SPACE/1401 engineering story in Endicott, 1958 – 59
George Ahearn, Jan, 2011
Posted for review Sept 12, 2011
This story begins in 2Q 1958. I was a Navy officer assigned to the National Security Agency at Fort Meade Maryland. Knowing my military duty would complete in July, I was anxious to find engineering employment. The economy that year was not generating demand for new engineers. I was particularly interested in employment with IBM because of my association with that company’s products at NSA. A colleague, Holgar Jensen, was completing an Air Force assignment and was scheduled to return to the Glendale Lab in the summer. Holgar was in contact with Bob Evans, the manager of a new project with openings for engineers. As fortune would have it, that contact resulted in a job offer for me to join Jensen on Evans project. I started as an Associate Engineer, one step above Junior Engineer in recognition of my experience at NSA. The new project involved development of a complex system in a contract with NSA. According to a book: "Body of Secrets" by James Banford, this was a difficult period when NSA was unsuccessful in reading encrypted Soviet intercepts. There was speculation that this project, named "Maidenform" would be related to the difficulty. I say speculation because only a small select group of management and architects knew the purpose of the project. Holgar Jensen, Peter Fagg, and Bob Evans were in that group. Security surrounding the project in the IBM Glendale lab was exceptionally serious. When the select group convened they met in a conference room, called the Blue Room which was essentially a tank built to government specifications and periodically inspected for leaks. Every one associated with the project was required to have a Top Secret clearance, issued after a successful lie detector test. A team of federal agents appeared at the lab one week in July to administer the lie detector tests. That was the week I was in the hospital recovering from appendicitis. When I returned to work after medical leave, Bob Evans greeted me and told me to get on a plane back to Fort Meade for my lie detector test. I was not allowed near the project until the test was completed successfully. Paul Farbanish and myself were assigned to the I/O section of the project in a department managed by Wayne Brook. Brook reported to Peter Fagg who reported to Evans. At that time the strategy for developing I/O was unsettled, except that ability to read and punch cards as well as print output was required. Brook tasked Paul and I to survey activity around the Glendale Lab to get an idea about current I/O equipment in development. The SPACE program in the lower buildings was particularly interesting because it combined the card and printing equipment in a subsystem already in the design phase. The idea of integrating SPACE as a subsystem of the larger project appealed to the management. In August 1958 Evans approached Chuck Branscomb to negotiate a merger of effort. Chuck needed engineering help and Evans sent Paul and I on a temporary assignment to Russ Rowley, the SPACE CPU manager. Russ reported to Jim Ingram. Others reporting to Russ were:
The CPU design involved the register set, the arithmetic and logic unit, cycle control, and instruction implementation.
Another department, managed by Al Miller and reporting to Jim Ingram, was responsible for the integration of the 1402 and 1403 into the SPACE system.
- Bill Murray…1402 and 1403
- Don McLeod…1402
- Ozzie Brooks…1402
- Tom Schappe…1403
- John Jones…circuits
The project effort during the remainder of 1958 was second level design represented in block diagrams followed by transcribing the blocks to and’s, or’s, latches, triggers, line drivers/receivers using a CTDL catalogue of available circuits. The engineers favored latches because they were convenient to use and they were invented by Ernie Hughes in the Glendale Lab. Late that year the engineers were beginning to express their design using the Automated Logic Diagram (ALD) format.
 by G. Ahearn |
In December of that year the project that Paul and I were permanently assigned to was cancelled by NSA. Then we were transferred to Russ Rowley.
Initials in the drawing include "CEB"=Charles Branscomb, "FOU"=Fran Underwood. Further details are here. |
The effort beginning in 1959 focused on building, debugging and testing an ‘A’ test model of the system, which consisted of a 1401, a 1402, and a 1403. While the project enjoyed a reasonable priority in the Lab, the schedule was tight particularly in regard to CTDL card fabrication, ALD encoding, and computer processing, and later board wiring on the Gardner Denver machine. After several months building the system, power was applied and the 24 hour a day, 7 days a week grind began. Engineers and technicians staffed the shifts to diagnose problems and generate design changes. This was an awkward, people demanding period. In those days there was no such thing as simulating the design before fabrication. All the problems had to be discovered and resolved on that single ‘A’ test model and the clock was running toward the deadline of getting the system accepted in Product Test. We thought we knew the duration of the ‘A’ test and we were aware that the product could not be announced before a successful completion of that test.
During 1958 and most of 1959 the SPACE development was managed at the first level by Russ Rowley and Al Miller. They reported to Jim Ingram and he in turn reported to Chuck Branscomb. Fran Underwood was the architect reporting to Jim Ingram. The engineers looked to Fran for interpretation of system function. The implementation in hardware logic allowed the engineer flexibility, however Fran frequently suggested alternative approaches, usually with a view to cost reduction and simplicity of design. Fran had a keen interest in logic design, somewhat based on his previous experience teaching courses in the IBM engineer training program. SPACE was a smooth running team with mutual respect and very little ego driven conflict. The experienced hands, Russ Rowley, Al Miller, Bill Murray and Fran were tolerant of the new engineers and willing and eager to help with problems. Lunchtime for a few was a trip up the hill to the building 6 cafeteria, for others it was a brown bag sandwich, a drink from the vending machine and a game of bridge or pinochle. The vending machines offered instant coffee for 10 cents a cup. It was awful.
Product Test was an independent organization within the Glendale Lab. Their engineers had little interaction with the SPACE people before they accepted the system for test. It was expected they would gain their understanding by reviewing specifications and checking the system for conformance. There wasn’t much in the way of software to exercise the product, let alone diagnose problems.
An engineering career at IBM followed the “dual ladder.” Most new engineers started as a Junior engineer. The first promotion would be to Associate engineer. For several years, the next step would be to Staff engineer.
However in the late 1950s the Senior Associate engineer title was inserted between Associate engineer and Staff engineer. At the same pay level a first line manager received the title Project engineer. Here the dual ladder begins. On the technical side the titles above Staff engineer were Advisory engineer then Senior engineer. On the management side the titles were Project Engineer, Development engineer then Senior engineer. In general a Development engineer was a second level manager and a Senior engineer might have more than one Development engineer reporting to her or him. Titles were important to the engineers, not only because compensation was related to the title but also because promotions were published on all the lab bulletin boards as well as in the company site newspaper.
Endicott IBM had a Country Club facility, which offered golf, bowling, swimming and other recreation opportunities at very attractive prices. These were available to all the employees as well as to their families. In the course of a year several festive, site wide events were held at the Country Club. It was said that this was IBM’s way of matching the benevolent personnel practices of the Endicott Johnson Shoe Company, which for years was the dominant employer in the Triple Cities of Binghamton, Johnson City, and Endicott.
The Glendale Lab had a cafeteria with appropriate lunchtime food at reasonable prices. There is a story about one of the 1401 engineers gaming the system. This fellow noticed that an egg salad sandwich at the cafeteria cost $0.20. Clever as he was he also noticed that a hard boiled egg sold for $0.05, and two slices of bread cost $0.10. All the necessary condiments were available complimentary after the cashier station. So this fellow made the sandwich himself, at the table for an outlay of $0.15. Such a deal!