First-Hand:39 years with IBM
Submitted by Joe Kuhn
I graduated in September of 1948 from Carnegie Tech. We had no ceremonies at that time. I never did get to the formal graduation ceremonies. I never interviewed with the large companies that came to school. I wanted to work for a small company. I did interview with an independent phone company north of Pittsburgh but didn't want to be that alone. In the yellow pages I noticed this company: International Business Machines. It sounded interesting so I arranged an interview. The manager, Jeff Davis, who interviewed me was impressed with one machine they had that had vacuum tubes in it. I was offered a job and I accepted as a Customer Engineer. That was a repairman. I started work October 18, 1948 and graduated from their school June, 1949. There was a key punch that needed no electricity, only finger pressure put holes in the cards. Accounting machines did all the work.
In 1954 we moved to Poughkeepsie where a small group was to learn the new computers so we could subsequently teach Customer Engineers how to repair them. At that time the 701 was the only large one in the field. Under development were the 702, a business oriented computer, the 704, a scientific follow-on to the 701, and a 703, a new machine tailored to sorting. The memory at that time was an electrostatic tube storing 1000 bits. That was the memory of the existing 701 and the new coming 702. The 703 was to have an electrostatic tube which would store 10000 bits. WOW! Sorting evidently took a lot of computer resources at that time and much new knowledge was being developed on sorting algorithms. The 703 never went into production. The few of us who had been studying that machine were assigned to the 702. That machine was short lived also because a new magnetic core memory was on the scene. Some 702's were produced but the emphasis was now on the 704, the scientific computer, and the 705, the business computer. The 702 and 705 had only 35 instructions. Everything was new. We all spent a lot of time learning just before teaching. It took a lot of effort keeping ahead of the students. As you might expect many intelligent people were in the classes. We got up early to prepare or duplicate material to distribute to the class. Eventually I taught everything about these computers from the basic circuits to programming and even the assembler and even the oscilloscopes that were used for analysis. I felt I knew everything about the 705.
I eventually became a manager. Mr. Wolfe, the head of all Education in Poughkeepsie, and Pat Beatts, a teacher's teacher, brought in closed circuit television for teaching. This was really new and we tried a lot of different things. Pat was really a leader. He helped all of the instructors preparing visual aids and teaching them effective presentations. Television was a forcing function. The instructors prepared especially well because so many would be seeing them. Television was very effective in debugging the machines. We would show a whole class symptoms of a problem on the machine. The class would ask to see other conditions and eventually solve the problem. Critiquing the approaches helped all in good techniques of debugging. We also made a kinescope that was distributed to the field to teach proof machine operators how to operate this bank machine.
New computers were on the way and I was transferred to Manufacturing Engineering in the plant. A new function was being tried. We held allegiances to a particular machine and coordinated the application of the various M. E. functions in putting that machine into Manufacturing. Transistors were being introduced now and there were transistorized versions of the 705 and 709. (7050 and 7090). Also the 7070 manufacturing was being transferred from Endicott to Poughkeepsie. Then the 360 family was coming into being. Five different sized computers all based on the same architecture. Very ambitious. Building blocks were common hardware but there were a lot of changes in getting those machines into Manufacturing. I was looking out for the 360 Model 40, the second smallest of the lineup but it was designed in Hursley, England and was released in Poughkeepsie. Getting the common pieces identified for Hursley to pick up usage and communicating with the five hour time difference challenged everyone. April 1964 those machines were announced and work was hectic until the first machines were delivered. The Model 40 had a transformer read only memory manufactured in Essex Junction, Vermont. Sequencing information between Hursley, Essex Junction, and Poughkepsie was a challenge. One very intelligent engineer designed and built a tester to unit test the main frame box. It could test the error circuitry better than the computer diagnostics. It didn't fly because computers being on the final test floor has to run the diagnostics anyway and engineering support to update the unit tester and it's routines might be excessive. Ever hear of GPSS (General Purpose System Simulator)? We successfully used it to verify that one tester would be sufficient to test the large (about 1 by 4 feet) core planes of the large capacity memory. We successfully permitted the merger in the field of computer units from the different manufacturing plants. Learned about learning curves as applied to final testing. In memory engineering I learned the insides of the manufacturing costs of a product over it's life.
In 1967 I moved to Essex Junction as a cost estimator. Ironic! Economics was the only subject at Carnegie Tech that I had a D. The product being engineered then was a flat film memory. It was fast and the last of magnetic main memories. Solid state memories took over. They improved fast in capacity and speed. Here again GPSS was used to simulate the processing of wafers to plan tools, space, and people.
Now the computer was used extensively to design the computer: documenting, simulating, designing the hardware, designing the testing, and releasing to manufacturing. My last activity before retiring in 1988 involved logic synthesis. One could map from one technology to another or from a software design to hardware. Verifying timing's and modifying to comply was part of the activity before releasing to manufacturing.
Today I still enjoy Spectuum and Computers. How things have changed. Twelve holes in a card, ASCHI, EBCDIC, UTF. 1000 spots on a cathode ray tube, to I don't know how many bits in a module and memory sticks that are finding their way to replace the hard disc of PC's.
A drum that never made it. Rotating disks that took the space of a refrigerator to hard discs of today.