Notice: We are in the process of migrating Oral History Interview metadata to this new version of our website.
During this migration, the following fields associated with interviews may be incomplete: Institutions, Additional Persons, and Subjects. Our Browse Subjects feature is also affected by this migration.
We encourage researchers to utilize the full-text search on this page to navigate our oral histories or to use our catalog to locate oral history interviews by keyword.
Please contact [email protected] with any feedback.
This transcript may not be quoted, reproduced or redistributed in whole or in part by any means except with the written permission of the American Institute of Physics.
This transcript is based on a tape-recorded interview deposited at the Center for History of Physics of the American Institute of Physics. The AIP's interviews have generally been transcribed from tape, edited by the interviewer for clarity, and then further edited by the interviewee. If this interview is important to you, you should consult earlier versions of the transcript or listen to the original tape. For many interviews, the AIP retains substantial files with further information about the interviewee and the interview itself. Please contact us for information about accessing these materials.
Please bear in mind that: 1) This material is a transcript of the spoken word rather than a literary product; 2) An interview must be read with the awareness that different people's memories about an event will often differ, and that memories can change with time for many reasons including subsequent experiences, interactions with others, and one's feelings about an event. Disclaimer: This transcript was scanned from a typescript, introducing occasional spelling errors. The original typescript is available.
In footnotes or endnotes please cite AIP interviews like this:
Interview of Ralph Gomory by William Thomas on 2010 July 19, Niels Bohr Library & Archives, American Institute of Physics, College Park, MD USA, www.aip.org/history-programs/niels-bohr-library/oral-histories/33902
For multiple citations, "AIP" is the preferred abbreviation for the location.
This interview covers Ralph Gomory's family and upbringing, undergraduate education at Williams College, and graduate education in mathematics at Princeton University under Solomon Lefschetz. There is discussion of Gomory's time serving in the Navy, working at the Office of Naval Research, and exposure to operations research. The interview then turns to Gomory's return to Princeton, and his work on integer programming, and subsequent hiring by IBM. The interview covers Gomory's time as part of the Mathematics Department with Herman Goldstine and Benoit Mandelbrot, and under Director of Research Emanuel Piore, and also Gomory's own tenure as Director of Research from 1970 to 1989. There is some discussion of the staff he oversaw in this position, and some challenges faced by IBM Research in this period, as well as about his public writing on subjects such as the Japanese challenge to American economic productivity.
This is Will Thomas, and I am with Ralph Gomory at the Sloan Foundation in New York City. The date is July 19, 2010, and we are doing a life history interview in conjunction with the [AIP] History of Physics in Industry project, but we will be discussing other topics as well. So why don’t we begin with your family background and just move straight up through your education from there.
Well, my father was born in Szeged Hungary, and he came to the United States after going to college in Germany. He went to the University of Berlin, and his first job after that was, I think it was, with something called the Trans-Atlantic Trust Company, and I think they offered him at some point the choice of Istanbul or New York, and he chose New York, and then he never went back.
What was his name?
Andrew Lewis Gomory. Eventually he became a very senior person at the Manufacturer’s Trust Company, which eventually became Manufacturers Hanover, which is now something else. So I was born and brought up in Brooklyn on Brooklyn Heights, which is just right across the river from Wall Street. It’s a very nice community. My mother’s family had lived in Brooklyn — they were reasonably well to do — for many years, and mother, born Marian Schellenberg went to Packer Collegiate Institute, which was just down a few blocks actually from Brooklyn Heights. So I was brought up there. And I always went to private schools, and I went to a really good private school called the Woodward School, which was really the only school I ever liked [chuckles].
I was reading one of your articles in which you had strong words about most of the schools that you went to.
Yeah. I’m not sure I want to repeat that! [Chuckles] But the Woodward school I think was a great… I think if it weren’t for the Woodward School, I’m not sure that I would have ever thought that learning and all that could be any fun. But they gave us great liberty, and encouraged us to do things. It was just a wonderful place. Other schools seemed after that very confining and narrow.
So you mentioned when we were at lunch that you had a History of the World in Ten Volumes. Was there anything in addition to that that your parents provided?
We had The Book of Knowledge as well. That was another series which everyone knew about at that time; it may not be familiar to you. It’s just full of how things are made, all sorts of real events. The loss of the Birkenhead — well that was the British ship that went down in some storm or hit a rock, but they all behaved very gallantly. And so there were many heroic stories… So I had a lot of good stuff to read.
So your parents were very encouraging of education?
Yes, I think so. They didn’t have an academic bias. I mean they weren’t training me to be an academic, but yes, they just thought this was part of the way you bring up children. You see what I mean, yes. And I went to camp from a very early age in the summers. In those days, of course there was no air conditioning, so there was [laughs] a tendency for the wives and children to leave the city one way or another. So I was sent to camp from a very early age, and went to a variety of camps starting at age 4 which was perhaps too early. However I was very fortunate, when I was about 8 I ended up with a camp that was really the equivalent of the Woodward School, it was a great camp, Wentworth Camp. It was up on Lake Wentworth. Lake Wentworth is an arm of Lake Winnipesaukee, and that’s where we were.
So in terms of it being like the Woodward School, I don’t imagine camp at all being like a school, so could you elaborate?
Well, you see it really is, at least the way it was organized. I don't know if other camps — well I think all the camps I went to, you basically had a schedule. You got up in the morning. Well, the bugle, we were always gotten up by a bugle, and you had a schedule. Like at 9 o’clock you might have archery, and at 10 o’clock you might have tennis, and at 11 o’clock… And you went to these activities, all the way up to noon. Then the afternoon at Camp Wentworth we went down to the shore of the lake and we had swimming classes, we had some boats, canoes and stuff.
So in the sense of having a schedule.
Yes, we had a schedule.
But it wasn’t an academic camp or anything like that.
Oh no. Such a thing did not exist. All camps were athletic or gardening or something. No, there was no such thing as a science camp at all.
Did you have much of a sense of science and mathematics as being a particular thing to do as you were growing up?
I did read a lot, but I think the thing that got me started in that direction was a very famous book that I think influenced a lot of people. I was called Microbe Hunters by Paul De Kruif. It was a totally inspiring book about the people who… I guess Pasteur was in there. But all these people who conquered disease. And so I immediately resolved to be a research doctor.
About how old were you when you encountered this?
I must have been somewhere in the 8 to 12 range. And that resolve continued until when I was in high school I had to dissect a frog — that ended it.
You did not enjoy the dissection, I take it. [No.] Did you have brothers and sisters at all?
I had an older brother. But his inclinations were very different from mine. So I went to school through 8th grade at the Woodward School, and then I went to something called Poly Prep, which was a very traditional school, which I really didn’t like. And then I went to boarding school, which was sort of an unusual school for its day in the sense that, first of all, it was a Quaker school, George School. Not St. George or anything like that; George School. Quaker school near Philadelphia; it was in Bucks County, Pennsylvania.
So quite a ways away from home, then.
Yes, that’s right. And it was a co-ed boarding school at that time — fairly rare. And that was pretty good. I played soccer there and wrestled in the winter and played tennis in the spring.
Is that sort of at the high school stage then?
That was high school, yes.
I have it down that you were born May 7th, 1929? [Correct.] So we’re in the 1940s now, I would suppose?
Yes. So the war was on. And of course Quakers are pacifists. And I took a physics class at George School, and I just loved it. After a while they said I didn’t have to come to class. For some reason or another, physics just came to me naturally.
You had had mathematics, I take it?
Oh I had some mathematics, yes. I guess I had both plane geometry and algebra. But I hated it. I didn’t get the point. I’m afraid that throughout my life I’ve always been sort of applied in the sense that I wanted to see a connection between what we were learning and something in the outside world.
That there’s a motivation behind what you happen to be doing.
Yeah. It was hard for me to think about it if I couldn’t see the point. Even in math, which I didn’t do particularly well in, once they got to word problems, you know, John can mow the lawn in six hours by himself, and Bill can mow the lawn in seven hours or eight hours. How long does it take them to mow it if they can mow together? I couldn’t miss those. So that’s just my thing.
Did you do well in math, regardless of your liking for it?
Well, it’s hard to say, because I did well in my subjects generally. But I do know that I apparently didn’t have much aptitude for math because we had some guidance counseling, and when I told my guidance counselor that when I went to college I wanted to study physics, he was very dubious because he said I had so little aptitude for math.
But at the same time, the physics course had gone well that you had taken.
Yeah. So I didn’t listen. [Chuckles]. Discussion of the interview schedule.] I think the physics thing was important, and also what they told me about my mathematical ability! [Laughs] Then after that, we went to Williams College. When I say we, actually a group of four of us, very unusual, decided to go to the same school and we got admitted. In those days getting admitted to college was not the struggle it is today. If you could pay for it and you were a good student, you tended to get in. Actually the four of us, who were good friends, one of them was Steve Sondheim, so we were good friends in high school, and to a lesser extent in college. One of the odd things I remember is visiting Oscar Hammerstein’s place, because it turned out that the Hammersteins had a place not far from George School, they had a country place, and Jimmy Hammerstein, who was Oscar’s son, was at George School — a couple of years younger, I think he was, than we were. But at any rate, we were invited over to the Hammersteins one weekend for dinner or lunch or something, and so I met Steve’s mother who was there, and Oscar Hammerstein, who turned out to be a very big man with a very bumpy face, but very kind — seemed like a very kind person. I mention him because I admire enormously the Rogers and Hammerstein musicals, and I think some of the combinations of words and music there are just —, you know you just wonder how they can do it. I don’t easily feel that way about things, but I do about that. So anyway, then we went on to Williams. I really liked the physics a lot, up to a certain point. You see, I really like understanding the world around me, and after a certain point in physics as we were being taught it, it didn’t do that anymore. Instead of explaining why objects around me were falling to the ground, we were getting deeper and deeper inside the atom or inside the nucleus (we hardly got into the nucleus in those days). Well, I didn’t feel that strongly about the atoms, unless I could connect it with something, if you see what I mean.
So is this quantum mechanics that we are talking about?
Yes, I guess it was quantum mechanics. Although I’m not sure. You have to realize that Williams, though it was a very good college, and I didn’t realize that when I went there, it wasn’t like a university. My senior year we were taught advanced calculus. We never had epsilon and delta, at all. And the same was true in physics. So there was really a gap between that and what most people went to graduate school knowing. Now today Williams is a totally different place; it is quite research oriented. At that time, teaching students was what it was about. But I was very fortunate. First I was unfortunate, had a person who taught who was actually an OR person, [Hugh Miser], but he was not a very good teacher for me. Then he left right in the middle of the term or something. He was somehow tied up with OR.
Yes, he would have been with the Air Force, I think, at that time.
I don't know whether he left to go to them or whatever it was. [Ed. Hugh Miser was at Williams College from 1946, when he had received his PhD in Mathematics at Ohio State University, to 1949, at which point he joined the Operations Analysis Branch of the US Air Force.] The person who was the head of the Math Department filled in for him, and his name was Donald Richmond, and he was just great. He was just a very inspiring person for me. So I became interested in mathematics. And because physics didn’t seem to be explaining the world to me anymore, I reached the erroneous conclusion that it was mathematics that explained the world, which I don’t now think is correct but I did at the time. Because I found that the same differential equations that described something in physics also were being used to describe say the beating of the human heart. So I got very interested in ordinary differential equations because of that. I was just fascinated by this paper by van der Pol entitled “The Beating of the Human Heart Described as a Nonlinear Oscillation.” So I thought now I’ve got the real key to understanding things — its mathematics after all! So in my last year I became a math major. Also, passing through there, I took a course in economics from Kermit Gordon, who was a well-known person at the time but died young. I found him very inspiring. I was very torn, because really it was too late for me to be an economics major, but economics came to me very, very naturally, like physics. It just seemed like it was common sense and you hardly needed to be taught it practically. And that does relate to what I do now, you see. So my senior thesis was about van der Pol’s equation, a non-linear differential equation
I noticed that you have a few publications on the van der Pol equation.
