return to IBM 1401 1950s Team Bios
return to Fran Underwood

Remembrances

Fran Underwood, May, 2008

Bio: Fran Underwood (Francis O.)

Born: Omaha, Neb. 06/16/1926

Father was an Engineer and Architect, worked for Union Pacific Railroad. Mother was a registered nurse. Father died in 1930 from gangrene induced by burst appendix. No cure at that time. I lost a potent Mentor. Raised by Grandparents in Binghamton, NY. Mother received RN in New York, Master Degree in Public Health Syracuse.

In those days, we hadn’t heard the word ‘technology’. Leading edge stuff of the day was the radio, the Automobile, Steam Locomotives and Telephone. We used to play in the streets; "tag", "hide and go seek","kick the can", "Ring-a-levio", until it was bedtime. No thoughts of predators. We walked to school in all kinds of weather, no other way to get there. Binghamton was one of the "triple cities’, the other two were Johnson City and Endicott, home of IBM.

When it was time to enter High School in Binghamton, I was offered a chance to take a 3-year Technical Course (Mechanical) designed to prepare one for a job as a Toolmaker at IBM. I was taught drafting, operating manufacturing machinery for tool makes , sheet metal, math/physics engineering Very prestigious. There were only 22 accepted out of 2000 applicants. Graduates would be offered a job to IBM. However, WWII occurred so after High school I joined the Naval Air Corps in 1944. I was sent to college instead of going to Flight school, where I majored in Mechanical Engineering with a Minor in Aeronautical Engineering.

The War ended before I could have finished the degree program but would have had to stay enlisted for four years of sea duty. So upon returning home in 1947, I was able to land a job as a Tool Designer at the firm Modern Design & Engineering, designing tools on contract for IBM, Scintilla Magneto, Link Aviation, and others. These tools were mostly Drill jigs, Milling fixtures, punch and die sets for the IBM 407 Printer. This job only paid 75 cents per hour so I applied for a job at IBM. Instead of a job as a Tool Engineer, they offered me a job as a Customer Engineer to be trained in a special new course they wanted to try.

So, in 1948 I joined IBM Endicott. Ten of us were enlisted into training course spending at least one full year on the production line, each of us individually learning how to assemble every product in the GPD line (026, 031, 035, 511, 405, 077, 513, banking machines) and to do the final testing. We had to complete at least five examples of each type. It was a lot of fun except the wiring diagrams were difficult to interpret until one day on final 402 testing on 402, the final tester made a simple comment, perhaps on how they were used, and it all became clear.

 

This kind of experience was quite foreign to me and I had considerable trouble trying to figure out how the machine was being used. Then one day, on the final test line for the 405 my teacher made some casual remark and suddenly every thing fell in to place and I had no more trouble.

The class used to gather in a classroom with student CEs and sing "Ever Onward, IBM" and "Hail to Thomas J, Watson, Leader of the IBM". Really!

I remember one day at one of those meetings when a new Electronic Calculator, the 604 with about 2000 tubes, was demonstrated to us. I was overwhelmed. I asked "where are you ever going to find people to service such a device?" The answer was that we were going to do it! Remember, I had absolutely no experience or knowledge of electronics and I assumed that no one else had, either.

After completion of the CE training in mid 1949, I was sent to Washington DC. I was reasonably happy there, except that I was only making $1.75 an hour. I spent my spare time designing two new product proposals of mine: a Calculating Key punch with relays (add/subtract/divide) and a card interpreter using wheels for printing. I took these designs to my CE manager and asked for an interview with engineering. He resisted, and came up with lots of excuses about why I wouldn’t get an engineering job. I persisted and finally got the interview. I was immediately offered $475 per month, and joined Endicott product development in 1951. My manager couldn’t believe it (and later transferred to the 1401 program).

My first assignment, working for Ben Durfee, a former CE in Russia, to design a modification for the 405 to handle Extended Capacity Cards (ECC). These cards were to use 6-bit binary coding, two characters per column and I think Poughkeepsie was working on keypunches and printers. That project didn’t last too long. I don’t remember why.

