Home > Articles > Software Development & Management

An Interview with Watts Humphrey, Part 9: Family History, Phase Plans, and Labs Around the World

  • Print
  • + Share This
  • 💬 Discuss
From the author of
In this transcript of an oral history, Grady Booch interviews SEI Fellow Watts Humphrey. In Part 9, they discuss Humphrey's family history, the advent of charging for software, creating phase charts, and running a global business.

This interview was provided courtesy of the Computer History Museum.

See the entire interview.

Family History

Booch: Your wife, who I was delighted to have on this as well, too, pointed out that I had forgotten to look way back in your past, and apparently there are some fascinating people as we reach back down the chain of DNA that became Watts, so tell me a bit about that.

Humphrey: Okay. Okay, well, first I'll talk to my Humphrey grandfather. My grandfather, who was the first Watts Humphrey, my father, the second, I'm third, and I've got a son who's the fourth (he's got two daughters, by the way, so the line stops). But my grandfather Watts actually was an officer in the Civil War, which is kind of hard to believe -- the Michigan Cavalry. He was a lawyer in Saginaw, Michigan, but he was an officer in the Michigan Cavalry. I don't know what battle, but he was wounded by a chain and ball canon. He was riding his horse. The horse was taken out and his leg was rather badly damaged. He came to in an army tent and the surgeon was approaching him with a saw. What you may not know is, in that timeframe, fatalities from amputations were close to 100%.

So he asked the doctor, "What are you going to do?" and the guy said, "I'm going to amputate your leg." And my grandfather said, "No, you're not." And he said, "Oh, yes, I am. I have got to amputate your leg." Well, my grandfather pulled out his revolver and said, "No, you're not." <laughter> So they did not amputate his leg. It later turned out to be his good leg. He had a problem with the other leg. Without that, we wouldn't have had the whole Humphrey clan. So that was the grandfather Humphrey story. His second wife was my grandmother. He married again when he was older, but it's hard to believe that my grandfather would have been in the Civil War. So generations go back mighty far and fast.

Booch: They do, indeed.

Humphrey: My father's elder brother was George Humphrey, and George Humphrey worked for the M.A. Hannah mining company, and he ended up being the first secretary of the treasury under Eisenhower. That was when Eisenhower was elected in '52. He was secretary of the treasury then. I remember once going and visiting them in Washington and my aunt said, "Would you like to go see your Uncle George? He's in his office." And I said, "Sure." So she had a car pick me up and it was U.S. license plate number two. And so here I was, this little guy, college guy in the back seat, I was still in graduate school, and every time we stopped at a light or anything, everybody was looking in the windows. They didn't have tinted windows. But he pulled up between the west wing and the Treasury and there was a little door there and he said, "Why don't you just go in there?"

And just to tell you how different the times were, there was no guard or anything there at all. It was just a door there. So I went out and opened the door, it was down a couple of steps on the Treasury there on the side closest to the White House, so I went down a couple stairs, went in and went up these stairs and knocked on a door and a voice said, "Come in". I walked in and I was in my uncle's office. No guards, no nothing. It was the view out there over the Mall and the White House and everything and we sat and chatted for awhile. A wonderful guy, but the security and the whole situation was so different. But, in any event, he was a great guy. At home, he drove a Chevrolet. They didn't have fancy cars or anything. But he, as I say, was the first secretary of the Treasury.

On the other side of the family was the Strong family. My mother's maiden name was Strong, and her father was Benjamin Strong. He was a financial wizard. Turned out, he'd been involved with a whole lot of stuff, starting way back in about 1900. 1909, he was involved in a crisis with the Morgans and everything, a banking crisis. He actually never went to college. His older two brothers both got degrees, but the money ran out. He ended up, in 1913, when Woodrow Wilson signed the Federal Reserve Act, establishing U.S. Federal Reserve system. My grandfather was the first governor of the New York Federal Reserve Bank. That was the number one bank, and so he was involved in all the crises and the war, financing for World War I, the reparations after World War I, and all of that.

There's a book written about him, Lords of Finance. Actually, during his travels he contracted tuberculosis, and so he suffered terribly from that, and he actually died when I was one year old. He was 55. As they say, if he had not died, the whole financial system results might have been quite different because, in the U.S. government, there was no strong man to take over. He had been a strong man, running the Federal Reserve and there was a jurisdictional dispute between the Federal Reserve governors in Washington and the board that he chaired in New York. As long as he was the chairman, he kind of overrode everything. Everybody sort of backed down and he did it. He knew what he was doing, he was pushing through, he was aggressive. So you don't know where I got it, right? But, in any event, he died right at the crucial point in 1928, October 28th.