Yes. Don Richmond then asked me to stay on for the summer after I graduated, and together we wrote a published paper, [“Boundaries for the Limits Cycle of van der Pol’s Equation.”] That was my first published paper. And I would say it was kind of typical in the sense that I found a simple way to do it — I like simple, I really do! Then I went on over to England, and that was purely accidental.
This was after you graduated?
Yeah, I graduated from Williams, and I spent the summer I graduated  with Don Richmond and his wife. They lived up in Williamstown where Williams was. Then I went over. My mother died after a long illness when I was a freshman, and she was a wonderful person.
I don’t think we got her name earlier.
Marian. She died very young. I owe a lot to her, my personality and everything else. Then I went over to England to spend a year at Cambridge, and that was accidental, really, in the sense that my dad knew the provost of Kings College, Cambridge. How he knew him I don't know; somehow through business. Because my dad was in the international part of banking. He ran the international part of Manufacturers Hanover, Manufacturers Trust and then Manufacturers Hanover. So he was always traveling to Europe, and in fact — I guess this is worth mentioning as well as my interest in history — he thought it was very important to see other counties and especially to know other languages, I think because that had played a large role in his career, because he was born to speak Hungarian, and they all learned German in those days, and then he went to the University of Berlin, learned English, and he spoke pretty good French, and that was a big help to him. So he sent us, my brother and me, in the summers when I was in college — see, my mother died when I was a freshman, so there wasn’t that much to go home to. So I spent a summer in Geneva and a summer in Tours in France. Then I was in England at Cambridge after I got out of Williams, then I spent that academic year in Cambridge. The summer after that academic year I was in Innsbruck in Austria learning German. Cambridge was a great place in many ways, but I found out at Cambridge that mathematics did not do what I wanted it to do… it was like what happened to me in physics. Physics wasn’t about rocks dropping on my head and interesting things like that, when you really got into most of it. Well, the mathematics that I was learning over there was about mathematics, and I liked group theory and things like that, but it was my first exposure to rigor, epsilon and delta, and really proving things. I certainly could learn to prove what a continuous function was, what really made continuity, but that’s not what I wanted from mathematics. I wanted mathematics to increase my ability to understand things, rather than teaching me to prove what I already knew… I felt I knew what a continuous curve was, so what’s the big fuss? And I didn’t fully appreciate the power and usefulness of learning that methodology. So from a mathematical point of view I didn’t enjoy being there. The first term I had a very rigorous kind of tutor, and he didn’t think much of my attitude or my ability, and I didn’t like what he was trying to do. But I was very lucky with my second term, Philip Hall was my tutor. You know, in England the tutor is the main thing at the colleges. You go to the lectures, but any teaching you get is from the tutor. And Philip Hall was wonderful. He was a group theory person, and I liked group — either that or I liked him, I can’t tell which. Anyway, I felt with group theory I did learn something, rather than unlearning the fact that a continuous curve was really continuous. I had a hard time with the lectures. I had trouble paying attention to them. My notes were terrible; I took terrible notes. And so, somewhere along the line, I thought I wanted to go back to the U.S., probably go to MIT, because I thought MIT is more applied and that would probably suit me as a graduate student. But then to my surprise, I received an invitation from Princeton to become a graduate student. What had happened is that Don Richmond, my professor, had run into Solomon Lefschetz, who was the head of the Princeton Math Department, and had given a good description of me to him, and so Lefschetz wrote and asked me if I wanted to come to Princeton. I don’t think I particularly wanted to go to Princeton, but on the other hand they were asking me to come and I hadn’t applied anywhere. So I decided I would go to Princeton. But then the summer before — we’re now reaching the end of the term at Cambridge — that summer I went to Austria to learn German. I had had German in high school, so I had a good foundation. And it turned out very funny, because this wasn’t long after the war, so these places were still kind of a little bit bombed and a little bit disorganized. So I arrived at the Innsbruck railway station carrying everything in my suitcase — I had one big suitcase. I had a family I was going to live with, a pension as they called a family that would take in a guest. And in the railway station I put down this big suitcase and went to ask directions or something, and when I returned it was gone. I wasn’t away very long at all. I was thinking of the US, and this wasn’t the US. So I had nothing left at that point. And in the suitcase (I’m sure the people who stole it must have been very disappointed) there were all my lecture notes, which would not be very valuable. First of all, they were terrible notes; second, there’s no real market for these things. So eventually I found my way to the place where I was going to stay. But I had been planning to learn all this stuff which I hadn’t learned; I was going to learn it in the summer from my notes. Those notes were gone. So what I did is I wrote to a friend of mine from Cambridge and asked him if he would loan me his notes, because he’d learned them although I hadn’t, which he was very kind and he did. I had been very uncomfortable not understanding the material, and so I decided that I would copy his notes into my notebook, but I wouldn’t copy anything until I felt that I really understood it. And that’s what I did in my spare time that summer, when I wasn’t bicycling or practicing German on the poor innocent people around, I copied this stuff, and by the end of the summer I really understood it. That was very helpful in my further work, because I really, really understood all those dull, horrible, techniques and I could make them work for me. It was a very nice family I lived with. Not long ago I got a note from the little girl of the family, now a grandmother, who had seen something about me somewhere and figured out that I was the person who had lived with her family long ago. It was wonderful to hear from her.
There wasn’t any particular program besides that that you were in?
I don’t remember any program.
You were just living there?
I think so. That was true even in France. When I was at Tours, they had a program. When I was in Geneva you were supposed to go to classes, but I never learned much from the classes. I learned by talking with other people. You know, with languages it’s an interactive thing. There may have been something like that in Innsbruck, but it obviously didn’t make much difference. So then I went to Princeton. The Princeton Math Department in those days was a very funny thing.
It was very famous in those days as well.
Yes it was, it was very famous, and I think it is still is. But the way they ran it, there were no courses. I mean you didn’t have to take any courses. They gave classes, but they didn’t give you grades, and I don’t think they cared if you attended. But after two years or so, you had to take the general exam, so basically you only went to class to learn how to pass the general exam, which was hard to pin down what it was. So it was actually quite a nerve-racking experience because you didn’t know how to prepare.
Were you working with any professors in particular at that point in time?
No, not at that point. I retained my interest in differential equations. So I struggled to get ready for this thing, and it was pretty nerve-racking. So much so that with my lack of interest in the rather abstract math I was learning, I actually left school for a while, but then I came back. But in retrospect, I now see how they ran the place, but I didn’t then. So I took the general exam, which I was just sweating huge beads of sweat about for two years, I went into the room and there were these three professors, one of them was Lefschetz, who was a most remarkable man. He had started out to be a chemist, and in a prank someone had rigged up his lab and something had gone wrong and blew off both his arms, so he had wooden arms and hands, but he could hold the chalk between the fingers. He was an obscure person, and he went to some university in Kansas, and there he did such incredible work that they invited him to Princeton. So he was quite a guy.
Was he initially European?
That’s a good question. He may have been, but I’m not quite sure. I can’t remember now whether he spoke with an accent or not. [Ed.: He was born in Moscow and was raised in Paris.] But he loved his work, and for some reason he was always very, very kind and encouraging to me, and I appreciated it. But at any rate, I remember going in there, and there were three professors. It was an oral exam. I was sort of stunned; I thought the questions they were asking me were incredibly easy. I thought well, this is an interesting way to do it. In retrospect, I think they did it to minimize their own effort — they didn’t have to prepare; they just went in there and asked whatever went into their heads. So the whole thing was one to minimize their effort. That’s why they didn’t have classes and grades and anything else. It was up to us. I remember when William Feller, a wonderful probabilist and a nice person gave me a language exam. I thought I’m not going to have a problem with a language exam! So he was going to exam me on my ability to read a French math paper. You had to be able to read a paper in French and a paper in German. This was going to be the French exam, and he hands me this paper. I looked at it. I said, “Professor Feller, that’s not French.” He said, “Right, it’s Italian. You passed!” [Laughter] So they did not deal much in the formalities.
I take it one had to be very self-directed in that program.
You had to be self-directed. And I think we were a pretty select group, so they sort of thought, “Hey, these guys are okay, let them learn, or something. And besides, we’re busy doing our research!” [Laughs] But actually, that was a reasonable fit for me. But I was so relieved to get through the exam, because I always felt my strength is research, I just always felt that way, whereas many of the students, they didn’t know what to do once they’d done their exam, because they weren’t getting much guidance about how to do research. But hell, I had already published a paper, and I knew that I had a lot of ideas at that point.
Do you think that really must have given you a leg up, having worked with Richmond?
Absolutely. But it was my natural inclination. I had a leg up and a leg down. Learning stuff was not my game. Research is my game. Some of them were there because they’re such good learners. I was just lucky that Don Richmond had run into Lefschetz and told him something good, and so that I was there.
And that you’d published with Richmond.
Yes, the thing we’d published. I mean it was really my stuff, I wasn’t helping Don do something; I had all the ideas and he was just very supportive, and he could make it in the right form. So then I started to work on my thesis, and I produced either two or three publishable papers quite rapidly at that point. So I wasn’t having any probable with my thesis.
One question that occurs to me is that a lot of graduate students might not necessarily know what would constitute a contribution in a field at any given point in time.
So you would have been familiar with the available literature as well.
Well, I knew something. But I just saw a lot of things that seemed like natural questions, so I went after them. And Lefschetz just kind of cheered me on. If I had gone off into something that wasn’t worthwhile or something, he would have definitely said something, but it didn’t work that way. But he was great. Occasionally I would write things on the blackboard when he gave a lecture, because he couldn’t write very well, so I would be his scribe. So that was nice, very nice. But the time approached for me to get a job. I had to decide what I wanted to do.
You finished there in ’54?
Yeah, ’54. I got my degree in ’54. I was there I guess three years. See, I was in England a year and then three years for the exam and writing my thesis and a couple of other papers. I forget which one was my thesis. They were all published.
I have a list here; I’m not sure if it is a complete list, but I grabbed it off of one of the citation indexes. The Richmond paper came up twice. Then there a few papers about trajectories in three spaces.
[Looking through and discussing the printed bibliography.] Then “Critical Points in Infinity and Forced Oscillations.” Yep. “Trajectories Which Tend to Limit Cycle in Three Space,” right, and “Trajectories Tending to a Critical Point in Three Space.” I think you’ve got them. That’s my career in nonlinear differential equations, which ended when I got my degree because I decided that I didn’t want to stay in doing that. I was offered what I guess we would call a tenure track position, but in those days you knew what you were going to get, I guess it was going to be at Brown or something like that, and I thought it isn’t what I wanted to do. I didn’t want to spend my life writing papers that maybe there were ten other people in the world who would read. It just wasn’t what I wanted to do. So I joined the Navy. [Laughs.]
I was wondering about that. I’ve seen that in your biography, and it’s unusual. Sometimes you’ll find someone in the military, especially in that period, before their graduate work.