Ben had designed relay circuits for the huge sequential calculator (SSEC) computer to add, subtract, multiply and divide in decimal. These circuits involved thousands of relays and he would draw the circuit diagrams with a drop compass, including every contact, in ink with three little circles for every contact point. While I worked for him, I was fortunate to be a student in the first class given in IBM in Boolean algebra for circuit logic design, by one of Claude Shannon’s student Fred Foss. I showed my manager what I was learning and he was amazed.

When the course ended, the Director of Engineering Education, Perry Perrone, asked me to teach an engineering training class for new engineering hires on how to be IBM engineers. Frankly, I was petrified, and begged off. I had no teaching experience, and would freeze up if I ever had to speak to a group of people. But after the first class, I was cured! During this time, I was also asked to teach an IBM 604 Calculator principals and circuits course. For this, I had to plan each lesson and learn the content the night before class. During that time, I was working on the design of a new Interpreter, the 519. I designed the Zone unit for the print mechanism.

 

 

I had to teach the "Switching Circuits for Engineers" class without the aid of a textbook, as there were none at that time. The class included relay and vacuum-tube logic circuit design and accounting machine operation (electrical, mechanical, operations.) One of the students who really liked the class, Mitchel Marcus, did so well that when I went on to do other engineering work, he took over the course. Transistors eventually superceded relays and vacuum tubes and this led Mitch to upgrade the course. As a result of this work, he wrote three editions of his book "Switching Circuits for Engineers" in the early 60’s and 70’s, a standard textbook for years and sold countless copies in 37 nations. [Mitch, an expert in probability and statistics, also corrected an age-old solution to a probably problem in Martin Gardner’s "Colossal Book of Mathematics"]

With a strong desire to design an advanced, cutting-edge system using future technology to expand the frontiers accounting machines, I was hired into Endicott’s Advanced Systems Development Division (ASDD) in 1956. Mitch and I found ourselves working side-by-side there.

For my first project in ASDD, I was the lead architect on VIDOR (Video Document Reader), drawing all the system diagrams. VIDOR was ….. plugboard controlled.

Next I worked on the Transistorized Accounting Machine (TAM), doing flow circuit diagrams.

During this time, I became acquainted with two brilliant mathematicians in ASDD, Nelson and Armstrong, from whom Mitch and I learned a lot of esoteric mathematics.

Von Neumann came as a consultant to answer questions, deep mathematical problems, he would sit back think he was sleeping, spring up and this is the way it is. Working on system sorting records on tapes (patent) 4 or 8 ASDD two mathematicians scheme for sorting tape records

Attended meetings in Poughkeepsie to discuss the WWAM.

 

Ralph Mork appointed to head of accounting machines, recruited me first to come up with a new design for an accounting machine. had little contact with him, he would invent scenarios about what was going on – but acted as wall to allow me to do my own thing. Looked at WWAM, terrible expense of plugboards, cost of interface electronics. WWAM folks were nitpicking on this transistor vs this one. Idea for word marks came from Ed Grenchus working on some kind of deadend project. My objective was to make a stored program computer that was mine, single-minded. Val Adams, designed 602A San Jose

 

After that I was asked by Perry to teach the Circuit design course and then a Computer Architecture course, including princpals, binary addressing, opcodes. At some time about then, I suppose the WWAM project was initiated overseas, but I was not part of this effort. I was working in the Advanced Systems Development Department on unrelated special systems. [WHICH ONES? MENTION MTing w/ vonNeumann]

IBM maintained two floors in the Beckman Towers in NYC to house students of the Systems Engineering Course. The school was held in a building directly across the street from the UN building. I was fortunate to be able to attend this 3-months course twice. The last time, to fulfill a requirement for a "term paper", I wrote and demonstrated a program which turned the 1401 into a Turing machine. The printer provided a graphic display of the tape. Fascinating to watch.

[1401 STORY!]

When Branscomb/Ingram/xx. Bob Evans was the quintessential manager. He held weekly meetings with about 25 people, and he would always come prepared with about 100-150 items on his agenda. He demanded answers and action from everybody. Impressive to watch.

