This interview was provided courtesy of the Computer History Museum.
Skiing with Tom Watson
The Endicott Lab
Humphrey: I agreed to take the
lab in Endicott. He said I would no longer be a division vice president,
because, you know, "We've abolished those jobs. You could stay here and in
that job. I don't know what you'll do," but et cetera.
And so he was quite direct on what I ought to do. So I agreed to take the
Endicott job. And so I got a few stories about what happened there when I ran
it. It was up in
So I decided that we had to re-organize the place. We had an awful lot of work to do. I had all the intermediate systems, the printers, the OCR/MCR stuff. It was a very important lab and we had several billion dollars worth of products. The mid-range computers were an extremely big market. It was billions and billions of dollars in hardware. So that's what we did, and it was about 2,000 engineers, and we were working with a big manufacturing plant there in Endicott. And so I put together a committee of the line managers, the people developing the products. And I said, "What I want you to do is to go through the list of all the departments in the lab and tell me which ones you need to do your job." And so they did. And they went down this list. And there were a whole mess of them that they indicated that they didn't know -- "Why have we got those? What are they doing? We don't need them." And so I started down the list of them to have each of those other departments to justify why are you here? What is your job? What makes your work essential to IBM?
And we eliminated. I
mean, we kept the technology stuff. No one said they needed the tech stuff, but
I said they did. So Sy
And it was amazing we were able to do that. And we did, basically, requiring that people justify why they're there and how they're supporting the products we're developing. That was why we were there. So it was really a hell of a good reorganization and it worked very well. Also, there were a bunch of battles with Bob Evans. Bob had a habit of going out to various labs and looking at what ad tech was doing, and if somebody had a wild idea, they'd bring it to him. And somebody had an idea for a disk printer, a little printer where you have little flexible plastic disks. And you'd hammer the disk onto the paper. And he showed us it would be real cheap, easy to make, et cetera. And so Bob essentially directed us to develop a disk printer. And we disagreed with it. The printer guys didn't want to do it. The printer guys were really difficult on that. The manager was a guy that I had a lot of battles with, who worked for me. And the basic problem was that he had expected to be the Lab Director, instead of me. And he'd been with the company a lot longer and that sort of thing. And so we did not start off well. And we never did work well.
As a matter of fact, he turned out to be a real problem later. So we had a big battle with the disk printer. They were never able to get it to work, and they tried all kinds of stuff. It was an enormous expense and big failure. A little bit later, as I said, I had this battle with the printer manager. And I ended up pulling him out, and I wanted to reassign him to run the OCR/MICR stuff, but he wouldn’t have any part of it, and we got into a big flap with that.
And so, I had a bunch
of problems with people, and a lot of them were old timers that had a lot of
background that we had some real trouble with. It was a tough environment to
work in because anybody could go open door if they disagreed and any of that sort of thing. You couldn’t move people without all kinds of
justification. It was a very difficult way to do things. But in any event, we
also had the OCR/MICR Group, optical recognition, which was in
Why RCA Failed
Humphrey: Well, Pat Beebe was my boss then. Actually, he was my boss’ boss. All of a sudden, I was his boss, and boy, that didn’t sit well with him at all.
Booch: I can imagine.
Humphrey: Well, that was the case with several people -- that basically here was a guy from lower down that came in. Of course, I’d been running programming. So, I had been in all these big jobs. So, I rocketed past all these people who felt that they had earned their keep and I hadn’t and so. I had all kinds of problems with them. So, Pat ended actually resigning from IBM. He quit and I had arguments with him when I worked for him because he [had ideas that] turned out not to make a whole lot of sense and basically said you couldn’t make a general purpose programming system, operating system. He said they all had to be special purpose. He was a special systems guy. So, he was just plain wrong on some of this stuff. Again, one of these people that had a conviction and essentially made up facts to support convictions instead of testing convictions against fact.
Booch: And, he had grown up primarily from a hardware perspective, isn’t that correct, where there was the legacy of building specialized machines.
Humphrey: Well, he actually was the guy who ran the SABRE System development. And so, he had that background -- the special systems stuff and how complicated they were and all that sort of thing. He came from that community; a hardware guy, but also through the SABRE stuff. So, he had a pretty good extensive background and he was no dummy. He knew how complex these big systems were. So, that was an issue. He had been in charge of the early development of that.
So in any event, Pat quit. Do you remember Orville Wright was my marketing manager when I ran TSS?
Humphrey: Orville Wright also quit and they both went to work for RCA. There’s an amazing RCA story here because Pat, when he quit, said, “I’m going to RCA and we’re going to clean your clock,” and was really aggressive on that. He had taken the job as the VP of development. Orville Wright was head of their computer operations. So, he took it over at RCA. He had a big job. What was interesting was RCA had come up with, as I said, the Spectra 70, which they had gotten going, but they were replacing the Spectra 70 with a new line. What they had done, which I learned later, they made some just god-awful mistakes in a marketing and policy basis because they established a series of long-term leases to sell their machines, and they were actually pricing them with fairly long lives.
So, they were able to compete pretty aggressively with IBM, and they were getting business. They were doing reasonably well. And so, when IBM came out with a 370, they had a problem because they were developing a new replacement system also with higher performance technology, and that’s why Pat was brought in -- Pat Beebe and Orville Wright -- to go in and fix that and bring this new system out. But the new system was delayed, and it was pretty seriously delayed. And so, what they decided to do -- Pat and Orville -- they decided that instead of coming up with a new machine, they would fake it. So, they would say, “Okay. Here’s a new machine.” They would take the old Spectra 70 machines, bring them back and repaint them and do a little bit of clean up to them and ship them back out as a new model at a much lower price. It was kind of fakery.
