This interview was provided courtesy of the Computer History Museum.
Building a Computer Group
Humphrey: I was trying to build a computer group now.
Booch: Now, what year would this have been?
Humphrey: This would have been in 1956, I think it was.
Booch: Wow, I didn't know the B-1 was in concept way back then. I mean, it takes decades for those kinds of things.
Humphrey: The B-1 was, you know, electronic warfare.
Humphrey: What do you call it? Yeah, it was jamming stuff and everything else, for the B-1 bomber. It was actually being built then. That's what my brother was working on. He'd tell me these wonderful stories. I remember he was telling me about it, and they were testing it, and these people coming in for a demo -- the management team, not outsiders -- and they rigged up a little tube that would run through the machine and he would take a big puff of cigarette smoke and puff it through this tube and a puff of smoke ends up coming out of the top of the machine, and he said management went wild.
Yeah, they shipped
the first B-1 electronic countermeasures -- that's what it was: ECM; they
shipped the first unit down to
So one day a guy came
to see me, a marketing guy, and he said, "Hey, we got this bid I heard
about, that the Navy and the
So he and I raced
Booch: Oh, my goodness.
Humphrey: We won the MOBIDIC and
UDOFT contracts. As I say, the other bidders included RCA and IBM and Philco, and so I was kind of staggered that we won this. And
all of a sudden, here's this little group, we had two big computers to build,
so we did do a mite of recruiting pretty quick, and we got a good team together,
and they did a marvelous job. Delivered them both, got them both working, we
had them up on the floor. So I learned a lot from computers. That's why I
didn't learn to program, because I was the architect on both machines, and the
architecture of the UDOFT was already designed. I mean, the
Booch: Any surviving copies of those machines?
Humphrey: The MOBIDIC, I think there is one. I'm not sure where it is. I've given talks and a bunch of papers on it. There's not much on UDOFT. I don't know what happened to UDOFT, I really don't know, but I did get to fly a jet aircraft, but it was in a simulated aircraft on the UDOFT. That was a rather funny experience. I wanted to get in and fly this thing. It was an F-104. They had two airplanes. That was the real test of it, we had to build this machine -- it was a pretty powerful machine. It had a 10 microsecond add time and a 20 microsecond multiply time, which was pretty damn good in those days. It had separate memories for instructions and data.
Booch: How much memory would these things have had?
Humphrey: I think the UDOFT
was like 10,000 to 20,000 words -- it wasn't big. It had, I think, a 20 bit
word. MOBIDIC had a 38 bit word, and I'm trying to remember how big the memory
was, but MOBIDIC had to be designed to be an expandable system. And so it had
to be able to grow, in terms of memory size and stuff, but nothing else. But
there were a bunch of other versions of it, some of which we got, but IBM and
RCA and Philco and others were all still trying to
get some of the fallout of this. It became an enormous program, lots of people
working on it. And it was doing great, and we built several machines for
Booch: Very good. And, of course, back then, programming was still at the assembly language level primarily, was it not?
Humphrey: Right. We were writing assembly language programs and all that sort of thing. One of the computers we got out of this, the fallout of the MOBIDIC project, and there were really two kinds of fallout from that. One was a program-- a machine that had to have multiple computers running off multiple memories with multiple channels because it was supposed to be able to fail soft so you’d be able to have one computer and we didn’t have to program it. We had to figure out how to build it and that turned out to be a bit of a challenge, but we did it and we could do it, so real-time connection-dependent computers would sense and check each other and all that sort of thing. So we actually built machines like that, designed them. They built them after I left. We designed them in 1958, ’57, this sort of timeframe, but we learned a hell of a lot about how to do all kinds of stuff then.
Booch: And would it be fair to characterize it, most of these machines are really not general purpose, but they were specific purpose things then?
Humphrey: Well, the MOBIDIC was a very general-purpose machine. It was built basically like an IBM 704. It was much more powerful than the 704, and it was the most powerful machine available at the time. It was transistorized. It was, I’d say, the first large scale general-purpose transistorized machine. It was built with unique components, though, unique transistors, and I had a very good crew working for me. I hired a PhD who ran our circuit design work and he was very good. We had a whole bunch of folks. It was a marvelous group.
Booch: Were you guys manufacturing your own transistors? I mean, even those things were not common beasts back then.
Humphrey: No, no. We would buy transistors, and we found that commercial ones were available that were able to meet the military specs. The MOBODIC was in a trailer and it had to meet military conditions. We had the same problems that we had with the other one -- to build a very highly reliable computer that could run with shock and vibration and temperature and humidity, all that stuff. And so we had to build some very reliable circuitry and packaging, and so it turned out to be an enormously powerful machine and very effective. So they were basically running without any problems. Those machines were just remarkably reliable.
So we were trying to
get it used on BMEWS and trying to convince the
And I had my MBA background and stuff and I said, “Well, it is going to be about 10 years before you get real money on this because it will take quite a bit of investment.” Not a quick and easy thing and ten years may have been optimistic, but I thought we might be able to. We had the machine design. It did work, and so when I said 10 years, he looked at me and said, “But I retire in five.” <laughter> So I said, “Okay, thank you.” I wrapped up my stuff, left, and went out to call my dad and figure out how I apply for a job at IBM.
Applying to IBM
Booch: So what did you know of IBM at that time? They had been another bidder it sounds like and what was your impression of IBM?