You ask me how I selected variable length instructions and variable word length. Back then, when the cost of memory was really significant, the choice was obvious. Today, the cost of memory is zilch, and it is hard to imagine the pressure to use as little as possible, back then. I was not inspired by WWAM or the 704 or anything else It was so obvious. I may have studied the plugboard concept and tried a design, but subsequently decided that this was too costly and very limited in function. The market we were after was Alphanumeric oriented. Binary was an exotic, foreign concept to these people Again, obvious choice. However, if high-level languages and compilers were as common then as now, we might have considered Binary.

You can read my interview of 1968 to find out the particulars of the 1401 development. I should probably mention that I was very naïve about large corporate business, totally unaware of the political aspects of product development. Looking back, there must have been a lot of politics at play in the WWAM program. For me, it was like being in a candy store. I could just do my thing, creating and innovating with little direction and interference.

I should have said that I was not management material. I had no training whatsoever, in spite of the fact that management training was an absolute requirement. Somebody made a mistake and gave me the management position, supposedly as some kind of reward, recognition or promotion.

Big mistake!

 

When the 1401 was announced and my part was over, I was fortunate to be transferred to San Jose to work on Process controllers. The result was the 1800. During this program, I had a chance to work with Arnold Spielberg, the father of Steven Spielberg.

Re: 1800. Arnold Spielberg was not an Engineer on the program. I think he worked in Marketing or some such. I consulted with him. The Engineer before I came to San Jose was Chuck Propster. He was very knowledgeable and laid considerable groundwork before I got there.

 

A little side note here: When I was in Endicott, IBM provided certain special features for some machines. Some of these were devices that mounted in 19-inch relay racks and could provide some kind of low-performance process control. One of these boxes was about four inches high, with a switch and a light on the front panel, and with a three-wire cable to a plug on the back. This unit was sold by IBM for $2500! This is an example of IBM’s pricing formula which was so devastating to the development of Personal Computers. After the 1800, I worked in ASDD in the Los Gatos lab developing Educational Systems. This is where Bob Treseder and I invented the floppy disk. We knew nothing of Sony’s prior effort. We incorporated it into an Audio/visual unit that would sit on a student’s desk.

The disk had a capacity of 32K bytes and was loaded with a program to execute APL, a programming language conceived by Ken Iverson. It was quite popular at that time. In a way, it looked and performed like a PC but with a Selectric typewriter for key entry and printed display. The performance was abysmal and the cost was high, but we had the glimmer of an idea and wanted to pursue it further. However, Educational Systems had another project that usurped our funding. Did I mention that Alan Shugart was the manager of the program? IBM had taken our floppy disk drive idea and given it to the Disk drive Development department. There, the idea of enclosing the tape platter in a paper envelope was born. When this was being readied for production, apparently somebody in management realized that it could be used as a very effective I/O device, cheap, portable and fast, and this could spell doom for IBMs multi-billion dollar card business. The program was killed, and a letter was sent to all that this device was never to be connected to any IBM product. Imagine that!

Alan Shugart, seeing a great opportunity, left IBM, probably with a good deal of the documentation, and started Shugart Associates. I don’t believe that IBM thought it was worthwhile to patent this device, so they didn’t.

Another aside: Sometime, around 1960, or maybe later, IBM had developed a new, powerful disk drive called "Merlin". One night, when the one and only model was left unguarded, outside of a loading gate, it disappeared, along with all the documentation. IBM knew who the 12 guys were. (Called the "dirty dozen), and tried to sue. IBM lost the case because of some technical error and because IBM could not show that they had made any attempt to safeguard the Drive. The "dirty dozen" formed a large and very successful disk drive company. This is a widely known story. As a consequence, IBM issued another edict: from that point on, one had to have a "need-to-know" clearance before you could find out what the guy in the next office was doing. Needless to say, this slowed everything down and probably increased costs.

About 1969, I was approached by a small outside startup and offered the job as V.P. of Engineering. I was naïve and vulnerable, and so I quit IBM and joined this firm, IOTRON, in 1970. The main business was Back Panel wiring. We had two Gardner-Denvers, run by two very competent young ladies. Unfortunately there was no way that these two girls could produce enough income to support the high salaries of the five principles. This is an interesting story on how not to do business, but I won’t go into more details here. The business failed after two years, so I had to ask IBM for my job back.

