An Interview with Bill Masson

Bill Masson
BM = Bill Masson
GAM = George Michael




GAM: We are interviewing Bill Masson today. Bill, why don't you start by telling us something of your background and specifically, when you came to the Lab and what you were doing and so-forth?

BM: I joined the Laboratory, actually in Berkeley, in 1960. I came to work in August, 1960 in the procurement department in Berkeley and my initial responsibilities were in dealing with computer-related acquisitions. I was involved initially with the LARC, particularly the LARC spare parts acquisition, and then the resolution of the STRETCH negotiations between the Laboratory and AEC and IBM, where we were involved in the reduction in price from the initial price of the STRETCH down to a much-reduced level because of it's failure to perform to specifications.

GAM: I remember that.

BM: So, it was through these contacts that I met Sid Fernbach and he asked me to come to Livermore. I did in 1961 and have been involved in computational activities from that time till 1990, when I retired from the Laboratory.

GAM: Beautiful! So, you spent most of your time in the Administration part of the organization, we didn't even really have a Computation Department when you first showed up, Theoretical Physics and Computation were both under Sid.

BM: Yes, it was a Division at that time, and Sid was basically the Division Leader. He had two rolls; he had Theoretical Physics as well as Computation and it was a kind of a loosely assembled group of people. We had a number of what we called Supervisors in those days and each one had varying sizes of groups and, of course, applications people and operations people. And my initial roll was as kind of, well, an assistant to Sid, but then it was kind of financial management. I was responsible for the budget and for many of the contracting issues that were then becoming more and more the responsibility of the Division. In a very early period, much of the computational procurements were handled by Berkeley, and

GAM: Who was that there?

BM: Well, that was Don Bruce, the procurement manager and then Charlie Blue was made the procurement manager in Livermore. And he handled procurements for Livermore when the administration of the two facilities was split. In moving those procurement functions out to Livermore, I played a key role in that because of my past.

So the earliest assignments with Sid dealt with developing a financial structure. We had budgets and things like that, but, there was some degree of informality to that. We were part of the Weapons Program, the Weapons Program funded 99.9% of the computer facility. So little, if anything, was done, at least in the early days, to support the other programs and their support was just peripheral to the Weapons Program. As things continued, and after the early 60's, then, of course, there was increasing dependence on the other programs at the Laboratory. It never became overwhelming until much later on.

GAM: Well, can you contrast the informality of the early years with the increasing structuring of these procurement efforts as they went on.

BM: Oh, yes! In the very early days, we had a very easy and well understood partnership with the AEC, with the University, and with the Laboratory. It was easy to define needs and, once these were defined, then it was just a total collaboration to achieve what ever that objective was. So, if you wanted a machine, or needed a machine, and the Weapons Program could defend it's need, and it's availability, that is, that the manufacturer was at least developing something in the direction that would meet your needs, it was a very clear, (I wouldn't say simple) problem, because we had to convince a lot of people. We had to convince Congress, we had to convince various elements within the AEC. But, it was a very positive effort, there were few obstacles put into our pathway. The greatest obstacles that we faced in that time were really the manufacturers. It's not that they were unable to sell us things, it's just simply that the technology, their plans for providing things, sometimes did not meet the schedules that we had set for ourselves. Many of the machines were a little bit late, or quite late, and slower than what we had anticipated. They had not reached the performance expectations we needed; even Control Data later on in the mid 60's had troubles. The compatibility issues from machine to machine were not solved. So, there were delays and that kind of thing, going on. But, the climate to acquire equipment in those days was nothing like it is today. It's almost a hostile climate today, as opposed to the period of the early 60's and the mid 70's.

GAM: Well, I suppose that's a natural consequence of age, but I really feel that the people they have running around there today, are not imbued with the same kind of missionary zeal that we all had in the early days. It was, sort of, really exciting then.

BM: Yes, I think you're right. I think that's one of the major issues; the major explanation, is that in those days the Laboratory had a clear cut mission and those missions were understood by the people we had a deal with. Even the Congressional staffers that we had to deal with understood the mission of the Laboratory well. Because of the classification, there was a relatively small group of review processes at the Congressional level and, so, the people you dealt with were the same people you dealt with the previous year and the year before that. So, your history, your performance, your accomplishments, were understood and appreciated. That's not necessarily the case today where you have many, many, more committees. Many, many, more people that are looking at various aspects. Within the DOE structure and from the very beginning of the DOE organization as opposed to AEC, things changed and they became much more regimented, they became much more subject to, I wouldn't say critical review, I think they became subject to review by people who were considerably more disinterested than they were in the early days.

GAM: Yes, in that Washington connection, can you outline or describe how you guys interacted with these congressional committees?