That's when the banks collapsed, and there was nobody there to do what Bernanke and these guys are doing now, because the people on the board in Washington were all political appointees and knew nothing about banking. All they knew about was you had to have balanced budgets and you wanted to make sure there wasn't inflation. So they were back to traditional controls and they did essentially everything wrong. So that's one of the key things. So it was a tragedy that he wasn't there, but he died of tuberculosis and complications that came with that. So there's a little bit of a family anecdote.

Booch: Amazing. Banking is in your blood in a variety of ways, isn't it?

Humphrey: I guess so.

Booch: Your ancestors from sort of the management of it and you from the automation of it.

Humphrey: Yeah. That's right. Connected in multiple ways.

Booch: It is. Well, I'm sorry, go ahead.

Humphrey: I just want to say, in returning to the story, I had gotten to where we were talking about, I believe, Bob Evans coming in to the re-organized divisions. And basically my career running programming was winding to a stop, and I need to drop back in here with appropriate dates for things that had happened before that I probably should mention.

Booch: Absolutely. So let's rewind to around '64, as I recall, where some of those anecdotes begin.

The IBM Board Meeting on Program Pricing

Humphrey: Right. In 1964, I was on corporate staff. I was Director of Systems and Applications Engineering. At that point, the IBM 360 was just getting started. I think it was '64/'65. RCA was making noise about their system. They were going to make a system compatible to the 360. RCA Specter or Spectrum -- I don't remember what it was.

Booch: The Spectra 70 is, I believe, what they called it.

Humphrey: Spectra 70.

Booch: They announced it in '64.

Humphrey: Right. Okay. That was the big concern that IBM had -- that RCA was going to steal a march on them, and so I was put in charge of a task force to figure out what do we do about it.

Humphrey: So we went through that. We had a technical study. Had the lawyers involved. There was a great deal of concern about antitrust from the legal community, and I remember we had basically no notes, so it was all talk and all on chalkboard, didn't have many whiteboards then, so we were all talking about it like that. The lawyers and everybody concluded we had to present the story to the IBM board of directors. So I've forgotten who the top lawyers were at this time, but they advised me that I could make a presentation but I couldn't use any charts and the board members couldn't make notes.

Booch: And, of course, when we speak of charts back then, you're talking the old overhead foils, if I recall. Certainly predating the days of PowerPoint.

Humphrey: No, these were actually paper flip charts.

Booch: Oh, my goodness. Even before foils. Okay.

Humphrey: Right. And so we'd usually go in with great big wads of flip charts, like three and four inches thick and backup charts and stuff, but I couldn't use anything. I had to just talk. And so I had 13 points to make and I remember going in. We were talking about what the various choices we could do fighting this and clearly one of them, which turned out to be the only one that appeared to be viable from all sides, was we would have to price our programs [i.e. charge for software].

So I arrived in the boardroom to give this presentation and I told them that I had 13 points to make, that I would use no charts. I had a blackboard brought in so I could write with chalk. I advised them that they should take no notes, so I would have to go through it this way, and we'd have to just inform them but we could not have anything in writing. Nothing about this meeting has ever come up since. Not in any of the depositions, nothing. It never showed up so I don't think anybody knows about this meeting other than the people who were there. In any event, I did go through the whole thing about what each of the choices were and what the consequences were. I did it all from memory and I did get all 13 points. I was sort of really kind of worried that I had missed one or something, but I got them all. I remember one of the board members, at the end applauded me. I got them all. But at least we got across the point there, with a rather complete background, as to what the consequences were and the only real choices. People really could make compatible systems. We had no real way to protect it.

 Basically, it said we had to start thinking about pricing programs. As the cost of programming went up, there was no viable alternative. That was in 1964, so that was very, very early. I think it's just as well we did it because the whole community was sort of ready, and this was something we ultimately had to do. The next anecdote, if you remember, I talked about early on, I'd had the crisis with the FAA and then the following year, this was in November of '64, I had the timesharing crunch where the 360 was out and we had the problems with Bell Labs and all of that sort of thing. Well, at that point, we were going ahead and developing the timesharing system. We'd gone through and had the upgrades.

This is after we had the GM battles and all of this sort of thing. And there were the debates about how things were going to be organized. At that time there was the division president of the development division -- I believe was John Haanstra -- and he’s the guy who later went to GE. But he was the development division president reporting to Dick Watson. And I was reporting to Dick Watson. And everybody was pushing to somehow get the timesharing stuff put under development and pull it out of corporate staff and move it where it ought to be and get it in place. And I remember meeting with Dick Watson who was then a Senior VP of IBM. And he and Vin Learson were sort of parallel VPs. He had the development division and Learson had marketing.