And so they did that, and what they didn’t realize was that all of their long-term leases had granted an out. You had to pay a termination charge to get out of the lease unless you were buying another GE machine, like getting another Spectra machine. So, it turned out that all of their customers could now take the machines they had, essentially trade them in, and get one of these newer models at a much lower price without paying a nickel. So, the new RCA operation, all of a sudden, they were running their factories like crazy, retouching all these machines and cutting the hell out of their revenue.
And so these guys went down to the RCA board meeting one day to decide what to do about this, and overnight, they wiped out the RCA computer division. The board just said, “No. It’s gone.” Oh, and it became a money pit. They were into the hole for like $100 million in no time. And so, that was sort of the dead end for those guys. Orville came out well. I don’t remember what Pat did. I heard of him running some smaller company somewhere.
But in any event, we also had another big flap. Learson now had taken over for Tom Watson. Tom Watson had retired. I was still at Endicott. When Learson took over -- this was about 1970-1971 -- they had to cut people. And so Learson basically came up with this thing that you have a requirement to cut out so many people. They levied a number on each lab. I got a number. I forgot what it was -- 70 or 80 people I had to get rid of -- but I had to follow standard IBM ground rules. I could only fire people for cause. It was ludicrous. I mean I had to go around and figure out some way to fire good people that were doing good work. I totally objected. I had big battles with my immediate management, and so I got in real hot water over that again.
My boss at that time
was still Bob Evans. He reported to a guy named Spike Beitzel
who was the group executive, and I really got into a terrible flap over that
thing because it was unfair. We were being dishonest with these people. If we
were going to fire them, let’s come up with the layoff number and they were
unwilling to do that. So, we had to somehow make it up. Learson
was driving that. While I thought he was an extraordinary and marvelous man, he
sure had this hard edge to him on this. This was a toughie. But, that was a
real problem. So, I got in a real battle over that, and then had another
re-organization -- IBM kept re-organizing -- and decided they were going to put
a whole bunch of things together, restructure it. One of the big debates we had
at the time -- and this will come up a little bit later -- we had a big meeting
with Bob Evans and all of his lab directors. He’d had these lab director
meetings periodically, and we had them in everybody’s labs. I remember we had
At this one meeting, Bob wanted to talk about how should we structure the division in terms of the way we lay out the logic for the organization. The option basically that he was proposing and pretty much everybody else wanted was we will organize around computer systems. We have a large systems group and so each of the systems would be the center and they would each have their own programming group and that sort of thing. So essentially, we would fragment the whole programming community. That was sort of what had been done before, but it hadn’t been broken up. And so they wanted to really now essentially break it up so everybody had their own programmers. All of the hardware system managers were now the system managers, and they would have their own hardware and software. The file people would get the file software. So that’s what they were going to do. I objected. I said that makes no sense at all. I said that what you really need to do is to have an operating system focus instead. So, focus on what the users are doing and do it that way. Well, the whole hardware community would have none of that. So, they basically decided to re-organize that way. The whole programming community was essentially splintered and that happened, as I said, right about the time of the re-organization when I went to Endicott in January 1970. They later asked me to go down to corporate staff and take a job as director of policy development. This was about in 1972.
Booch: If I may ask, around that time when you were still the director of programming, what was your span of control, because you told me these stories of all these groups you’re working with.
Humphrey: Oh, okay. I ran
programming from 1966 to 1970 and had a whole lot of people under me. I had
Fritz Trapnell, who was the director over the OS 360
-- later became Don Gavis and Fritz, by the way --
went over to
Booch: Got it.
Humphrey: So, Fritz reported to
me. Jim Frame had the intermediate systems, DOS and that sort of stuff. We had
Dick Bevier and all the labs reported to him. The lab managers were different
than the system managers, by the way. There was a
But, the big operating systems, the large and intermediate and the small, they reported directly to me, and Dick Bevier had the labs. I also had my own very small staff. So, I had about five people. I didn’t have a big crowd. Dick Bevier had quite a list of lab directors, although the European people were all kind of pulled together in one group under By Havens. But other than that, they were pretty much all under Dick.
Booch: Okay. That helps.
Humphrey: So now, from 1970 to 1972, while I was running the Endicott lab, I was no longer the programming director.
Humphrey: You remember now the
programmers were all reporting to their local laboratory management chain. When
I went to the Endicott lab, all the programmers were doing that as well. They
had like the
While I was in the
IBM was being aggressively attacked at the very low end. It wasn’t PCs yet, but small computers - DEC and others. And so the hardware guys had concluded they really needed to make a small system that was not 360 compatible. They were going to break the line and go off on their own -- the System 38 I think they ended up with. I was irate. I thought this was crazy. So, I wrote a blazing letter to Spike Beitzel about compatibility and what we would do, and I got Dick Case, who had worked for me as the 360 system architect, to bless it. He went over it and commented on it. I didn’t anticipate the Internet, but I basically said that compatibility, interchangeability -- that there were various levels of it and you really had to have the ability to move data and programs back and forth -- that people were going to do this stuff dynamically and we had to be able to interchange between all these systems, and that going off with a completely different system was a terrible mistake.
I never kept copies of any of it. I’m not a pack rat, so I don’t have much of that stuff. But in any event, it created no affect at all.