BM: Well, in the early days, most of our interactions were with the controller of AEC, John Abadessa, and he had a great deal of influence at the congressional level. And so, it was almost like the same role that Duane Sewell played at the Laboratory. If you convinced Abadessa, just as when you convinced Duane Sewell, things would happen. I mean all you had to do was to inform Abadessa sufficiently as to what your needs were, and to what your plan was, and your expectations for success, and if he was convinced he sold that to Congress. And we didn't have to go before a congressional committee and testify, because John Abadessa would carry that message for us, because of his influence and because of the respect Congress held for him, those things were done in a rather straightforward fashion.

GAM: But you did deal with staffers and stuff like that?

BM: Yes, we did.

GAM: They come out to the Lab ever?

BM: Oh yes, they had a small, a very small, organization. I can't recall all the names, but John Kirskey is the one that I remember. He had a small group of people, (today they are represented by about 50 or 60 people.) He had maybe four or five people that handled all of the computer tasks for AEC. This is, of course, before the Brooks bill and the other limitations that were imposed on the acquisition of computers. In the scientific area, computers were considered to be strange beasts, but something that was acceptable. Once they became more prevalent in the administrative area, they took on a different rule. Congress became fearful that they would be abused. That they would increase in numbers astronomically, and people like Brooks, Jack Brooks from Texas, thought that they had to be controlled. And once those controls, as they evolved out of the Brooks Bill, were put into place, the freedom of acquisition and the speed of acquisition were slowed down immensely. It got to the point where many organizations, not AEC so much because of it's scientific base, but other organizations like the military, in particular. Their initiation-to-acquisition-to-use was running seven to eight years as compared to the few years, one to two years, that it took the Laboratory and the AEC to achieve the same things. It was really a very, very, difficult time for a number of organizations. In fact, it got to the point where many of those very organizations, the Army and other organizations would come to the Laboratory and seek ways, or try and examine ways, for us to do something for them, to bypass because our pathway was so broad and so uncluttered compared to theirs.

GAM: That's interesting, I didn't know that. That's an interesting possibility. We didn't do much of that though, did we?

BM: No, because we really were very protective of the priviledges that we enjoyed. When I say that, what I really mean is that we tried to keep our process as clean as possible. We could have easily enjoyed buying things for other people and taking a little bit off the top, but if we had we would have lost our uncluttered pathway.

GAM: Your purity!

BM: That's right, our virginity would have been lost.

GAM: Well, Los Alamos did, at least they have always had the DNA computer there. It always pissed me off that we couldn't have something like that at Livermore, I think that we would do a better job, but then...

BM: Well, part of it was we felt that we really had to meet our own needs and we tried not to corrupt it. I think that they were really being abused. That is, these other organizations because of interference by Congress and by people like Brooks. And so we just really did try and keep a more pure approach to it. We made enough mistakes as it was, we didn't need to compound it by any means.

GAM: I understand! What about the relationships, let's see, Sid had a staff that included you and Wayne was there, but I don't really know anybody else, well, they had a secretary, but

BM: Well, that really was the organization. Within a relatively short period of time, Wayne Hudson became Sid's assistant in Theoretical Physics and so he no longer involved himself in Computational activities at all. So within about a year after I joined, Wayne Hudson and I became counterparts.

GAM: I understand.

BM: He was in Theoretical Physics and I was in Computation, and so, it was for many years it was Sid who was Department Head or Division Leader, or whatever you wanted to call them in that day, and then I was his assistant. Of course, Sid had a number of technical supervisor's who had various roles throughout the organization.

GAM: It was a fairly flat organization then, right?

BM: Yes, it was. Yes, it was.

GAM: I think that's very good.

BM: Well, clearly, Sid was the head of the Department and there were many people who made major, major, contributions, but it was easy to deal with a structure like that because he was clearly the head of the Department.

GAM: Yea, well, you knew where the buck stopped.

BM: That's exactly right!

GAM: So, how about the relationships with the Director's office? What's going on there?

BM: Again, we were somewhat blessed. Because Computation was not an orphan. Sid was the equivalent of an Associate Director under today's structure.

GAM: yea.

BM: Because he ran the computational activity and answered to people like Duane Sewell and whoever was running the Weapons Program. And so, he had pretty well, throughout the Laboratory, full responsibility for the computational activities.

GAM: Yes, I once asked (then Laboratory director) Mike May, why they never made Sid an Associate Director because, clearly, he doing a fairly effective job, even if I didn't always agree with him.. He said, "Sid will never be a Director because he doesn't have a Program," and I wasn't smart enough then to realize that was maybe not the whole story. The reason they never made him a Director is because he would have overmastered the other ADs. He was so much more effective.

BM: For years and years, all computational activities were controlled by Computation, it wasn't until considerably later that even the most powerful organizations got their own computers.

GAM: Well, I think there were, (I don't mean to leap ahead to a conclusion,) two basic mistakes that Sid made. One was that he wouldn't pay enough attention to the small computers; he thought they were beneath his attention. And secondly, he got us two STARS (CDC STAR 100s) instead of just one and I think that those two situations together destroyed his empire.