Humphrey: Well, my dad had
been, as I said, on Wall Street, and he was the chief investment advisor and
treasurer for what was called General Reinsurance then. I believe that became
AIG, but I’m not sure. And he made a whole series of recommendations to the
management back in the 1929, ’30 timeframe for the stocks they ought to buy,
one of which was IBM. He said that’s the only one they didn’t accept. They
bought all the others. And he said, “If they had bought that they would have
been worth more than all the others put together.” <laughter>
He didn’t buy IBM stock but he advised several good friends to do so. And so I
had been buying IBM stock all along. I had been putting my time in
Booch: And thanks to Google I see that in December of 1960, IBM stock was selling about $20.00 a share.
Humphrey: In 1953 it was pretty good. <laughter> So I joined Sylvania in ’53 and I started buying IBM stock right away whenever I could, and my wife at one point, Barbara, said, “Why are you spending your time at Sylvania and your money at IBM?” And so that got me thinking; when this lab director said this, I said, “Well, maybe I ought to think about IBM.”
But in any event, I got a hold of my dad and said, “I’m thinking of applying to IBM. What would you do?” And he said, “Well, I’ve got a friend who I think might be helpful. He knows the IBM people pretty well, so let me send a letter. Can you send me a write up on yourself?” So I said, “Sure.” So I sent it in. I’m trying to remember his name. I’ve gone blank but it was -- yeah, my dad said he was a wonderful guy. In any event I’ll remember his name at some point here soon. But my dad sent him the resume.
Booch: Now, what year would this have been?
Humphrey: This was in early ’59.
Booch: ’59, okay. And IBM at that time according to my calculations had about 94,000 people in it, so it was a much smaller company back then.
The Watson Interview
Humphrey: Yeah, it was still
pretty big. I mean, it was a lot bigger than
Booch: Tom Watson, Sr.
Booch: Oh, junior? Okay.
Humphrey: This is ’59. Senior
died in ’56. So this was Tom Watson, Jr. who was then president and CEO of IBM.
I said, “Well, okay.” I didn’t know what was going on, so I went down to
But he was talking about what they were going to do, and he asked me a whole lot of random questions. He was a very nice essentially -- oh, it was Fales -- Fales was the man that my dad sent my resume to. It turned out Herbert Fales was one of Tom Watson’s boyhood heroes. Fales was an adventurous pilot. He’d done all kinds of stuff and he had for years just been buying IBM stock because of my dad, and so he was a big stock holder of IBM and they all knew him well. And so when he sent this thing in, Tom Watson paid attention, and so he had a real in with Tom. I didn’t know that but in any event so it was an extraordinary bit of luck. So Tom had one of his assistants take me to see Jerry Haddad, the President of their Advanced Systems Development Division.
The Advanced Systems Development Division was a new division they’d just set up, and so he introduced me to the president of that division and I went up there and talked with them. And at the end of the day the personnel people got a hold of me and said, “What would you like as an offer?” And I said, “Good question, but I’m not sure I ought to tell you.” I said, “Why don’t you tell me what you think, and I think that’s the best way for me to understand what you think I’m worth.” And so I got a very good offer, and I went there posthaste and so that’s where I was -- I learned a lot but I’ll tell you. I had published a book and I had taught the courses and I done an amazing amount of stuff for a young guy. This was ’59. I was 32 at that point, and I’d just been having fun, but it was a great time, so that’s where I got.
Booch: So that’s how you came into IBM and so you relocated then to where?
Humphrey: We moved -- we were
Booch: I got to ask if you don’t mind me asking a personal question. How much did a house there cost in those times?
Humphrey: The first house we bought was the one in Cochituate. It was a Martin Cerel house. It was a whole new development going in. It was a marvelously built foundation, plaster walls. I mean, you don’t get that much these days. It was a beautiful house. We put $100 down and got it -- I think the price was around, I don’t know, $20,000, something like that. And we got it out with a 4 ¼% GI mortgage and everything. We put it on the market and we cleaned it up every way, which way, didn’t have a speck of dirt on it.
With four kids, it was, kind of surprising. We patched the wallpaper anywhere where things were marked, so the house was in pristine shape. I’d redone the basement and planted a mess of trees, built a terrace and screened porch and it was a lovely looking house. I must have planted 50 trees in that place, and so we put it on the market and that day we sold it for our asking price. We had somebody across the street-- it was a real estate agent across the street debating whether to ask us -- offer us more than we were asking.
So it went like a shot. We made a little money on it -- not a lot -- and in Chappaqua we bought a house again. The house was kind of rundown, but it was a lovely, basic house on a very big hunk of lot -- hunk of land. Yeah, it was right across the street from where Bill Clinton’s living. It was in a reasonably nice area.
Booch: I guess so.
Humphrey: Yeah, and we bought
that house -- I think we paid $42,000.00 for that and did a lot of work on
that. We later sold it for a bit of a profit. We kept outgrowing our houses. We
needed a house that had six bedrooms, so we got that built and we moved over to
that in ‘63 when our sixth child was born. That was too much to get into our
old house. And so we built a new house there and we moved into that. And we
finally got into our third house in Chappaqua, which was bigger still, a big
old castle, and we did pretty well on each of these. We sold them for pretty
much what we had paid for them and then we moved around again. I moved up to Endwell,
and they moved me up as a lab director up in
Booch: Well, didn’t they say that IBM stands I’ve Been Moved? So--
Humphrey: Most of the moves we did just because we grew. You’re figuring I had -- basically two moves were caused by IBM but the rest of them weren’t.