Yes, there weren’t a lot of Ph.D.s in my class (at Officer Candidate School, Newport, Rhode Island). And you have to understand the structure. At that time there was a draft. None of my classmates ever served in the military because while you were a student you were exempted because you were a student, and then you took up a teaching position at a university or a college and you were exempted because you were a teacher. That’s what everyone did. But they knew where they were going, and I didn’t! [Laughs] So I joined the Navy, because I forgot to tell you that during my youth, when I was at George School and my mother was still alive, we rented summer places out on Long Island. One in particular was on Shinnecock Bay, which is a bay on the south shore, and I loved sailing. I’d read about it in some of my childhood books, books which had a great influence on me, I should have mentioned them, called Swallows and Amazons. It was all about kids who sailed small boats on English lakes and camped on islands, and boy it was really great. So I got very much into sailing. I loved to sail small boats. So I joined the Navy because I loved to sail. It’s very rational. And I never got on a ship, because their approach was, “Ah, we’ve got this guy! He can help us with technical stuff.”
Before we move on to the Navy, I wonder if I should ask you about some of the other people who would have been around you at Princeton. You’ve mentioned a few names.
I was very friendly with Herb Scarf, and Herb will reappear later. I guess Milner was in the next class or something.
I know we’re coming up [on] when you learned linear programming. Alan Goldman would have been there roughly then?
I didn’t know Alan. Alan does appear. I’m glad you mentioned his name, because he will appear in a funny place. And Al Tucker.
Kuhn and Tucker were still there, though of course you weren’t involved in…
It will turn out that I don’t encounter them until I come back to Princeton after the Navy. [Some discussion]
Then the other name you’d mentioned in one of your papers is Lloyd Shapley.
That’s later, too. That’s all after I got out of differential equation stuff. See, at this point I’m doing differential equations and doing it very well, but I don’t want to make a career of it. So I decided to join the Navy because I love to sail. Very good reason, right! [Chuckles] You had a choice in those days. You could either serve for two years if you just enlisted, or if you wanted to go to Officer Candidate School, then you signed up for three years, or three years and four months, because you went to four months of OCS (Officer Candidate School), then three years after that, so it was really three years and four months. And I got married. As soon as I got my PhD we got married, and then I joined the Navy. So they sent me to OCS.
Had you met your wife at Princeton?
No, actually I met my wife… I mentioned that I was sent to Europe several times. I forgot to point out that when you were being sent to Europe in those days you didn’t fly, you went on a boat, and I met my wife on the boat on my first trip when I was going to Geneva. She was going over with her family because her family had a lot of Swiss connections, and her mother was a friend of my mother, and in fact they had helped find the family that I was going to stay with in Geneva. So we were on the same boat and we met on the boat because our parents knew each other. And then we got married. While I was a graduate student I would go and visit her…they had a sailboat, too, I sailed with her family.
Was your wife living in New York?
Her family lived in Roslyn, Long Island, which is on the north shore. So we got married at that point. Where was I? I was going to OCS, and that was a really good experience. For some reason I felt quite exhilarated by it, so I felt I was doing the right thing. It was tough. You never got enough sleep. That was on purpose. But we lucked out, because that year there were two hurricanes, and when the hurricanes hit we were in Newport, Rhode Island. They couldn’t do anything —, we couldn’t march to class, we didn’t have class. And so we could catch up on our sleep, and believe me we were the happiest group you ever saw. But it is very interesting and very instructive. For example, we each had a locker. We had no possessions other than our uniforms and things, but each piece of the uniform, underwear, everything, had a fold — had to be folded a certain way — and they had to be in a certain position in the locker, and they inspected to see whether they were. And so we spent a lot of time learning the folds. And then four weeks into it I think it was, they changed the folds — they issued a new set of how you were supposed to fold. That’s the sort of thing it was, and it was quite deliberate, and for some people, it was just too much chicken shit or whatever you want to call it. But it didn’t bother me particularly because I felt I know why they’re doing this. And so we had a very good group. I had three roommates — four to a room — and we still meet. I guess we met about three weeks ago in New York. So anyway, there were people there who could tell you how many hours were left before graduation. That’s how a lot of people felt about it. Not how many days; how many hours. And believe me, when we graduated, we were the happiest people you ever saw. We just kept looking at the gold stripe. But I was very disappointed because I was sent to Washington. I had applied for mine sweepers, which is the smallest boat. I wanted a small boat, —
you wanted to sail.
It wasn’t a sail boat, but it was the smallest damn thing that you could apply for. So I went to Washington and I was at the ONR, Office of Naval Research, and they assigned me to the physics branch. And they did all sorts of odd things, and one or two things that were fun. But that had to do with the OR branch, which was down the hall, that thing about the weapons system and so forth. But I did one thing in the physics branch that was fun, and I was just lucky to get this assignment. The astronomer, Fritz Zwicky, had thought there ought to be a way to detect incoming ballistic missiles, because when they hit the atmosphere there was a big shockwave. He predicted that any astronomer who was looking at a star, the shockwave would cause the image to jump, to displace, and he thought that might be a good way to detect incoming ballistic missiles. Well, it seems to me it’s late in the day, but what the hell, they didn’t have much way to detect it. So somehow that got handed to the physics branch to take a look at, and they handed it to me and Sid Reed, who was a civilian employee, Sidney Reed. I don't know whether we decided or we were told to test this out, but the way we were going to test it out is we were going to fly out to the Naval Ordnance Test Station in Inyokern [California]. Inyokern is out in the middle of the desert on the border of Inyo and Kern Counties, and believe me, there’s nothing there, and then we’re going to do a little experiment, which I will describe to you.
Did you know John Williams at the Rand Corporation, mathematician?
I think so. I think I did.
He was the head of mathematics at Rand, and before he went to Rand he had gone to Inyokern, and finding no housing there, he went to I think Pasadena, the Pasadena branch, and it was from there that he was recruited into the Rand Corporation, allegedly because he was so lazy he could always find simple ways to solve hard problems, which is what Warren Weaver said.
That sounds like the person I remember, yes.
Anyway, that’s an aside.
I can see why he couldn’t find any housing in Inyokern. Well we flew out to LA in a DC3, and then we flew to Inyokern. That was a two-prop thing, as I remember. All these things are so educational. Inyokern is a collection of Quonset huts, basically, and I thought, god, I’ve really reached the desolate end of the world! I was looking down at this big collection of Quonset huts. Just desert stretching out to the horizon and some mountains, that’s it — nothing else. But as soon as we landed, because Sid and I were going to do this great experiment together, I was talking to some of the people that were stationed there — it was all Navy people — and I said, “God, you know, you weren’t lucky to get this post, were you!” I’m sure I said it a little more tactfully, and they just thought this is the greatest place in the world. They loved it. They had their own community. Nobody worried about what their kids were doing, they were right there. They thought this was the greatest place on Earth, and that really made an impression on me. At any rate, when night fell, Sid and I trudged out into the desert to get away from the lights. Sid set up a telescope on a tripod and trained it on a star. I got out my Navy — I can’t remember whether it was a .38 or a .45, but anyway a big caliber, horrible thing. I don't know if you’ve ever fired one of those things. The recoil is so big you can’t hit anything with them, they’re just awful. So Sid trained the telescope on a star, and I was blazing away into the sky with this gun, and nothing happened. The experiment was a total failure. Finally we trudged back, got on the plane the next morning, flew back. It was a couple of weeks before… I don't know whether we figured it out or someone else did. The bullet from those goddamn pistols is subsonic — there was no shock wave. If we’d have had a rifle it would have been another story. But by that time everyone had lost interest, no one wanted us to go back out there, it was over. And that made an impression on me too, that that’s the way things work. There’s timing. And that was also a lot of fun. So, anyway, I eventually got out of the Navy, and again Princeton was very nice, they invited me to come back to Princeton as a Higgins lecturer.
Now before we do that, should we talk about your exposure to Operations Research while at ONR? You were in the Pentagon?
No, no. This was a temporary building. I think it was a World War II temporary. You know where the Washington Monument is? A corner of that big thing, that big grassy thing the Washington Monument is on was covered by the ONR, which was a temporary building. Across the street from us (not across Constitution Avenue but whatever the street is) was Main Navy. The Pentagon, I don't know whether it was full. But anyway, Main Navy was a World War I temporary, and a lot of Navy people were in that. And then I went down the hall where they did some OR things, and there I heard there was something called Operations Research. What it consisted of, as it was described to me, was mathematics applied to everyday problems: problems of a company, scheduling problems, blah, and blah. I thought, oh, that’s what I want to do! So my original intention after getting out of the Navy was to get a job in OR somewhere, because at that time companies were forming OR groups, almost all of which have subsequently disappeared. So I started going to school at night, and I think I took a course from Alan Goldman, though I’m not sure.
I was just reading that he just died I think a few months ago.
I think he was teaching down there in Washington. I’m pretty sure.
It wasn’t that far out. Anyway, somehow he had something to do with my education.
At the time I think he was at the National Bureau of Standards.
I think that sounds more like it. So I took a couple of economics courses, because I’d always had this interest in economics, and I think I took a linear programming course, but I’m not sure, but somehow I learned something about it, and I think it was by going to school at night. I think Alan, maybe he taught the course, I’m not sure, but he was in there somewhere. So while I was on this path to find a job as an operations research person, I got a note from Princeton from Lefschetz inviting me back to be a Higgins lecturer. I don't know, I talked to them, I called them up or what, but I said, “Look, I don’t want to go on with nonlinear differential equations.” That was his field. He said, “Doesn’t matter.” He would have me back anyway. That was typical Lefschetz. He was just a big person, not a narrow person. So I thought well, I really don’t know anything about operations research. Maybe I should spend a year at Princeton and learn OR and have this prestigious background that I was a Higgins lecturer. So I went back there with the intention of learning something about OR. At that point, then, I ran into Al Tucker, and Harold Kuhn was there, and people who were interested in game theory and stuff like that.
Nash was younger; I think a year or two behind. He was there. But he was disagreeable. We used to have tea in the afternoons in the common room there, and people tended to try and outsmart each other, show each other how smart they are, and he was more extreme than most at that, and that was never my particular style, probably because I wasn’t good at it. I’m more the kind of the slow and deep rather than quick. So I went back. And Al Tucker was always very kind to me and very helpful. And somehow I had retained this tie with the Navy that they wanted me to help them with some studies, so I did that, I would go down every now and again. I was paid, so I was sort of a consultant. I wrote about it later in the book “Early Integer Programming”, The Navy people were making these models of task forces and coming up with fractions of ships, so they were asking me is there some way to get — we need whole numbers of ships. So it immediately struck me that when you solved ordinary linear equations, there is a sub-branch in that called Diophantine equations, which is about getting whole number solutions from linear equations. Linear inequalities are what linear programming is about. It’s plausible that there’s some equivalent thing. So I went back to Princeton and thought about it. I can’t improve on the description I give in the book. I could repeat it; well maybe it’s better to have it in one piece. Anyway, I struggled with that because I wanted to get the answer. [Laughs] So I tried all sorts of things related to Diophantine equations, and I could not get any progress. One day I said to myself, supposing they gave me an actual numerical problem and said, “Get the answer.” What would I do? So of course I said to myself the first thing I would do is I would solve the linear programming problem, the linear equations problem and not worry about the integer constraints, and then I’d go on from there because I’d be in the right neighborhood. In addition, I said to myself something like if we’re trying to maximize the objective function (the score (or criterion) you’re trying to maximize ), and it came out seven and a half, and we’re trying to get the maximum score, and we got seven and a half with all these fractional variables. Well then at least I’d know that the integer thing — because all things had integer coefficients, so the score was going to be integer as well as the…
That means you couldn’t get more than seven, if everything had to be whole numbers. And that seemed pretty obvious, so I said to myself how the hell did I know that? Because it is obvious, but I wanted to know why is it obvious? It’s because if you have something that has to be a whole number, you maximized it on the inequalities, the constraints, the polyhedron, whatever you want to call it, then if it comes out seven and a half, you can push it in down to seven, or whatever it is. I said what? Why do I just push in the objective function? I could construct lots of other things there and push them in. And that eventually led to what are now known as the Gomory cuts, which are still the most widely used cuts there are.
So linear programming at this point is not a particularly old field, but at the same time it’s established in the late 1940s through the work of Dantzig and others. And of course at Princeton (this is all elementary if you happen to have studied this, but given that this is an American Institute of Physics interview I feel the need to go over it in a little bit more detail). And of course at Princeton you have the question of the duality of the two-person game and the linear program, and that’s Kuhn and Tucker. So the question that I wonder about, is this sort of a logical extension of that, or is it something wherein you have to go out looking for it?
I think that people understood that there… I think there was a desire to solve the linear programming problems in integers. I don’t think anyone was working on it to speak of, because I think it was regarded as hopeless. One of the reasons I think that, is first of all I’d never heard of it [chuckles], and secondly, we had a seminar that met every couple of weeks on linear programming and stuff, and Martin Beale, E. M. L. Beale, an English person, was visiting, and I knew Martin.
I can’t remember where he was from; Martin Beale was a very good man. So I signed up to give one of the seminars, and he said, “You’re giving the next seminar. What’s it about?” And I said, “It’s about solving linear programs in integers.” He looked at me and said, “You can’t do that.” That I think reflected that. So anyway, that sort of put me on the map. And Al Tucker was very, very helpful to me. He gave me his time at some meeting to present this stuff, and a whole lot of things like that happened. “ Oh I’ve forgotten to say my oldest son was born just before I got out of the Navy. He was born at the Bethesda Naval Hospital. We were living at Princeton over a store, my wife and I and Andrew, my oldest son, and I still have this desire to do practical things. So IBM is just starting up its research, and Mannie Piore was appointed the head of it. Now I know Mannie Piore as a name because he had been the head I think of the ONR, or if he wasn’t the head he was the lead civilian person.
A couple other names from the Navy period. Thomas Killian?
Killian, yes I think that sounds very familiar, but I think he was a civilian.
He would have been the chief scientist there. And I think in one of your articles you mentioned Frank Isaacson.
Aha. Frank Isaacson was the head of the physics branch. He was a very good man. I really liked working for Frank Isaacson.
And operations research, which would have still been the period when Jay [Jacinto] Steinhardt was the head of that group.
I could be. I’m not sure I ever met the head.
I didn’t want to press you on it too much because you mentioned at lunch that you weren’t really conscious of the organization.
I don't know, I just sort of went down the hall and did work for them. It was quite informal. So I never quite knew who they were. But I was just attracted by the nature of the work.
And you said it was sort of weapons analysis.
Yes, like that thing I talked about with the infrared.
Just for the sake of the recording, I think that was at lunch we were talking about that.
Well I think it was them. Anyway, they were studying a weapons system which involved flying over the Soviet Union and finding out where you were by looking at the infrared emissions from power plants, because boy they emit! You can’t miss them in the infrared.
Unless it’s cloudy.
Unless it’s cloudy, right. And I was uncomfortable with the analysis. It wasn’t the OR group had done it, but the proposer, which was one of the big airplane companies from Texas. But everyone thought that we can’t get detailed cloud data from the Soviet Union. But I went down to the Library of Congress one Sunday, and lo and behold, they have piles and piles of huge green books with cloud data from all over the world. And I just thought I’ll look and see what there is about the Soviet Union, and it turned out it reported everything, so you could get the cloud data over Kiev every day. As soon as you looked at those books you knew that system couldn’t work. They had cloud cover I think around 1,200 feet, solid clouds for a couple of days, and then it would break up. In other words, clouds came in I think from the east and covered all of western USSR and then move on, and during that period you weren’t going to see anything. So that’s something I did for somebody. I think it was that group. But it was very interesting. I love doing that stuff. So anyway, we moved up to Princeton, and I heard that IBM was starting this, and at the same time Herman Goldstine — now here’s a person who was very important to me... Herman Goldstine was the associate of von Neumann, and together they wrote some famous paper which outlined digital computers. It was von Neumann and Goldstine. Herman was at the Institute for Advanced Study, and Mannie hired Herman to start a math department as part of the IBM research. Herman had a long history at Princeton, and he invited me to join this organization, which I did.
This is in ’59?
’59, yeah. I guess I returned to Princeton in late ’57 and I left in ’59 and joined this new organization, which was housed in something called the Lamb Estate, which had once been a home for alcoholics or something. And I was the head of a group of four or five or six people, one of whom was Benoît Mandelbrot. So we worked on operations research problems, but it was hard because we were very isolated out there, and so people tended to just do their own thing anyway and not have the interaction that I had hoped for. But we managed anyway. And Paul Gilmore, who was one of the group, and I got involved in a series of papers on stock cutting. If you go to a paper mill — and I recommend going to paper mills, I’ll explain why in a minute — you’ll see that what they make in a paper mill is an endless stream of paper that comes rushing out of the machine, and it’s about ten feet wide. They roll it up on big ten-foot spindles, so that’s what you get. So what comes out at the end are big rolls, like ten foot high, and you have to cut those up into the sizes that people want and the quantities that people want. So you might have orders from a newspaper we want so many things of this specified width…
And it doesn’t add up to the total.
Does not add up to the total, you’ve got it, exactly. And they have other customers with other widths. So they have this assortment of widths and quantities, which they are somehow supposed to make out of all these ten-foot rolls. So that was called the cutting stock problem in the case of paper. So Paul and I got interested in that. We struck out (failed) first on some sort of a steel cutting problem, but we seemed to have some grip on the paper thing, and we used to visit the paper mills to see what they actually did. And I can tell you, paper mills are so impressive. I mean they throw a lot of junk in at one end, like tree trunks or something that’s wood, and out the other end comes — swissssssh — paper! It’s one damn long machine, like a hundred yards long. They smell a lot, too. We were quite successful. They didn’t have computers; believe me, no computer in the place. So we helped the salesman to sell them the first computer.
So is sort of the rationale of what was going on is that in having a mathematics group that did operations research?
You have to realize that there was a lot of misunderstanding about what research was for. I’ll talk more about that when I become Director of Research. But the general belief in the upper management and pretty much among the people was that somehow mysteriously you left them alone there and something good would come out.
Was this Piore’s attitude as well?
I don't know. You could never tell what Mannie was saying because he very seldom said anything. He used to grunt a lot. But he had a lot of insight, so I don't know what he thought about that particular question. At any rate, we really figured out how to do this paper stock cutting problem. But see, as usual, there are other things other than the rational. A lot of these paper mills had older employees who instinctively knew how to cut up these things, and it turned out there was a lot of difference in skill. Some people could do this and some people could not do this. And all the older people were retiring, and so the mills wanted to have something they could count on. That made it easier to sell our method, our algorithm. They would give us a month’s worth of output and say, “Well what could you make out of this? Here’s the list we want.” And we could show that there was less stuff thrown away, less leftover strips. I don’t think they would have done it, though, if it weren’t for the fact they also felt more secure if they had something under their control.
Because they had the employees who had the instinctual control over it. That’s odd, because it kind of matches up with what I’ve seen in other areas when you’re talking about inventory theory and production control, for example, that’s sort of an effort to express explicitly what seems to be known implicitly, and then to improve on it sort of at the margins.
That’s right. But if you can express it explicitly, that sometimes does something for the management and for the other people, even beyond whatever you say, because they can reproduce it. And if Tommy gets run over by a car, then they can go on. So we had a lot of success with that, and we went on from there and learned to cut glass, which was different because it was in two dimensions. We got the Lanchester Prize of the Operations Research Society [of America] for that work. It was exciting. That was why I went there, but it was not typical of what was getting done at that time.
You didn’t do much consulting outside of the research?
Well, who would you consult? People didn’t have math problems. I remember visiting one part of the company once where they were having some kind of problem, and their attitude was, “I don’t think we would have a math problem that we couldn’t deal with somehow. That just isn’t where our difficulties are. We can somehow solve our math problems.” So they weren’t rushing around wanting help. Later on you’ll see they never do want help.
How deeply or not were you kind of keyed into the ORSA/TIMS communities at that time? At this point you’d already been to Rand, right?
Oh I forgot about my whole trip to Rand, which is too bad.
Was that before Princeton or after, or in between?
When I got out of the Navy I went to Princeton. When I was at Princeton, I invented this integer programming thing. When the people at Rand heard about that, they invited me out for the summer, so that was I think the one summer that was sandwiched between two years of being at Princeton. So I flew out to Rand. I forgot to tell you, how was I actually solving these things? I may be getting too detailed, but I’m telling you the things that were fun and meaningful. Well, nobody had computers to begin with, so I was doing these things mostly by hand, tiny problems. And I think I may have had an adding machine. But then the Statistics Department had a computer, called the E101, Electrodata 101, and they let me use it I think after 1:00 A.M. in the morning and up to something like 6:00 A.M... Not a very desirable time, but to use the E101, which was about the size of this desk, and its rate was one execution a second, one add a second. It clicked. The operating things were relays, so you could hear click, click, click. And oh, that was so great, you couldn’t believe it. And you programmed it with pin boards. It had an internal memory, but I think the internal memory only had about 120 characters, so you didn’t put the program into that; it just used that for computing of stuff. But you had pin boards which somehow told it what next to do. You could put three or four pin boards on it, so I you could give it a sequence of about 100 instructions on the pin board, which meant that your program could not be very long. So I wrote the first integer programming program that was ever executed on that. But you see, what you had to do was I had a set of boards that would do linear programming, it would do it, it would solve the linear programming problem. Then I had pulled them off and put on a couple more pin boards which generated what are called cutting planes — that’s the new addition to the problem. I’m not sure quite how they stored the cutting planes. Then I put the linear program pin boards back and run it to the next round. That’s how I got the first results, which were very promising and very small.
In terms of the ease or rapidity of the calculation?
They were small in the sense that they would have maybe three, four, five variables. By hand, I don't know, maybe I got up to three. Of course they didn’t make mistakes, either, which was wonderful. Anyway, that was very exciting. Oh, I forgot the whole thing about computers. The real reason I joined IBM goes back to when I was at camp (Wentworth) and I read some of the works of H. G. Wells. He goes off into the future somewhere in which people don’t have to work because the machines did all the work, and that was a very attractive idea to me. I thought that’s really fabulous. It would be like being at camp all the time, which I loved. So I puzzled it, well why can’t we do it now? Because we had very intricate machines. In factories they had machines that did all sorts of things. I finally realized that what was missing was the intelligence. You see, his robots, they somehow knew what to do, so they did a variety of complex things, whereas the robots I could see in the factories they just did one thing repetitively. So when I first heard about computers and started to realize what they could do, my reaction was wow, this is what I’ve been looking for! So I didn’t join IBM for any other reason. I mean it was a big company and it was fortunate they were starting a research arm, but I really joined IBM because I thought computers were really going to be a big, big thing, the missing intelligence. And that turned out much better than my idea that mathematics was the solution to everything. That did turn out. So I guess there I was. We’re now back to the Lamb Estate and by that time my wife and I had two children.
This is prior to the Yorktown Heights opening?
Yeah, it was before there was a Yorktown Heights, right. But then eventually we did move to Yorktown, so the Math Department grew. Herman Goldstine was the head of it. I guess when Herman left; I guess I became the head of it.
Martin Shubik was there for a little while, wasn’t he?
Martin was there briefly. He’d been at GE previous to that. Martin was there briefly. And Phil Wolfe, Phil was a very good friend to me in addition to being very well known in linear programming. He was out at Rand when I was, and he took care of me. I wasn’t allowed to go into certain parts of the building, and he would take my programs for me, and he was very hospitable. And I think Harry Markowitz was out there too, and he was another good guy. At a much later date he came to IBM for a year or two, but I think he didn’t like it. And Phil, when I became head of the Department, or maybe earlier when I had a larger operations research group, I invited Phil to join IBM, and he did. And he was always a very good friend.
So is that distinct from the mathematics group then, which the operations research group becomes?
Well there’s the Math Department, and then within it was the OR group, and I ran the OR group. Then I think when Herman retired I ran the Math Department. Something like that. I took a year off at one point.
You mentioned Mandelbrot was there.
Benoît was there from the beginning. He was in my group of five. We have a really neat picture. It’s worth getting. It’s probably in there somewhere [Turns on computer.]. It’s worth a minute or two to see…
Was he working on OR type things too? Because obviously he becomes famous for the fractals and that.
First of all, Benoît was a very admirable person in my opinion because he always had great ideas, which no one appreciated, and he stuck with them. He made thousands of enemies because he wasn’t tactful, but his ideas were always really, really, really good, and he knew I appreciated it. He’s always been very good to me and mentions me in forwards and things like that. I admired his work and supported it from the word go. So I really like this picture because it shows me at the blackboard, and Paul Gilmore is sitting there and Benoît is in the background. That was the Lamb Estate. [Ed. A copy of the picture is printed in Gomory’s foreword to Mandelbrot’s Fractals and Scaling in Finance]
Before we got onto Rand I was asking you about your involvement with the ORSA and TIMS crowd.
Of course at that time I went to all the meetings. I was a regular person. Somewhere along the line I got the Von Neumann Theory Prize, but I can’t remember when. So I had become a recognized big-time OR person, so I was very much into that. I took a year off at one point because I was getting a divorce and it was very disruptive, but then I came back and became head of Department again. We had a few Directors of Research, and then I guess Mannie was figuring that I should become Director of Research because they offered me a job over in Armonk, the corporate headquarters So I was over in Armonk for about six months, then I became Director of Research after Art Anderson, and I remained there for about 18 years.
In the 1960s you mentioned that everyone sort of did their thing.
More. But it depended. You see, when I said that I meant the Math Department. But if you were in the Semiconductor group, you were trying somehow to do something… You see, there are degrees of difficulty. The advantage you should have if you’re in a company is that you should have very good access within the company. Access to outside the company can still be extremely difficult. For instance, our access was to paper mills. Well, that wasn’t within the company. But the whole access problem is non-trivial, as you will see when I start to talk about technology transfer But you see, it wasn’t really driven because the dominant idea was that if you leave these people alone — I don't know how they phrased it. I think I mentioned in one of these articles that the idea was you put an academic type institution on the top of the hill and the ideas flow down. That was the widespread belief, based on a misreading of history by the way.
So as far as say the Mathematics Department’s relationship with say the Physical Sciences people like Rolf Landauer was there at the time. You wouldn’t have…?
No, there wasn’t much relationship. So that picture, let’s see if I’m just lucky, then if I’m not we’ll stop pursuing this. But somehow that picture, there is something about that picture that captures…
I also have to ask you about integer program and pricing.
That was a paper I wrote with Will. At Princeton, Al Tucker introduced me to Will Baumol, who was a young professor. Pricing in economics had been tied to linear programming. It was that link.
Yes, I’m interested in that link, because historians of economics actually are sort of starting to become interested in some of the crossovers in these areas. And so I know that there’s the Dorfman-Samuelson-Solow book, but I don’t know much about it beyond that.
I don’t either. The idea was I’m just adding more cutting planes, so if we can price the cutting plains, somehow we’d get prices that would work with integers, or something like that. And so we wrote this paper, which a lot of people reference. I honestly don’t think it was that good a paper. I just don’t think it’s possible to really have meaningful prices, but I could be wrong. It was fun to work on it, but I never felt that we got where I wanted to get with that, so that’s why I passed over it because I forgot it. But on the other hand, it did lay the foundation for my working with Will, which I do now. So eventually, [after] maybe six months in Armonk, I became Director of Research, succeeding Art Anderson. So I was uncomfortable with the situation, because I did not believe that this putting people on a hill was going to last. There was no coherence. We had a book of stuff we were doing as a division, but it was this, this, this, and this.
When we interviewed John Armstrong, he had sort of referred to not really feeling what is sometimes referred to as the “golden age” of industrial research as not being that at all, because it is I think of the opinion you’re expressing.
I should certain hope that John Armstrong, after 18 years of my running the thing, should express that opinion. [Laughter] Yes, completely agree. So I was very concerned that this thing just wasn’t going to last. It was fundamentally based on an illusion. And I talked to Frank Cary, who was my boss, shortly after — When he was first my boss he wasn’t the CEO, but about two years later he did become the CEO, but he continued to have me report directly to him. He was in my opinion a person of good depth. So I asked Frank, “Look, what do you think we’re here for?” ‘we’ meaning the research division, and he said, “Well, I think the feeling is that sometimes all sorts of new technologies come along, like the computer itself, and we want to know it’s coming.” So it was a sort of a lookout post. And so I thought that’s not going to do it. I can do that with 20 people; I don’t need 1,200.
Which is what you had at the time?
About that. Something like that. And I knew enough about things at that point, I figured we’re not going to make it unless in addition to doing scientific work of merit, we will do something for the company that makes a difference. So I tried a lot of things. I remember my earliest attempts; I would gather up my directors, that was the next level under the Director of Research, the Directors.
So as Director of Research, where are you located?
Then there is a laboratory director beneath that?
No, there was no lab director.
Because you’re over Zurich and Almaden as well?
Yes, although it wasn’t Almaden then, it was the San Jose Lab. That’s right, we had a lab director for San Jose and a lab director for Zurich, but the Director of Research and the Lab Director at Yorktown are the same person. So when I say I gathered up my directors, I meant the head of the Math Department, the head of the Physics Department, the head of the Semiconductor group, the head of the Computer Science Department, the heads of the San Jose and Zurich labs and we visited all the development labs, one after another, and saw what they were doing, and we asked them what did they think, how could research help them? Which seemed to me a natural thought. It isn’t, though, it’s not. If we didn’t know much about them, they knew less about us, of course, and so the things they wanted out of research were usually impossible, unreasonable, and unimportant. Useless. Their requests were fundamentally useless, and it was because the things they really wanted they were doing, so if they had some sort of left over dreams that they couldn’t afford to spend money on, that’s what they wanted.
The really kind of out there sorts of things, pie in the sky.
Yeah, usually pie in the sky, or unimportant. So that approach, which seemed plausible at the first sight, wasn’t. So we started to work away at this year after year, trying various things. We had a lot of successful individual projects with the labs. And I also then invented something I refer to as The Invention of the Wheel. I invented the wheel chart, and the wheel chart was a chart, but it was circular, and it had the different areas, like the semiconductor area, the physics area, the math area, the printer area, and within these areas we tried to develop coherent programs, and we gradually realized that the kind of program you wanted to have had two parts. You had a program dedicated to improving the in-place technology (this is really a key point; I want to come back to it), and a program to look at radical alternatives. Both.
This is what you decided upon as a goal, or this is what existed?
No, I decided that we should organize ourselves that way, but once you organize yourself that way, you get more. I mean there were things that didn’t fit in, but I didn’t worry too much about that. We could pretty well, over time, organize the division in that direction, but that doesn’t solve all our problems. For one thing, it was very clear to me that the development divisions were not expecting anything from us, or counting on us, and that also we were naïve about what you really had to do to be successful, because for a long time my guys would go out there and say, “Gee, we’ve shown them how to make a better such and such for the printer, but they’re not taking it up. They just don’t get it.” And I believed that for a while. But that’s because they didn’t understand. One of the things I’ve really learned is if you think the other guy is stupid, it just means you don’t understand. But if you’re developing a printer and it’s on a two-year development cycle before it goes into manufacturing, and somebody shows up one year down the cycle and shows you how you could have made a better head, that does not help you at all because you couldn’t possibly switch. So if you’re working with people who have that kind of life, you’d better be there at the beginning. Secondly, you learn, like a memory or something like that, it’s not just good enough to have something that switches state. There are many things that have got to be able to switch a hundred million times back and forth without degrading. They have to switch fast. They can’t use much energy switching. Blah, blah, blah. And so you have to know all the requirements of a real memory, and the physicists don’t. In those days, “Oh look, we’ve got a two-state thing here.”
How did you, coming from a Mathematics Department, kind of adjust to having to deal with all these varieties of concerns?
Well, they all are different… You just learn. At the very beginning I certainly asked some very naïve questions, like, “Oh, why do you guys have to do these experiments, firing electrons into the surface of silicon? Can’t you compute? We know where all the darn silicon atoms are, and we know what an electron is. So why can’t you just calculate?” Well it turns out that most of the things that one can know about it you couldn’t calculate; it was at that time out of reach. Now someone at San Jose was trying to do it mathematically eventually, but it was a lot easier just to do it by experiment. So we learned that, in physics, experiment tended to lead the way, and then you try and explain after the fact what happened. You have to learn that kind of stuff — that’s the way it really works. It’s not that hard. If you can relax and listen to them and be unafraid to ask the questions in your mind, after a while you get the sort of picture which you need to get. You certainly never know as much as the people. So anyway, we organized along those lines, and that helped a great deal. First of all, it made clear to everybody why we were there. But one of the parts of the wheel was science, and I would say, “You guys, you don’t have to be practical.” My view was you have to have a foot, you have to be involved in IBM, and you have to be involved in the university community both. We want those contacts, we want new people coming into the universities, and we want a research reputation.
Did the IBM fellows have a particular role to play within this?
Later, they come later. But really, an IBM fellow is to give a goal to the researchers that is an honorable one. When you become an IBM fellow — which reminds me, they did make me an IBM fellow I guess somewhere before I became Director of Research, so it didn’t come along later. They’re supposed to just let them do what they want to do, and that was people like John Backus who invented FORTRAN and people like that. But that wasn’t main line. They do whatever they want to and we would support them, but that was a side line, really. So it was a long struggle, but we kept getting better and better at it, and more and more realistic about what it took to actually help the development people. We were very helpful to them I think in some production crises they had, so we became more and more credible because we actually did things they couldn’t do every now and then. And if you can help people in a production crisis, it makes a big difference.
There was a big one in ’79 or ’80, somewhere in there.
Yeah, right along in then. I forget if that was the cracked stripe or the multilevel ceramic, or something. Anyway, there were problems. In the beginning when we would show up, they didn’t want us around; they thought it was a waste of time — we’ve got a crisis here. The thing is, take physicists. Physicists know more about how the goddamn material is put together than the engineers do. The engineers do the thing that’s been working. But if it suddenly develops cracks that shouldn’t be there for any reason, just because electrons are flowing through it, come on! They don’t know what to do next. Our physicists do. So the physicists have a greater depth of knowledge, and it’s interesting that most of them, when you get them involved in doing that, they love it. They spent their whole lives doing academic things, and some of them just love it; some don’t. But having someone respond, trying to solve the problem, it’s exciting for them, too. So we were able to pitch in, and that helped our credibility a great deal. And on one of these occasions or trips, Jim McGroddy, who was the second long-term Director of Research (because it was me and then it was John [Armstrong] for two years or something, and then it was Jim for ten or twelve [Ed., McGroddy was Director of Research only from 1989 to 1995]) came up with the idea that eventually we called the Joint Programs, which is that we should have something organized between us and the people that would be an organizational structure.
Is this the Advanced Technology Laboratories?
ASTL was the first one.
Advanced Silicon Technology Laboratory.
Yes, it was the silicon technology one. And that was fun, because I think I even managed to persuade Erich Bloch, who was the one who became later the head of the National Science Foundation… Eric was a big semiconductor guy, and you see, I managed to work out that deal with Eric because our credibility had improved. So we did the ASTL first. I think Bob Hanley was the head of it. He was a very good fellow. And the formula we gradually evolved was that most of it was paid for by the Development Division, but not all; we paid for our guys. We usually had the head to be a research guy, but I’m not sure that was always the case. But the real point was it was their guys and our guys together, and so we didn’t have a transfer problem. When we had the stuff it would be timely because these guys weren’t interested in doing it if it wasn’t. And it was theirs, and they were in the Division! And also it was very good for them, because in the divisions, development division or manufacturing division, manufacturing is the priority because they’ve got to get the stuff out the door. Behind that, development is the priority for the next product, because it’s got to be ready on October 3rd so they can manufacturer it. Behind that, they always had something called Advanced Technology, and that was way down the totem pole, so it was odds and ends. So a lot of that became part of the Joint Programs. The management of the Development Division was happy because now it was organized and had good leadership, and was tied into what was happening. And we were happy because it was our way in. So it worked. So we started with the ASTL, and by the time I retired from IBM we had 16 or so.
This was all over the course of the 1980s.
This is all in the ’80s basically, about ’80 to ’89 was the period I’m talking about. We learned how to do that, and that was what really, really worked. It took a long time to get there, and it took a long time to figure out just how to do it. That’s sort of the conceptual history. Now of course there were real events that happened along the way, like the design of the RISC computer and the invention of the relational database out in San Jose, things like that. And we won a couple of Nobel Prizes.
So how are we best to approach a history like that from your perspective, I suppose?
The thing is, I did some things wrong, some things right. But I did the big things right, which is my concept for the Division was right, that it had to do these things, but I also kept it tied into academia. So we didn’t just invent the relational database; we won Nobel Prizes, and Nobel Prizes were useless to IBM.
So there must be a certain realignment of priorities after you try and tie in the research activities more to the development laboratories.
Yeah, but I mean the guys doing semiconductors, they were mostly dealing with semiconductors rather ineffectively, if you see what I mean. But then they also did research on things about semiconductors, and I was perfectly happy with that. For a long time we had both, but we understood they did different things. Now when IBM went under great stress, which was really after I left, then there was a lot of stress on the budget. Although Jim did a very good job of fighting against it, he went much farther in the direction of eliminating the scientific work. But he was also very good at tying them into IBM. But he did a good job. See, the concepts—I could live with this concept, which I sold to the company, but not in these words. At one of our early annual meetings of the division leadership I introduced the “chicken concept”. Someone gave me a chicken tie as a memory. I said there is no such thing as an all-meat chicken. If you have to have a live chicken, you’re going to allow some space for the claws, for the beak, and for all the rest of it — you don’t eat the whole chicken. And that was my view of the Research Division, and I wouldn’t let them take off the beak or anything else. I could sell this to Frank Cary, financial people always struggled against it, but 18 years in a row I got my budget. Whether I would have been able to do that during the crisis I don't know.
One of the other things that I noted is one of the policies that you had in the Division is before a project would be discontinued it would have to come up to you.
Yeah, that’s true. Where’s the cube? Aha! Let’s get this around.
We are looking at… “IBM research goal: A Research Division famous for its science and technology and vital to IBM. Principles: Excel technically, know IBM, know the technical world, provide technical leadership.” It’s encased in plastic, but it is sort of a three-dimensional shape. [Gomory showed Thomas a clear plastic cube with the above goal and principles written inside of it.]
You see, that is very clear, and everybody had these things. This is my concept, that we should be famous for science and technology, and vital to IBM. In later years they dropped the “famous for science and technology” part. That I don’t agree with. They may have had to because the times were very hard and the thing was not well managed. I don’t mean the Division but I mean the company. But that expresses my philosophy right there, and it worked. And the Division is still there, which is more than you can say about practically any other research division of that period. That’s because we could do both.
So if we were to talk about some particular research programs, I don't know exactly what would be the best ones. Josephson springs to mind.
Of course that was a learning experience. That’s $100 million that IBM paid for my naiveté. I didn’t understand. The Josephson program was an attempt to use super cooled switching of Josephson devices. They work faster. I remember talking to Bob Henley, very experienced silicon guy, and I remember Bob said to me, “Look, silicon is a worldwide technology.” Bob never said very much. And I said, “Of course it’s a worldwide technology. But how can anything new come in?” That exchange indicated what I missed.
This was early on?
This was early on. We were used to the idea that silicon came along and displaced whatever we had before that, and transistors replaced this and that. Memory: memory went through all sorts of peregrinations. So we’re used to the idea of new technologies coming along. So I said, “Look, how can a new technology come along? It has to compete against a worldwide technology.” The point I completely missed is that when you talk about competing technologies, you have to consider the number of people working on them, and you have to be able, with a small number of people, to overtake one that has a large number of people solving its problems. The trouble with silicon and trying to beat it with Josephson was that silicon hadn’t hit any kind of a wall. We were just going to be faster and better, but it was still moving fast. I would never make that mistake today, but I sure made it then.
There has to be something that allows you to catch up and overtake it.
You’ve got to somehow overcome the handicap of numbers. If you could do something that’s so simple that you can just do it with five people and blow 3,000 end users away with it, that’s okay. But you’ve got to keep in mind that Josephson had plenty of problems, and we had a big group, like 90 people finally working on Josephson, but it never made it.
And this was something that you felt you were doing specifically because it would be a benefit to development?
Oh yeah. That was the other part.
We discussed about excellence in science and technology on the one hand…
Look at our strategy for logic and memory at that time, [it] was improve silicon, maybe do some gallium arsenide. But behind it, Josephson is the radical alternative, so it fit our program. And we were able to fund the radical program much higher level than normal because NSA was funding it, or a lot of the funding for it, because they wanted a super-fast machine. So that's a program that looked reasonable to a lot of people, but it wasn’t fundamental, if you understood more than I understood at that time, or very few people understood. Bob understood it, but he only understood it viscerally; he couldn’t express it. So that was probably my biggest catastrophe, and it was big. But we did a lot of good stuff.
So why don’t we talk about some of those programs.
Well we made great contributions to the printers, to silicon itself. The whole scaling of semiconductors. Bob Dennard is a huge name in that. We designed the RISC computer, which we called it the 801, what’s now known as RISC. Very important design. Scanning tunneling microscope, of course that’s one of the Nobel Prizes. It could see the atoms at the surface.
That was in the Zurich lab?
Yeah. And the other one was also in Zurich. You see, Karsten Drangeid was our Zurich Lab Director, a very modest guy. Gets zero credit, but he produced two Nobel Prizes. He was kind of more a leave it alone guy, but he left the right people alone. I shouldn’t tell you this, but this is a story. I very seldom lose my temper. You’re not supposed to. But I did once. I was visiting the Zurich Lab, and I knew about the work on the thing that eventually became the scanning tunneling microscope. It’s a probe. And I knew the guy that was doing that, who was an unusual guy, not a typical guy and he had no help, no technician. I said, “He has to have a technician. This stuff could be terribly important if he can really do it.” And the chap who wasn’t giving him a technician was the head of the Physics group, who was I guess Alex Müller. So I kept showing up over there, and every time they didn’t do it. Finally I really got fed up, and I said, “You have to do it.” A week later Alex resigned as head of the Physics group, and he went on to do the high-temperature superconductivity thing and win the Nobel Prize. But I think my outburst of temper did two things. It got the technician which enabled him (I wish I could think of his name; very nice guy), he won the Nobel Prize because he could make the scanning tunneling microscope finally and work; and Alex was so annoyed at being overruled by me that he went off and won a Nobel Prize.
Where did he go to?
He stayed in the lab.
Oh, he was there, just not as Director of the Division.
But he stopped managing. His strength wasn’t — He did a good move. But I always claimed that one burst of temper produced two Nobel Prizes.
Then you had Esaki.
Leo, yep. Leo was a very good guy, and I toured Japan with him, which was a very interesting experience. The real work that he got the Nobel Prize for he had done before he joined the Lab.
You mentioned Japan, and I know it’s a big theme in a lot of the articles that you wrote, so this might be a good time to bring it up.
Yeah it’s a good time to bring it up, but let me just pause and see if we’re missing anything really huge. [Break]
So we had exchange agreements with the JCMs, as we called them, Japanese computer manufacturers, patent cross-license, and these were renewed every five years. Part of the renewal process was that a team went from IBM and visited the advanced work of the Japanese.
This was in the ’70s?
That was in the ’70s I think we started that, yes. They visited our advanced work because that bore on the terms of the patent agreements, because they were paying us because we had more stuff. So they had a motive to show us stuff, to show that they were going to contribute, and we had a motive to show them stuff. It was very unusual. Usually if you went to see a Japanese firm, they’re not going to tell you what they’re doing in the future. So we went, and I was the head of the team, and I had Bob Henley with me and other people from the development divisions.
Sorry, Bob Henley’s position was what?
I think Bob ended up being an IBM fellow, but he was an important leader in the semiconductor side of IBM. But he was a real technical guy, and very experienced and very sensible, and I learned a hell of a lot from him.
So you went over for the cross-licensing?
Yes. But since we mentioned Bob, Bob is the person who told me something very significant. We were having one of these manufacturing crises having to do with semiconductors, and I said, “Well, Bob, our guys still have a lot of ideas about how to fix this thing.” I mean the research guys. And he said, “Ralph,” and I never forgot this, “in development, you never run out of ideas. You just run out of time.” And that’s a good thing to know and to think of. A lot of situations are like that where yeah you’ve got a lot of ideas, but you only have that much time. So we went every five years. The second time, the transformation was amazing, and the next time after that I’m scratching my head! How can they move so fast? There were a lot of reasons, but one thing was the emphasis on manufacturing. The development people, as I remember it, then they were all trained first in manufacturing, then in development, so they would develop something that could be manufactured. And their development cycle was much quicker than ours. You could caricature: when we did things badly, it was like we developed something and we threw it over the wall to manufacturing. Describing it that way was a way to criticize it. They didn’t. They had really good understanding of when they developed it, it would be manufacturable. So they did a lot of things right. They had very good government support, helped them with their R&D. I remember we saw a pilot line I guess at Oki better than anything we’d ever seen at IBM, despite of the fact that Oki had practically gone bankrupt a couple of years earlier. So I was very impressed with what they could do, but I also dimly remembered from my economics that I had taken both at Williams and at night when I was in the Navy in Washington that the idea sort of in trade was, well, if you lose a lot of industry to someone, then your wage goes down and you become more competitive. And the productions functions they dealt with and how much labor you had to put in to get something out were linear. But in semiconductors, if you get out of it, getting into it is a big deal. I mean it has economies of scale; you have to have a lot of knowledge. I mean you can’t get into it on a small scale; it becomes a big operation. So I thought those models can’t be applying here — if they beat us in this industry, we’ll never get back in! So I stored that away in my mind, and I said when I get out of IBM I’m going to look into this. And I did.
This is the source of your globalization work with Baumol?
Yes, all the papers I’ve written about that came directly out of that experience. And the funny thing is that after several years on this thing, it turned out that you could have that same problem without economics of scale. I’m still struggling to explain that one to the economists. They’re not happy with it. So we learned a lot. I learned a lot from watching the Japanese do that, and I was impressed by two things. You could, if you did it intelligently, make very rapid progress from a very low level, but it required a lot of investing; it wouldn’t be easy to come by. And second, we did not have the understanding in the US of how that worked. And China of course today is a different thing because it’s our own companies helping them over there.
You started writing articles then about this in the 1980s, which is not something that you’d done before.
Right. Yeah, it stirred me up.
And of course it becomes a very big topic of conversation, the Japanese challenge.
That’s right, it did.
I’m curious, because of course your main point is against the naïve notion, again, that research and government support of research [is the source of competitiveness], which is on the surface of it anyway, curious from a person who is a director of research, but given everything you’ve said it is much more understandable.
You understand, yes. First of all, I was old enough to remember the Second World War, and that’s been a big help to me, because we didn’t win that with research; we won that with manufacturing. A lot of people have rewritten that history. And there’s a lot of rewritten history that says that America became rich because the nations of the world were prostrate after the Second World War and we didn’t have much competition, and that's nonsense. We became the richest country in the world way before the War, and it was because we had tremendous manufacturing ability. We used more electricity per person than anybody, more steel per person than anyone, etc. And we won the war not because our planes were better. They were good, but the Germans had jets, but only had a few. Well we had a lot of these other things. We just drowned them. We built ships faster than they could sink them, and so forth. The size of the navies that we sent out, and I saw endless footage of that when I was in OCS in Newport, and you cannot believe the size of those navies that we sent out. Then it’s sort of been revised that science somehow did this. You know, it’s the atomic bomb and radar. And radar was good, but it was a British thing anyway. And America didn’t become rich because of scientific inventions made after the war; it was already rich. So it always annoyed me. I bow to no one my admiration for science. It’s the way a lot of things start, and a lot of the big things start. But your everyday country has to do all these other things. And the Japanese demonstrated that. At that time they weren’t innovating, but they had a totally different philosophy. They would show us something. They would stand in front of showing us this neat product, and they would say, “I got this idea from here, I got this idea from here, and I put it together and here, look at this!” The American view was, “It’s not my idea, and I have something better. But I [inaudible, possibly a reference to the “not invented here (NIH) syndrome”].
You also mentioned that they had less interest in market research, that they would create a specialty project.
That’s right; I remember that, they seemed to belong to the school “if you build it, they will come.” That was quite striking. Inkjets were an example of that. We had inkjets, they had inkjets, but no one could figure out quite what to do with the darn things. But they realized you could write on parcels and stuff with them, but it never was a big business. But that didn’t deter them. They were willing to just go ahead and do inkjets. It’s different. It’s not always a winning way to do things. I mean they spent a lot of money on the Fifth Generation Computer, and that was a complete fiasco.
I’m curious to what extent you see your thoughts about different topics as intertwined on a conceptual level. For example, that example of finding a particular set of technologies that can be repeated or expanded onto a certain scale by trying them out, and I noticed in another article you sort speak the same way about software, wherein there are certain lines of code that you can then pick out on sort of an empirical basis and then expand. So I wonder if you have a conceptual basis that you…?
Yeah, I try things. I’m glad you brought that up. A lot of the times, you don’t know what to do. You sit around and think you don’t know what to do. All your ideas seem no good. It’s much better to try something, and you try something and it’s no good, but in the process of trying you learn, and then you might go off in a totally different direction, but you’re more informed, and then you fail again. But you keep finding. It’s much better to try than to think, when it’s true that you don’t know anything. People tend to underestimate the complexity of the world, and there’s much of it that can only be found out by doing. That’s I think the basic thought there.
Perhaps related is the idea of competitiveness, wherein the Japanese would aim to take a field and then be competitive in it. And then when you talk about funding basic research, you say that the United States should be competitive, or should be at the head of their particular field. So I notice a sort of similar thematic similarity in that, and I’m wondering — one could almost say that there is a relationship to the idea of optimization. So I’m wondering if coming particularly from the field of operations research, and I know operations research has been linked to decision theory and management theory, and I don't know how more or less you personally have been interested in those connections, but I’m curious what your thoughts are.
I think my thinking is too simple to require that support. In other words, I would look at the science budget and how it is thought about pretty much the way I would look at the Research Division before I had my plan or vision of where it should go. So I asked myself the question what is the purpose of the scientific enterprise from the point of view of the United States? Just as I asked what is the purpose of the research enterprise from the point of view of IBM.
So then coming from your background as a mathematician, where do you think you pick up sort of your managerial understanding from? Is it just something you had with you, or were there particular influences?
Of course I was very involved with — I mean we had some contracts from the government. But I was on the Council of the National Academy of Sciences, I was on the Council of the National Academy of Engineering, so I see. And I was on PCAST, The President’s Council of Advisors of Science and Technology, and I was on the COSEPUP, Committee on Science, Technology [sic, Engineering], and Public Policy of the academies. So I see how everyone is thinking about the big things, and it wasn’t coherent, so I finally came up with… I got COSEPUP to consider something which eventually came out as COSEPUP report entitled “Goals for the Scientific Enterprise,” or something like that. The thought in that was the thought I think you were quoting, was that, well, we certainly want to do basic science because you never know what you’re going to find. My view has always been yeah we do do basic science, and we do non-basic as well, both. So I basically divided the science into science that had proved that it had consequences to a larger world outside of science, and that which had not. And the goal of our scientific enterprise should be always to be among the leaders in the things that have not proved anything, and the leader in the things that had, but we should be in everything, either as the sole leader or as one of the leaders. That was basically it. And that’s a fairly simple thought, actually. So you won’t miss anything, and you’ll help make new things happen. You want to be right up in the top group, but you don’t have to be the absolute leader. But in something which has already shown it has an impact, like solid-state physics, and then you want to, if possible, be ahead. So that was the idea.
I think I got a similar impression when speaking with Marc Brodsky. He mentioned particularly he actually learned quite a bit from you. It’s always seemed a little curious to me but probably understandable that people who come from scientific and mathematical backgrounds, when they enter management — Brodsky discussed when he came to AIP how he would have three bullet points for the organization, and that this would be sort of your governing strategy. It always struck me how night and day it is coming from the rigors of science to just laying out a series of bullet points as a managerial strategy, and so I’m wondering, it must be sort of a night and day attitude towards things, that management is one thing and science is another.
I understand what you’re saying. I think that may be for some people. But you see, I didn’t come from mathematics, like I was a born mathematician and I just always wanted to do mathematics. I was much more a born thinker who was constantly trying to understand different things, and happened to be in mathematics because I thought that that would explain a lot of things. So for me, I’ve never felt that transition. I try and think sensibly about both of these things, and I never felt the difference. But I think you’re right…
Different sets of problems, but similar mental process.
Exactly. Now that’s not always the case with people, so I think your observation is also correct. For a lot of people, it’s really different. But because it wasn’t like I came from mathematics, it’s rather I came from something else and did mathematics.
[Discussing the remainder of the interview] I mentioned at lunch that we have been working on this project to create database of physicists, and we’ve looked at some of the people who are at IBM, and of course this is limited to the people who are in our database, but it is a good place to start, the people who you would have been around. So I think what I’ll just do is I’ll throw names out there.
Are they physicists?
Mainly. They might venture into engineering or whatnot. But we could start with Lewis Branscomb, for example.
If we’re going to name people’s names, I want to make sure we don’t leave out some who weren’t physicists who were very, very important.
Yeah, that’s incredibly important, because it’s a semi-arbitrary list, but it’s a good start.
John Cocke — he really is the father of the RISC design and of many other things. An eccentric person, very kind hearted, thought about computers night and day, and knew just more about computers than anyone could imagine.
Where had he come from?
I don't know, which planet? [Laughter] I think maybe he was at Duke, but I can’t tell. But the thing is, he was just kind of an inspiring person. It wasn’t just that he knew every technical detail. That wasn’t the point. But he knew what was important. He was a wonderful person. You named Lewis. Of course Louis was the Chief Scientist for many years, and he was a very helpful sort of person.
What’s the relationship between the chief scientist position and the director of research?
It varied over time. [Drawing on paper] For instance, when I was there we had the CEO, that was Frank Cary, and the Director of Research was me, and then the various labs and stuff. Then we had a staff position, which was the Chief Scientist.
And that’s coming off of the CEO, I see from your chart.
That’s right. So he reported to the CEO to give him input or whatever. Now actually my last two years at IBM I became Senior Vice President for Science and Technology (I didn’t like the name Chief Scientist), but the division then reported to me. So it’s varied a lot over time.
So the new director was John Armstrong after you?
Yeah, John reported to me, and then he became — I don't know who John reported to.
Director of Research, then Vice President of Science and Technology. [Armstrong’s position]
But over the succeeding years it was all over the place. I’m pretty sure that Jim [McGroddy] reported to the CEO. Probably John did. But in recent years there have been times when research was somewhere down here several layers. I don’t think that’s good. But anyway, during the period when Lewis and I were both in the company, that was the arrangement.
So what would that role be...?
What was Lewis’ role? You’d really have to ask him. But he did a great deal of outside relations. He was very visible from the outside world, things like that. So crudely speaking he was Mr. Outside, but I ran the inside. But I didn’t report to him and he didn’t report to me; we both reported to Frank.
It was a separate sort of sphere.
Oh poor Praveen, you know he died too. Well Praveen was a very good guy. For a while he was the head of the Physical Sciences Department, which was one of those — so when I’m talking about my directors, he was at one point my Director of Physical Sciences. Earlier than that John Armstrong was my Director of Physical Sciences. Praveen had a lot more vision than most people, but he also annoyed people sometimes. Eventually he morphed into the Director of the Brookhaven Lab, where as far as I can tell he did a lot to save that lab from being really essentially destroyed by the Department of Energy. But in that process he annoyed an awful lot of people at the Department of Energy, but he saved the lab. Now I’m getting that from him, but I believe it. He and I always talked very frankly. Then he returned to IBM and unfortunately he died. He was a very good man, in my opinion. Outstanding. Had a much better grasp of the big picture as well as being a good physicist.
You mentioned Arthur Anderson.
Art Anderson, yeah, my predecessor.
But he stuck around for quite some time.
Yeah, but he was doing other things. You see, Art really wanted to advance in the company, and so when he left research he became the head of the Development Division that made disk drives — very important job out in California. That’s about all I can tell you. But he was a good person to work for. I liked working for Art.
He was Director of Research for three years, ’67 to ’70?
That sounds right, but I couldn’t testify to that. I think Rolf Landauer was his Assistant Director.
Yes, that’s what I have. How about Rolf Landauer?
Rolf had a lot of good insights. He wasn’t easy to deal with, but you learned how. His quick reaction was usually somewhat negative, but by 24 hours later he would have thought it through and could give you a well thought out and reasonable response...
I have it that he was one of the fellows.
He probably ended up as a fellow. But he was Art Anderson’s right hand man.
Oh okay, in that period.
He was a good guy, basically.
Ah, Dean. I guess Dean must have been the head of the Physical Sciences later. Another guy who was a very good physicist. I think he had more problems with the management side. But he was a very straight shooter.
He was working under you while you were Director?
He never reported directly to me, I don’t think.
But then you worked a great deal with John Armstrong.
Yes, I worked quite a bit with John and with Praveen. I think those were the two main ones.
We’ve talked about Jim McGroddy.
Oh yes. I think maybe Jim was in there too as head of Physical Sciences. Maybe not.
I have him as Director of the Semiconductor Science and Technology.
Ah, that’s it. He wasn’t in the Physical Sciences; he was Director of a different thing, the semiconductor part.
Then as Vice President of Logic and Memory in the ’80s.
Yeah, that’s right.
Those are the primary names that I have.
Those are good names.
But if there was anyone else that you remember?
I would like to mention Karsten Drangeid.
Okay, I know nothing about him.
Everyone knows nothing about him! Long-time Director of the Zurich Lab gets no credit for anything.
Oh, right, you were mentioning him earlier.
But he did well.
Those labs reported to you?
Yes. And another person who was very effective as Director of the Almaden Lab was Frank Mayadas. Frank worked for me in that capacity, then he became the secretary of the CMC. That was really the corporate management board. They put promising people in as their sort of secretary to that board. That consisted of the CEO and the top few people. And he did a great job at Almaden Lab. He is retired. He came down with me to — Both Hirsh Cohen was a very important figure in the Math Department, and Frank Mayadas eventually joined me in the Sloan Foundation. Both very good guys.
One figure I forgot to mention from your earlier day before Director of Research is T. C. Hu.
Oh T. C.! T. C. was one of the people in the Math Department, and he and I worked together on network flows and things like that. We even wrote an odd paper about sets of points in a plane. T. C. is a very original thinker, and very underestimated because he doesn’t do — You know, you can make a career in science by just doing the next thing. But it’s good to have people who can jump into different directions. He was that type of guy. And Frank made a great contribution at the Sloan Foundation because I wanted to have online learning. This is before there was a commercial Internet. I felt they could give real degree courses and so on. And I tried to get one of my program directors to do it and he couldn’t do it, and then Frank came down and he made it work. And so we really made a huge contribution that couldn’t have happened without Frank.
That was one of the other connections that I noticed from reading your papers, is sort of the critical attitude, again it’s the idea that there is a flaw in United States science and technology education, which is another one of those ideas that is against the grain of what a lot of people would instinctively say, but that education of particular skills, there is a lot of room for that, and you mentioned specifically the prospect of distance learning as a way to learn, and so that was something that carried through to your work at the Sloan Foundation.
Well I don't know whether I thought about doing distance learning much when I was at IBM, but I just saw the potential that you could do it, because I understood the technology. But I think misunderstandings about the US educational system have persisted. People believe that our K-12 system is handicapping us against… our children are not going to be competitive in a globalizing world and all sorts of nonsense like that.
The idea of what they used to call scientific manpower has been a rallying call for 60 years or so.
Yes that’s right, there’s always a shortage of them. You know, there never was a shortage, and I don’t think there ever will be. We fought the shortage idea quite hard; I haven’t heard it for a while. But all the time, we have these long backups of people with PhDs doing post-docs because they can’t get an academic job, and at the same time everyone is talking about the shortage. It’s not a misunderstanding. Companies would rather always have lots more people applying for jobs. But there is no shortage, and our educational system is a lot better than it gets credit for because they don’t realize that the average is pulled down by having an underclass. China and India, for example, don’t even take the exams because they score around zero on the average.
When you write articles about these things, it’s clearly in response to sort of a general conversation that you see taking place.
Yes, it is.
And I’m wondering who you have in mind as intended audience for articles of that sort. Do you think these are things that are well understood by people who are on the inside of industry?
We’ll have to take a particular topic and I’ll tell you who I want to aim at, because it’s not always the same.
Let’s return to the Japanese challenge, and whether or not particular industries…
I think that was aimed primarily at academia, but secondarily at companies, because what I was saying to academia is you want more money, but it isn’t more money in academia that is going to make us compete with the Japanese. They had a somewhat egocentric view that they’re doing everything, and I’ll come back to why they had it in a moment. It was also aimed at the companies, because they didn’t understand that it was the short product cycle that gave leadership, and if the Japanese could do it, why couldn’t they? There was a very good study done at MIT, what was it called, it resulted in a book that was basically about the Japanese auto industry, which was a very good book. Those were the circumstances that led me to found what became the industry studies program [here at Sloan], so in the end here we had 26 or so universities, which one of them would have the auto industry, another would have some other industry, so that they would develop ties with industry and have some understanding of these things. But with that stuff about the product cycle, I was trying to explain that it isn’t academia’s fault; it’s that we’re not focusing on development as opposed to science, and the companies aren’t doing that great a job. But why they think so I can explain in a nutshell. When I went to IBM, I thought the research was the whole thing. Then I had a lot of contact with development, and I noticed that development thought, well, research you just have ideas. We have to make it into a product —that’s the hard part.
Let’s talk specifically, you mentioned you had strategies for moving people around within an organization, and when I was speaking to Marc Brodsky he mentioned that he had worked directly for you as a fairly junior person because that was a program that you had where you had somebody who would come work at your office for a year.
Yeah, we had that. But let me finish the thought, because I didn’t finish it. And then I noticed those guys didn’t think research was the center; they thought that development was the center. And that one day I was going through a manufacturing line out in San Jose where they made disk drives, and we’d have very good relations with their development team, so I remarked to my guide, “You know, you’ve got a very good development team here in San Jose, making disks.” And he looked at me scornfully, and I’ll never forget what he said. He said, “Ralph, anyone can make one of anything.” His view was, hey, those guys can make a design. But make 100,000 of them? Hey, that’s totally different, and that’s what you have to do. His view was manufacturing. When you start dealing with the sales people, they don’t think the product matters that much at all. They think it’s the relationship with the customer, do they trust you. So each group automatically thinks that it’s the center of the universe. So it’s not surprising that academia does — so does everybody else! But academia is heard from more. If development was heard from, you’d find that they were the center of the universe. You wanted me to talk about something else, how you educate people in an organization?
Yeah, and to what extent can you shuffle people around to different parts of an organization.
You can, but I’m not sure I did that more than anyone else. I appreciated the value of it. I think I was more at getting people just to think straight.
Shall we move on to your transition to Sloan?
Is that part of the story?
On the phone you mentioned we might go up to the present day.
I think the main thing is that I picked up in the course of my IBM experience, and whatever else, I had a whole lot of things I wanted to get done, and a lot of them influenced by my experience in IBM about science, contact with industry, how you compete in the real world, all that stuff, and the possibility of online learning. So when the Sloan Foundation was looking for a president and they asked me if I would be interested…
Had you had dealings with them before?
No. But in those days, people did things differently. They didn’t hire an executive search firm. They just asked their board who do you think would be good. Frank Press was on their board. Frank Press had been the head of the NAS, and he thought of me. So it was more like that. So I said to them, “Well, I’m interested, but here’s what I want to do. Do you want this?” and they said yes. So it was a great thing for me because then all of a sudden I had the freedom to pursue all these things with money and get people to do them. So it was a very good transition in that sense.
So you were Director of Research for about 16 years?
18 years, and then two years I had that senior… that other position.
And then you were here for 18 years as president?
I think I read somewhere that you had passed up opportunities to move along within the IBM organization. It might have been in Piore’s book.
Yeah, I think that’s true.
So is it something that you consciously tried to do, to have fairly lengthy tenures in each of these positions?
I don't know how you go about having a lengthy tenure! [Laughs] That’s not usually up to you. But no, I felt I was in about the right place. When I was at Sloan I was offered a lot of jobs, too, but I thought you couldn’t beat the independence and money.
I suppose we’re probably getting along in the afternoon. We could talk briefly about globalization, since that’s your current interest, and how you got back into that, or how it extended from your other interests.
It really is just an extension I think of my belief that industries had to be competitive. So when I left IBM, I took up my original premise that there’s something missing in economics, and I was fortunate in having two old friends. First Herb Scarf, we’d been in graduate school together at Princeton, who taught me some economics. And so I started to do research on the economics of trade, and my suspicions were right — there were things to be done. There’s a very good book by a person who was famous in his day, Walter Lippmann, called Public Opinion. It was written around 1922. It’s mostly about the First World War, things like that, and how public opinion is shaped and all that. Of course it was shaped in those days by newspapers and stuff. But it’s very striking to read it, and to see how perceptions relating to fighting the war and keeping people’s morale up and keeping them fighting, the public perception of what was going on and what was actually going on — light years apart, just like now. So the perception, I mean all these people are talking about free trade. All economists line up solidly behind free trade. The idea that we have free trade with China, come on! The way their economy is run, do we? We don’t. And so on. So it’s interesting to me to see the gap often between reality in the areas where I know about it — sometimes I certainly don’t know what’s reality — and what people think. And it’s a very interesting thing that I think could bear a great deal of investigation. But the simplest explanation of these gaps usually is that you hear what you want to hear, and if you’re in a position running a company and you want to expand worldwide, even if it means moving your stuff, and you’re very, very vigorously rewarded for that, you will be a believer in free trade. And if you’re running a labor union and your guys are losing their jobs because their jobs are disappearing, you will have doubts about free trade. That’s a very trivial thought, but most of our political thinking in the end is shaped that way, I think. Walter Lippmann’s book is very instructive.
I guess I have just one last question. You’ve also gone back to some issues in the cutting stock problem, I think, with Ellis Johnson down at Georgia Tech.
With Ellis it wasn’t the cutting stock problem. I did the cutting stock with Paul Gilmore. I did network stuff with T. C. Hu. And Ellis I did later stages of integer programming.
So you found that coming back to this, one can.
Oh yes, you can.
What happens kind of in the interim? Had you been keeping up with it?
Well, first of all, the first few years I was Director of Research I wrote several papers, but then I think you can do what you want to do. I became very interested in what I was doing. Recently I actually have made some contributions to integer programming, and it’s not that hard. I think it’s a question of what you want to do. I’m not doing integer programming right now because I’m doing the economics thing.
To me it’s just a question of… I mean sometimes when I see that you can sustain a line of research in linear programming or in inventory theory, or anything of that sort, that they start from very basic propositions, generally in the 1950s, but that these are lines of research that continue today. It’s not something that I’ve investigated in any great detail as to how one sustains a line of research for 40 or 50 years.
Oh, you mean how people live in it?
Yeah, or how it develops for that long, and what sorts problems come up.
How the thing develops or how the people live in it?
I suppose both. I’d thought the first, but…
You see, it’s pretty easy for people to stay in the thing they’re good at. Most people are very reluctant to get out of it because they know all the wrinkles in it. The idea of starting over in something they don’t know, you need a lot of self-confidence to do that.
But ones doesn’t exhaust the subject matter.
Oh, you see, from an outside point of view you may—the subject matter may appear exhausted. But from the point of view of your peers in the field, oh no!
There’s always more constraints, or…
Yeh, exactly. Like when I said I didn’t really want to be writing papers on nonlinear differential equations for 14 people, that was the picture I had. There will always be something to do, and they make a contribution. And after all, most scientists don’t make big contributions, but they make a contribution, and they lead a decent life, and they have lots to think about, so it’s not a bad way of life.
If you have anything further to add please do, but I feel I’ve kept you plenty long.
No. If you have any further thoughts, I’m always happy to talk with you. And I think you’ve done a great job. You made it very painless for me!
I appreciate it. I’m glad. Thank you.
 This was “A Picturesque Tale of Progress” by Olive Beaupre Miller (REG).
 The Navy equivalent of the Army 2nd Lieutenant; has a single gold stripe on his sleeve.
 It was 17th Street.
 Gerd Binnig won the prize with Heinz Rohrer.