And so Dick was Tom’s younger brother. I was in this meeting with him talking about organization, and I made some comment, “Well, if I were you here’s what I’d do.” And Dick Watson absolutely exploded, and I couldn’t figure out what it was. I used a figure of speech. Well, the “if I were you” struck him as sort of presumptuous on my part. Here’s this guy who thought he could be me. And I thought, my lord, this guy had a terrible image problem. And he obviously did. He drank quite a bit, but he was a brilliant guy. He became U.S. Ambassador to France in Paris, and rather shortly after this he moved over there. But I really got on his bad side by that meeting. And so he and I were not on good terms at all. But he actually left IBM and went over to Paris. And I remember the stories were that when he called a meeting for 8 o’clock in the morning, at 8 o’clock they’d lock the door. And if people weren’t there at 8 o’clock in the room they couldn't get in. And so he really tightened up the way the embassy ran in Paris. And he got into all kinds of flaps with stewardesses and stuff. And he subsequently died from a fall. He fell downstairs. The rumor was that he was drunk. But it was sad because he was quite a guy, but he was really badly overshadowed by his older brother Tom, but a very capable guy. And really a very nice guy except for the problems of his ego, I guess.

Phase Plans

The next anecdote is when I took over programming in 1966, if you remember, we were in a crisis shape and we got everybody to put together plans. But I had no way to track progress.

Earlier on I had studied a lot of the IBM procedures to figure out -- this is when I joined the company -- what they did. And they had some earlier procedures from the old systems development division -- I forgot what they called it now -- but it was the large system division, one of the hardware divisions at the time. And they had something they called phase plans for development. And it was a very orderly procedure that they had in the procedure book. People weren’t really following it much, and I didn’t know that. But the procedure book was rather nicely laid out with about six steps on how you actually develop a program or a project.

So I adopted that for programming and required everybody to develop their plans against certain phases. So you have an early feasibility part. Before that, you get feasibility ideas and then you get real project funding. You could name a project. And then you had to get certain stuff before you could announce the program. So there were various steps you had to go through and things you had to accomplish before then. Not a lot -- it wasn’t too much -- but it did require certain things and certain approvals and certain reviews at each step. So I put that in place. And so the minute we got our schedules in, I had every project list their planned and actual dates against each standard process phase and track it. They’d track where they were, which ones they had finished, and then when they were projecting a date, they’d have a projected versus the plan dates and that sort of thing. We put that together and it was pages of stuff. So we had these projects, as I said, 4,000 people and I think I had 15 development labs at the time and dozens of projects in every one.

So there were hundreds of these things with all of these phase dates. I put that together and I remember early on in one of my reviews with Frank Cary who was the general manager over all development and manufacturing. He was a Senior VP at the time. It was before he became president and chairman. But I remember going in and reviewing it with him. And he was really impressed. You could see exactly where everything was, and we could track it, and we could flag things that were late, and it was marvelous, and it worked. And we did not have sort of wild guesses in there. We were putting in actual progress and dates. Also at that time in 1966, I was, as I say, Director of Programming and I went out to the IBM annual meeting -- stockholder meeting -- and it was near San Francisco in that year.

And so we went out there and at the meeting, I remember Tom Watson stopped me at some point and said, “Watts, what are we going to spend on programming this year?” And our budgets had been growing like 40 percent a year. And I think the projection was it would be 60 million -- the early projection -- but as I had gone through the numbers, it really had looked like a lot more. And I told him, I said, “Tom, I’m afraid it’s going to be close to $100 million.” So he actually referred to that in his remarks. He said, “I talked to Vin Learson about what we’d spend and Vin said we’d spend about $40 million. And then I talked to the division president and he said he thought we’d spend about $60 million. And I just talked to the director of programming and he said it was 100. I hope it doesn’t keep growing,” or something like that. But they gave us what I needed.

Booch: If I may ask a question sort of about the nature of programming back then, there was no sense of outsourcing back then. So was it fair to say that all of the programming staff was in the U.S.? Or did you have international groups doing it? How large was the average team? What kind of tools did they use? Can you relate upon that a bit?

Humphrey: Sure. The big lab was in Poughkeepsie. They had nearly 1,000 programmers. They were developing OS 360. And they had the control program work. They had everything. It was basically there. That’s where the architecture work had been done. That’s where the offices of all of the executives were. And the problem we had was that IBM had a ground rule that you couldn’t get more than so many people in a geographical location. So we had headcount ceilings in every location. And Poughkeepsie had been growing and was right smack against the ceiling. And we were pushing it, every inch. And so the location manager kept kind of cutting back the amount of programming we could do in Poughkeepsie.

The Data-Management Interface

And so a lot of stuff had been moved out. And I remember one thing we had set up (and this was ahead of my time), they had moved the whole data management stuff -- file support, all of that -- out to San Jose where the file hardware development was done. So we had the IBM San Jose software group which was growing, and that’s when they were building a big software lab out in San Jose, and I’ve gone fuzzy on the name of it. But in any event we had built a growing group out there, I’ll come back to a comment on that in a minute.

It was an IBM software lab. It was built specifically for the software. It was really very well done. They set up with a whole series of wings where you have projects; everybody is sort of grouped together. It was very well done and it was a fine operation. We also had a group in Time Life. We had a lab in Kingston and New York City. One in Raleigh, North Carolina. We had one in Rochester, Minnesota; Boulder, Colorado; Tucson, Arizona. We had a small lab -- more researchy in advanced development -- up in Boston at MIT. And Tom Watson had been a bug on international. So he had required us to set up labs in Europe. And they had done all of this.

This had already been done before I took over in ’66. And so we had laboratories in England -- Hursley, England -- that’s where they developed the PL1 compiler. We had one in La Gaude, France. I don’t remember what they had there. It was some communications stuff or something. They were smaller. The European labs were like several hundred people. The U.S. ones -- the big ones -- were several hundred up to a thousand. We had a very small research sort of lab in Vienna. We had a fairly big lab in Boeblingen near Stuttgart. And a modest sized lab up in Stockholm. So we had labs all over the place. And this is before international -- even telephone communication was shaky. And so I spent a lot of time on the road going meeting with these various people in terms of seeing what was going on and that sort of thing. So I had multiple trips to Europe and with seven kids, my wife would go with me and they would spend the week going around to various cities. We’d take two of the kids at a time. So I’d meet them on weekends in places like London and Paris and Vienna and Rome. So we got some advantages of it, but we had stuff everywhere. It was really horrendous. And it was a terribly difficult collection to manage. We had to keep sorting things out.

Booch: Was there anything in Asia around this time, yet?

Humphrey: No. It was all in Europe. And the issue that I had when I took over and I began to look at the plans, we had FORTRAN compilers everywhere. Everybody was developing all kinds of random stuff. So when we started getting plans we sort of -- if you wanted a sort developed you find somebody that had some people available. And so we stamped that out. I said “No, we’re going to have missions.” And so we put the sort lab -- we originally wanted to put it over in Stockholm. It ended up, actually, with some storage experts in Poughkeepsie originally and we had to bring it back. That brings up another yarn I’ve got to talk about on lawsuits. But in any event, we had all of these different labs all over the place. And typically, as I said, the big lab was Poughkeepsie with about a thousand [people].

Kingston was about 500 or 600. Boulder might have been close to that. The others were a bit smaller. And San Jose was also a big one, 500 or 600, something like that. And so that’s where we were. And one of the big issues we had was how do you pull this stuff out because when they had moved the data management work out of Poughkeepsie to San Jose, the system was not designed to have a totally separate data management system. And so there were all kinds of interconnections. How do you define the interface between what’s the data management and what’s the control program? And no one could agree on that. So they came to me with a big battle. And this was like in the first few weeks I had taken the job and they had these warring camps. They didn’t have an answer. They had a big meeting with me in my White Plains office with a whole bunch of guys from each location, and they couldn't agree. So I sent them back. I said, “Come back with a proposal.” I said, “Put together a story. I can’t make this up.” So they did. And they went back and they came back with a meeting a few weeks later and they had two proposals: the proposal from San Jose and the proposal from Poughkeepsie, and they presented them both to me. I was really annoyed because they had to make a decision, we couldn’t just sit there. Neither proposal satisfied the other one.

And I said, “Here’s what we’re going to do.” So I made up the answer. I knew least about this of anybody in the room. And I said, “Look, here’s what I think you need to do.” And they had gone through all of this stuff and so I carved up the baby. And I said, “You either go work that way and make it work or agree on a better way and tell me.” I said, “We can’t sit here forever so that’s what you’re doing to do.” So I designed the interface. I’m sure they changed it quite a lot and it’s probably way different now but I couldn't believe it. I knew the least about it in the room and I’m the guy who actually came up with the answers. But there we are.

Booch: You mentioned FORTAN popping about around this time. Did you have any interactions with John Backus as well?

Humphrey: I did not, actually. The only way I got involved with John, I ended up buying a house he owned. But I never had any connection with John, which was my loss.

Booch: And this was in your march to get bigger and bigger houses as your family was growing, it sounds like.

Humphrey: That’s right. That was our third house in Chappaqua, a big castle that John Backus had owned. It was a gorgeous house.

Booch: Well, that’s a connection unto itself. It’s good to know that you owned his house, cool.

Humphrey: I did get to know Jean Sammett who was a researcher at IBM, a language researcher. She actually worked with me at Sylvania earlier on, so I knew her from there, so I’ve known her all of this time. But in any event, so that was the story with the programming, the lab, as I say my role, and why we included planning.

  • + Share This
  • 🔖 Save To Your Account

Discussions

comments powered by Disqus