing Unit (CPU), a large L-shaped box with a panel of lights and switches which displayed the system status at any particular moment. Near the CPU was a card reader, a card punch and a printer; off down one side was a line of nine magnetic tape units. There was also another large box with a number of lights and switches which contained the circuitry for the machines two I/O channels. The IBM 709 was the first computer with channels which allowed it to accept new data into the "Memory" concurrently with the CPU accessing memory for computational data. The channels box was physically located about half way between the CPU and a large T-shaped box which was "Memory" or "Core" storage unit. Data were"kept" and retrieved from minute magnetic donuts, called cores, with three wires running through the center of each donut. Current flowing through the "x" and "y" wires could change the direction of the magnetic flow in the donut. The third, or "sense" wire, could determine which way the magnetic field was flowing to determine whether a particular donut was set to a "1" or "0." The last box was a Power Distribution Unit. It sent power to each of the of the other units through wires strung underneath the floor of the room. The floor was actually a platform that was about three feet above the real floor of the room. This hollow space allowed for signal and power cables to be strung between the various boxes. Cooled air was also blown into this space and up through holes in the floor and through the boxes to help cool the very hot vacuum tubes.

When I began working there the routine went something like this: a tray of punched cards was brought in from an assembly area where they had been submitted by researchers or students. On occasion there might actually be two or three trays of cards to be "input" to the computer. The card decks could be programs, data or both. A program was started that read the cards and placed them onto one of the magnetic tape units. This could take from fifteen to forty-five minutes, depending on the number of cards. After all the cards were read onto tape, the tape rewound, and there was yet another program that read the card images from tape and into the computer for processing. The results of this processing was written to another tape unit for later printing or for punching out cards on the card punch. After all the cards that had been read in had been processed or "run," the output tape was rewound and another program was started that read the images on the output tape and directed them to either the printer or the card punch. The routine went pretty much like this: get cards and for about a half hour read them onto the input tape; process the tape for one to two or three hours; print the output tape results for two to three hours, occasionally putting new paper into the printer when the box was empty. I had been working there only a few months when we hooked up an amplifier and speaker to the CPU. This really improved things for the operator since he no longer had to watch the lights to determine if the computer was running properly. He could sit with his back to the computer reading or doing homework while the low volume squeals and squawks assured him that the computer was still running. We had some extremely long compute-bound physics jobs that could run for four, five or six hours and operators learned to recognize what phase the job was in and how near it was to writing a tape or completing the run from the sounds coming from the speaker. There was a wall of windows across from the CPU. On the other side an observation room for visitors with this warning sign posted on one window.

Next

Number of Visitors to this page: Number of visitors to this page: