return to main page

Fran Underwood Interview June 25, 2008
During 1401 Restoration "All Hands" Meeting

Fran Underwood was the designer of the 1401 -

Source was audio file FranU...08-all.mp3 from Robert Garner

Self introductions by audience near beginning seems to have been edited out -
Picture of group

July 20, 2008 - Notes to this transcription are at the bottom: Links are shown as superscripts, this being an example(1).
Links to these superscripts are: 1, 2, 3.
The previously unidentified speaker is Dave Cortesi.

Key to speakers:

Fran U = Fran Underwood
Robert G = Robert Garner
Shel J = Shel Jacobs
Allen P = Allen Palmer 		Chris R = Chris Reid
Dave C = Dave Cortesi
Ed T = Ed Thelen
aud = general audience

Fran U Hello Robert -
Robert G Hello Fran -
I can hear you, can you hear me?
Fran U I can just make you out.
Robert G Is that better?
Fran U that's better
Robert G I hope this speaker phone can work better with the rest of us so Your talking to a room full of about 20 people
Fran U Oh my
Robert G Yeah - this is the 1401 Restoration Team. Sheldon Jacobs is setting about 10 feet to your right
Shel J Hi Fran
Fran U Hello there, how are ya doing?
Shel J Good, and you?
Fran U Oh, not so good
Shel J I just have one question for you - Do you still carry a plastic pocket protector?
Fran U No, I never did
aud (laughter)
Shel J Ah, come on, I talked you out of it -
aud (more laughter)
Fran U i took (????? noise)
aud (more laughter)
Robert G Oh, dear -
OK, so Fran what I thought would be a little fun would be just introduce your self and its OK if you just ramble along, we are not in a hurry or anything - ... just go as slow as you like -
Fran U want me to get started now?
Robert G sure
Fran U All right
Fran U from when I was just starting out?
Robert G sure! why not -
Fran U well, First of all I want to set the stage here,
I had a stroke (noise) three years ago -

It truly paralyzed my left side. With therapy I've recovered a lot of the mobility

- but subsequently I've had three more very small strokes. The last one affected my throat muscles - I have a tough time talking.

Robert G Understood
Fran U Bear with me -
Robert G Understood, it sounds fine to us though -
Fran U Umm - You know I joined IBM around 1948. Previous to that I'd been a tool designer - working for a small design company and we contracted tool design for IBM. So most of my tool design experience was designing most of the tooling for the 407 printer. Doing the tool jigs, assembly fixtures, and milling fixtures and so on - And I thought that since I wasn't getting paid very much I might as well go to the source and see if I could increase my wages.

So I applied for a job at IBM as a tool engineer. Well, it just so happened that at that time IBM was contemplating ?starting? a new customer engineering training course. They wanted about ten diverse bodies to attend this course and the course was - don't go to class, don't go to school, but go to the manufacturing site where all the assembly was done. and make machines, do machines, learn how they work talk to the experts on the assembly line and final test, and learn all we could on every machine in the product line. And at the end of that you would be a customer engineer.

I didn't know what a customer engineer was, but it sounded good to me. so I signed up for the course and over the period of a year I build five examples of every machine of the line, and learned all about everything (chuckle ;-)) And it was a really great experience.

I don't know how manufacturing managers felt about this, you know having a guy come in and manufacturing had to teach these guys every thing they knew, then they would get up and walk away and (chuckle) go some place else.

But it all worked out well - After the school was over, I was assigned to the Washington, DC office I had the usual customer engineering tasks to do - servicing all the different equipment. Eventually I was promoted to the refurbishing of accounting machines. I didn't have to go see customers anymore, I just stayed in the work place there in Washington and serviced accounting machines.

All during this time I really wanted to be a design engineer. So I had a couple of ideas, so one idea was - wouldn't it be great to have a calculating key punch that you could punch in the two factor fields and then the machine would automatically punch the answer either a sum or a difference or a product or a [quotient] or a remainder in the card. So at night I would spend my time designing the machine. I was pretty good at circuits by this time. When I started the customer engineering course I couldn't get my way through a simple simple two relay circuit, I just didn't know how to do it.

But at the end of the course I was able to almost memorize the 402 wiring diagram and fix the darn things almost without reference to the wiring diagram. Got to be pretty good at that. I also had another idea during this time. I never did like the interpreter, was it the 511 or something like that? 512, I don't know. The interpreter printed right on the card, so I designed another interpreter that used print wheels like the 407. And I designed the entire thing, and I had all these drawings, of these two machines.

I went to my CE supervisor and asked him to get me an interview with engineering. But he said "Oh no, I can't do that - they wouldn't be interested."

Well it just so happened -later - that IBM was increasing their effort in their engineering lab and needed new talent. So they sent interviewers out to the local offices to interview. Well, I managed to get an interview and (chuckle) it was only about three or four days later I got invited to the engineering lab for further interviews and then I came back to Washington.

And my supervisor had to come to me and say well, (chuckle) you've been offered a job in Endicott". And he was overwhelmed at the money that they were offering me. Really disappointed to lose me.

So I went and had a grand time there . First project was designing an increased capacity card line of equipment particularly the accounting machine. I used the 402 as the basis and designed new counters and new card feeds and I don't know what all - circuits. It would read this hundred sixty column card. I don't know if you guys have any idea what that was but 12 rows of punches in a card can be divided into two six bit rows and we used binary encoding in each column so we that had a hundred sixty column card. And the guys at Poughkeepsie were designing the keypunches.

Well, I worked on that for a while and it finally died for some reason, I don't what the reason was. Then, some time during this period, the director of education, I think Perry Pronie, got a hold of me and asked me to teach the engineering ???turn??? course. This is the course where we took a variety of engineers of various types and we hired these people and needed to train these people to be *IBM* engineers which is a different kind of engineering, which is a different kind of engineering than is done in the rest of the world.

So I said "No, I can't do that - First of all, I absolutely would freeze up in front of an audience - I just couldn't talk and I get so nervous and shaky". And he says "Well, don't worry about that", so he finally talked me into doing it. And, son-of-a-gun, I lost all of that fear teaching that course and I taught all of these guys how all of our machines worked - how relays work and what relay circuits are and how they work and how they are designed and so on.

Then IBM hired a guy fresh out of MIT and he was an expert in switching circuits so he came to IBM Endicott and taught one course in switching circuits. Then he was re-assigned so I volunteered to teach that course too. So I went on to teach several courses in circuits design and also computer architecture. So I had a great time teaching - in addition to my regular job. One of those jobs was working on the design of the 557 interpreter and I designed particularly the ?zone? unit.

So - after that I was assigned to the Advanced Systems Development Department and worked there on various proposals - primarily doing the systems design, drawing up all the flow diagrams and the circuit diagrams. I was pretty good at that.

And then came the World Wide Accounting Machine and that is a whole nuther story. But I did some preliminary work in the Advanced Systems Development Division and when Ralph Mork was assigned to be the Electronics Accounting Machine manager he picked me to do the design work.

I don't know if he picked Tim Engram and Russ Rowley, but anyway eventually, we were in the same department all though I had little contact with them. I think Russ had something to do with the core memory. Tim/Jim Engram was the circuits man, he really knew circuits very well. And after the - oh, we had lots of fights about which system was the best - mine, or one proposed in San Jose or the World Wide Accounting Machine proposed in Germany and France - we had a go-around.

But eventually my system was picked and they finally headed for production. And it was somewhere along here that Shel Jacobs showed up and poured new life into the program and really made it go. I'll never forget his contributions, they were really great.

Robert G Hey Fran
Fran U Speak up -
Robert G You had that story about T.J. Watson [Jr.] come into your office.
Fran U Oh yeah - well Shel Jacobs was a great proponent of the system, and he insisted that they were going to sell, I don't know, four or five thousand units but local gurus who made predictions like this officially thought that we were going to sell, I don't know, two hundred, two hundred fifty or something.

Shel said "No, No, a whole lot more than that" Well, Watson heard about this - Tom Watson Jr. - and as we got nearer and nearer to the announcement, he began to get worried about the discrepancy in the numbers.

So one day he showed up at my office and sat down and said "Now, tell me about this program." - And so I did, and then he said "Do you really think we are going to sell thousands of these things?" and I says "Oh Yeah, no doubt about it." then he says "Well, OK" and he accepted my word, and we had our program.

And it really did, it went very well - we had Chuck Branscom who was a great manager, and then Bo Evans who was supreme, he was an incredible manager. And we had Shel Jacobs who was great with the marketing end of it. And we were just rolling, tooling along just fine -

Robert G Hey Fran, do you remember how you selected the 1403?
Fran U No I don't
Robert G That just came into the scene?
Fran U That could have very well have been - Branscom and Jim Engram might have been instrumental in getting that evaluated - and become part of the program. *Although*, there was one other guy - I think his name was George - he was a sandy haired Scotsman, and I think he worked in marketing -
Shel J Are you thinking of George Barntique?
Fran U That's the guy! -
Shel J ? Hallelujah
aud (laughter)
Fran U George Barntique - oh - well - he saw the value of the 1401 system as an off-line printer system for the big systems. And he did the marketing work on that aspect of it -

George Barntique - goll-darn - he still around?

Shel J I've have no idea
Fran U OK
Robert G So Fran
Fran U Yeah
Robert G Reading a lot of the history, and a lot of people don't know, I was able to find copies of Fran's original design memos back in the IBM archives and I've been sending them to him.

But one of the original points of contention with the World Wide Accounting Machine was the cost of the plug-board. Did you do the analysis on that?

Fran U I think I must have - 'cause I was just appalled at the costs of a plug-board. Just to implement it - The attendant costs to the customer who had to put up a board for every application - there was no one board for every application - he had to plug a board for everything you wanted to do and the cost was enormous - Its just not the way to go -
Robert G and part of the problem was the transistors required to protect the signals - right?
Fran U well yeah - the transistors and diodes were really a serious problem -
Robert G yeah - The target rental for the machine was something like $1,500 per month or $2,000
Shel J for which machine?
Robert G the 1400
Shel J $2,500
Robert G the European World Wide Accounting Machine was coming in at about $5,000/month
Fran U that is correct -
Robert G so they were really off in left field - so the irony is that all those guys who worked on it - some of which are still alive today - like Maurice Papo and Karl Ganzhorn - they all love the 1401 because they think you took their ideas like two field addressing and core memory(1).
Fran U Well , It's hard to really say - I didn't work in isolation - I worked with those guys and knew what they were doing - I knew what was good and what wasn't good - but I
Robert G You had that sense of American "can do" attitude and maybe that what was missing?
Fran U yeah - ah - but ultimately I did it my way - you know - Maybe a different form of instructions, well I went for an alphanumeric machine rather than the binary - and that was new and unique - and this question of the edit function - sure, they had an edit function and it was used for check protection - you know zero fill and all that - but it was designed to work with the control panel - so an entirely different method or scheme was required to do it totally by stored program. And that was my contribution -
Robert G you called the project SPACE right? Was that inspired by the Sputnik program?
Fran U Yes, it was
Robert G and what was that other name for the project too - Roman Letters - I e-mailed to you - do you remember what that was?
Fran U Oh yes - Stored Program MBM Magnetic Core Memory and then DLH - VLI for variable length instruction, and then VLW for variable Word Length and then ACM for Accounting Machine(2)
Robert G yeh thats it
Fran U that was before WWAM [World Wide Accounting Machine] and that was the name I stuck on it when I was in Advanced Systems Development - for project number 4740
Robert G yeah - and do you recall how you first learned about core memory? was it maybe that person in the military group?
Fran U that was general knowledge in engineering -
Robert G Does anyone have any questions for Fran?
Dave C Well, I never programmed the 1401, but I'm just curious about what seems to be its most unique programming feature is the use of the word mark - wasting a whole bit out of every memory position for a terminator, how did that come about?
Fran U Well, I liked it (chuckle)(3)
aud (much laughter)
Fran U It appears to be wasteful - but - in retrospect it appears to be wasteful - but I still think it was the right thing to do -
Robert G One thing Fran, looking at the old documents and reading the history, is when you designed the machine you were not thinking of tape drives on it at all.
Fran U No - that was George Bonaque's act - contribution - he saw that hooking up the tapes to the printer all under control of a 1401 was a great way to have enough line printer for the big systems.
Robert G I think a lot of people bought 1401s with tape drives and started copying all their cards to tape and began the transition away from punched card. That was not a planned event but that what the effect was -
Fran U Yes - IBM didn't want to get rid of cards. When I was in San Jose , another engineer Bob ?Trachedure? came up with the idea of a floppy disk drive - and we developed it to a point and IBM didn't want it - so my boss at the time, Al Shuggart, saw the value of the thing and he ran off with it
aud (some laughter)
Fran U Started Shuggart Associates - and made a very successful, as you know, floppy disk drive. And IBM issued an edict at the time that this device will never appear on any IBM product. Well - they wanted to protect their card business - Multi-million dollar business - floppy drives just wiped out that whole thing you know - and that is what eventually what happened.
Robert G Then you worked on this 1800.
Fran U yeah - I went out to San Jose to work on that thing - I don't remember much about it I drew a lot of system diagrams trying to find the right way to handle everything that was required - but really (cuckle) I don't remember much about it.
Robert G do you remember - I haven't asked you this question before Fran - Do you remember how you guys were so effective in shutting down the 305 RAMAC follow-on and IBM chose the 1401 instead? Was it because it had a plug-board
Fran U Oh Yeah - the RAMAC group was out in San Jose and they also proposed a disk based World Wide Accounting Machine - but - when the time came to compare the costs and ability of the two systems, they lost by a long shot -
Robert G OK - so they were more expensive basically
Fran U oh yeah - more expensive and not as proficient - they also had a control panel -
Robert G and then you guys basically offered the disc drive on the 1401 your selves so you basically killed the project and took the disk - (Chuckle)
Fran U Yes
Robert G Well done -
Fran U but putting the disc drive on the 1401 is not the same as basing the whole process on the disk drive.
Allen P Fran, Allen Palmer - the question - did you stay with the projects when they went to the 1440 and 1460? And then after a while -
Fran U No I didn't - I did not stay with the project - after the announcement You know right now - I can't think what happened
Allen P What about all the features that were eventually added like multiply, divide, switching the tapes between two CPUs, and everything else, was that still under your domain at that time?
Robert G Did that happen while you were in Endicott?
Fran U Oh Yes!
Robert G I think he did because one the memos he wrote outlined the multiply feature - Fran did design that because I found a memo where he discussed it.
Ed T Overlap?
Robert G Oh yeah - Process Overlap - I think you were gone by the time Process Overlap was added to the 1401
Fran U No
Fran U I don't know - I transferred to San Jose in '62 or '64 and I don't remember when -
Robert G That's OK - it turns out that Process Overlay is this feature that IBM added to the 1401 which allowed it to do cards and tape at the same time and greatly complicated the logic design - its a massive EC to the machine that CEs would take days to implement - and its been a big problem for us in our German machine because are failures in the diodes and transistors this process overlap stuff just infects the entire machine so they really corrupted your original design - that's for sure.
Chris R My name is Chris, and I remember us getting a console for our 1401 - but I don't remember the circumstances that allowed you to have a console.
Fran U I can't remember either -
Robert G it may have been after he left - the console
Fran U You know, 4,000 bytes of memory - (chuckle) - is not very much - There was a definite demand for more memory.
Robert G A lot of people here don't realize - In his original design memos he only designed for 4 K - that was the maximum and - Did the 12 K add-on, was that done while you were there or after you left?
Fran U No, it was done while I was there - or after you had left?
Robert G The other question is you designed, I think, the very famous front diagnostic panel - the switches and lights
Fran U Yes -
Robert G What made you decide to show the data path of the machine on the front panel?
Fran U (chuckle) I liked it -
aud (much laughter)
CR? That was a great/ fantastic -
Fran U It made a lot of sense to me - it was a simple data flow - the way the characters flow around the machine - it made a lot of sense to me - I had a lot of fun with the industrial design people - talking about the dials for setting the addresses - whether they should be rotary switches or slide switches - or turn right or left - or whether they should point to the numbers or if the numbers should be on the dial - they went on and on about this sort of thing -
Robert G isn't that amazing
Fran U yeh - we don't worry about that any more -
Robert G yeh - we just play video games and just answer our cell phones these days -
Fran U push buttons -
Allen P Fran, a question - did the final cost of the machine come in at budget, under or over? You projected the cost to make the machine -
Fran U well it was within 3 % - I was totally amazed by that - but we did a pretty thorough job you know - it came out within 3 % - probably over -
Robert G were you lucky or where you really that good?
Fran U OH, I was GREAT -
aud (much chuckling)
Robert G the more we talk about it - sometimes the more memories come out. OK, Any other questions from the room to ask Fran before -
Allen P Did Watson come back into your office and say you were right? Did you ever see him again?
Robert G I think that you said in your oral history that IBM booked 5,000 systems in the first five weeks - which in today's dollars would be - if they were purchased - would be two to three billion dollars in just five weeks - which is remarkable -

The question was "Did T.J. Watson ever come back into your office and congratulate you for being right - or did you ever see him again?

Fran U No - never did - he was too busy counting his shekels, I think -
Robert G If there are no other questions - we have you number and e-mail address - so
Fran U Have Shel Jacobs come up to the phone - I'd like to talk with him for a minute -
Robert G He's going to come up right now -
Shel J I'm here Fran -
Fran U Hey Shel - how ya doing?
Shel J Absolutely great, you?
Fran U Wonderful - are you living in Saratoga now?
Shel J Still living there, haven't moved in 42 years -
Fran U WOW
Shel J Still in the same place
Fran U oh - I've moved all over the country
Shel J Hey - I'm under the impression that when you retired, you built an airplane -
Fran U Yes I did, but I sold it before I finished it -
Shel J did it fly?
Fran U I sold it before I finished it - and lost contact with the guy that bought it - I don't know
Shel J You know, as you talk you lit up more bulbs - you know they asked me to talk for a half hour - I'm going to schedule three days -
aud (much chuckling)
Shel J there is so much that you reminded me of -
Fran U oh good
Shel J they were great great days !

I don't know about the rest of the guys here, but I think we contributed a lot to our pension plan!

aud (more laughter)
Shel J Great hearing from you Fran - I hope things go well for you
Fran U I'm glad that Robert finally found you -
Shel J Oh Yeh - he found me and I think he's going to try to milk me dry -
Fran U and he should, you are a fountain of knowledge
Shel J Thank you Fran, I appreciate that -
Fran U Well - I appreciate all you did on the program, wonderful -
Shel J Thanks - good talkin to you
Robert G We are going to hang up now

- thank you so much for your time

- we will stay in touch - Good Bye
aud General Applause

Addendum notes to Fran's interview - linked from above -

  1. In Robert's question about possible influence of the WWAM, the WWAM design features that were purported to have had an influence on the 1401 were: "2-address instructions and the EDIT function" (not "core memory" as voiced).

  2. The long-winded acronym that Fran coined in 1957 before switching to the shorter, Sputnik-inspired code name SPACE (Stored Program Accounting and Calculating Equipment") was SPMCMVWLVILCAM, or "Stored Program Machine, Core Memory, Variable-Word Length, Variable-Instruction Length Calculating Accounting Machine."

  3. Fran sent me an email saying that this is a better response to Dave Cortesi's question about the origin of word marks:
    On “Todays’ talk”, I was asked why I chose ‘word marks’ as the means to define fields and instructions. I was not prepared to answer this question at that time. All I could think of was to give a flip answer and say “Well, I liked them.” Not a good answer. It is a good question and deserves a more thoughtful answer. I have given the matter some thought and this is what I have come up with.

    At the time I was putting the Architecture of the 1401 in place, I was working in ASDD. We worked with Marketing, trying to figure out what applications were going to be important in the next five-to-ten years, what technology might be available, and how best to match the two. Much of what went into the 1401 was known by the computing community: stored programming fixed and variable word lengths, various circuit families, etc. We did not anticipate integrated circuits, PCs’ etc.

    I had a coworker in the department named Ed Grenchus. He also was a systems designer/architect. He dreamed up the concept of ‘word marks’ for a product that never made it. I never claimed to have invented word marks; I only made a practical implementation of the concept.

    It has been noted that 1/8 (12.5%) of the very expensive core memory was devoted to word marks. How can this be justified? Well, consider the alternatives:

    1. Fixed word length. Accounting machine applications are, by nature, variable length. I could see that much space would potentially be wasted if a fixed length was architected. Name and address fields were typically 20-30 characters long, numeric fields were often eight characters long and control codes typically were a single character. Inevitably, there was going to be more than 12.5% wasted space in memory.

    2. Field separators. Punched cards typically had contiguous fields. Field definition was achieved though Control Panel plugging. The 1401 was NOT going to have to carry the burden of plugboards. The content of a punched card could not just be dumped into memory without field demarcation. Also, there would have to be some means for writing and erasing field separators. I judged that handling word marks would be easier.

    3. Some other method. Well, I didn’t find one.