When I came back, I was given the job of Technical Liaison on the big Laser Printer development program (3800). After that task was over, I was transferred to Rochester MN to work on a development program, ultimately known as "Triple Play" This was somewhat like a crude PC, and was finally announced. I don’t think it was a particularly successful product, but I lost track of it when the program was transferred to Austin TX.

I then worked on a Communications Product. I didn’t understand my function here at all well and consequently became dissatisfied with the position. In 1979, I was asked to transfer to Austin. That didn’t appeal to me, and since I had felt I had enough after 32 years with Big Blue, I decided that it was time to retire. I moved to Auburn CA and started my own Engineering consulting business. All during my IBM career, I keep busy on the side doing Tool design and Machine design, for friends who had machine shops and small manufacturing businesses, so I was intimately familiar with the world of links and levers and nuts and bolts and screws and cams and kinematics, etc. So, I was well prepared to design specialized mechanisms. One of my clients built scrap metal recovery plants. I did the design for about 15 plants that were installed around the world.

When this client moved his facilities to Nevada for tax relief, I found another client who was building simple Disk drive testers. I designed some more sophisticated products for him and wrote a couple of million lines of machine code to operate them.

Then, in 1988 I saw a want ad for a structural designer. I couldn’t really call myself one, but I was curious, so I answered the ad. I turns out that the company, Cable Transport Engineering Company (CTEC) designed and built Ski lifts. In early 1990, they merged with a European Company, and then merged again with an Austrian firm, Doppelmyer, the new name being Dopplemyer CTEC.

The owner had previously tried several so-called Engineers fresh out of college with no previous experience. None of them were acceptable, so he thought he would try me (at age 62, ordinarily unemployable). Thank goodness he did. It turned out to be the best job experience I ever had, and he was the best manager I ever had. I spent 15 years there, designing Ski lifts that were installed all over the world. I designed terminals, sheave trains, tower tops and chairs for both regular and high-speed lifts, installed in CA, UT, NY, MN, Canada, Chile and Australia. I was working as the only Engineer at the manufacturing facility in Sacramento CA.

When it was decided to consolidate all functions in Salt Lake City UT, I decided to finally retire one last time. [WHEN?]

I must start with a disclaimer. I don’t remember, reliably, what happened over 50 years ago.

Much of what follows can only be described as logical deduction. For instance, I have been asked how I came to choose core memory over other technologies such as drum or disk. I recall that Ralph Mork was given the responsibility for the development of Electronic Accounting Machines. Before this, he was directing the IBM Military Division in Owego. As I recall, Russ Rowley was working for him, in charge of developing top-secret processors for the Military. Of course, I never knew what the machines were for, but I suspect they were for encryption. I believe that core memory was a central Technology. I also faintly recall talking to Russ. I assume that he had influenced me to give serious consideration to core memory. Once I saw the light, it seemed obvious that a drum would be a bad choice. The 650 had a drum. I observed that when one wanted to program an application, one had to consider timing of the drum. This was difficult and time-consuming. There were a set of instructions that had nothing to do with programming the application, but only to keep track of the drum revolutions and when and where data was under the read/write heads. In my view, this was not the way to go.

It is interesting to note that there was a serious effort to replace the 650 with a higher-performance 660. When the 1401 showed vast improvements over the 660, that program was dropped in a hurry. Again, I only have a blurry recollection of this.

As to the WWAM plug board: it may be difficult for people who have had no experience with plugboards to appreciate the problems we had with them. Not only were they bulky, but they were expensive, error-prone and difficult to wire and store. The customer had to have a wired control panel available for every application they needed to run, plus several empty spares for odd jobs, plus a very large inventory of various length plug wires. It wasn’t just the cost to the machine to have a plugboard (although it was considerable), it was also the cost to the customer to carry the expense of many plugboards. The logical structure of an accounting machine built around the limited functions available through the plugboard was reflected in the limitations of the machine. This was not so obvious until the 1401 came on the scene and showed a clearly better way.

I can no longer remember the sequence of events that occurred between 1950 and 1980. I do know that I taught the IBM Engineering course soon after I left the field and joined the Lab. I also taught other related courses while attending to my regular Engineering duties.