BM: I think you may be right, I know that the pressure for dealing with local computational activities was not within Sid's realm of interest at all and, therefore, he gave it little or no attention. His total focus was on very large machines and large scale computations and he almost resented the diversion of funds to any other activities. That applied to not only the scientific side, but also to the administrative side and I think the Laboratory could have fared better in the total computational area if more attention have been given to those two aspects. The small machines and local computational needs, other than super computers, and then administrative computational needs which were put off for such a long time. Long after many other organizations had dealt with their administrative computational activities in a much more generous fashion than we did. But his enthusiasm for the big machines also stimulated the supercomputer market and if it hadn't been for that almost single-mindedness the process would have been slower. You can fault the STAR's, you can fault other activities but if they hadn't been undertaken, the process would have been slow.

GAM: I agree, I'm not faulting STAR's; the STAR was a very important machine; it just didn't perform well on our big design programs.

BM: Yes, I know that. I think that Sid felt very strongly that he didn't want to abandon Control Data and the STAR effort until it was a success. But, on the other hand, I think we all now understand and appreciate what the CRAY's effect could have been. We would have been better off if we had gotten the first CRAY 1 instead of Los Alamos. Because it would have matured much faster here. It was just one of those technical decisions that was based on Sid's intention that the STAR was going to work and he put all his energies into it, just like he put all his energies all the other projects that he tried to support.

GAM: Can you just generally describe an average procurement process, from the point that you discern you need a machine and you write a spec.

BM: In large part, much of the new acquisitions, that's the acquisitions of a new series machine, were stimulated by the awareness that something was coming into being and Sid's most frequent comment was that the shopping list of computational needs was so long that all you could do was chip off a little bit each time to expand the ability to make the computational calculation. So, maybe it was a larger mesh or a finer mesh or whatever it was you had to deal with it in small increments. So, if something came along that looked like it met those kinds of needs, then he would go and promote it. He would promote it first to the Weapons people and, because of their insatiable appetite, they were ready listeners. Super computers, or large machines, were the only thing they had lived with. We hadn't yet evolved to the point where they could get it on their desk so they looked to the supercomputer as the solution to their requirements, and they were an eager audience. Once that promotion started inside it was very easy to define; the next set of items on that long grocery list of things you would like to have but couldn't meet with the current capabilities. The process, then, was to write a justification and convince the people in Washington, the John Abedessas, and others later on, that you had that need. Once that was approved, then the formal negotiations that established the contract, the specifications, the delivery, and so forth, were started. In most cases, much of that was already being discussed. That is, we would pursue the process in kind of a parallel fashion. One was to get approval from Washington, the other was to start finalize the process with the manufacturer. In many of the early transactions, we weren't dealing with a competitive process, we were really dealing with whoever had the biggest, the best, the fastest available. In certain respects, it was easier because few people came out simultaneously with an equally capable machine. They kind of leap frogged one another so you could pick your candidate, and there was only one, and then you would just concentrate on them. So, you would take these parallel paths. One, the Washington path to get approval itself, the other to deal with the technical issues, how the machine would be structured, whether it would meet your needs, what it's price would be, when the delivery would be, those kinds of things. Once you got the approval, then all those things you had been doing kind of in the background could be brought into the foreground and they became the focus of the transaction.

GAM: Well, you set somebody to writing specs too, like Stu Stone, or I, for that matter, occasionally wrote specs.

BM: Yes, but you see the last real set of specifications we wrote were for LARC. We defined that machine and Remington Rand went out and built it. But STRETCH was an IBM machine, the CDC 1604's, the 3600's and the 6600's, were Control Data's machines. The 7600's were also Control Data's machines. The STAR was a Control Data machine and the CRAY's were Cray's machine. So they were defined, really, by somebody else, and molded by us. They were within a framework that would meet our needs and maybe we skewed them somewhat, but primarily we were dealing with somebody else's design to which we matched our problem. And so, the writing of specs really dealt with much of the peripheral stuff. Things like mass storage and graphic systems and those kinds of things where we tried to fill those special niches within our network that weren't really readily available in the marketplace. Or somebody would come out with a product and we would say "well, that just about meets our needs, but not quite", so we wrote a set of specifications that would meet our needs. But the large machines, I can't really remember anything that was done based upon "our specifications" that somebody else built it. It was really putting in specifications in a contract if it was a solicitation that would allow what we wanted to be responded to.

GAM: Well, you made the distinction between the solicitation and what?

BM: Well, of course, many of our acquisitions were really sole-sources. So that would be the alternative, a sole-source acquisition. The sole-source acquisitions were based on the fact that, for example, the 6600 was maybe 10 times faster than anything else in the marketplace. So, we did not try to write a set of specifications that would allow the 6600 to be bid and then allow somebody else to be bidding. We just simply said that we wanted the 6600 and we went out and got the 6600.

GAM: Yes, do you think that is possible today?

BM: I don't think so.

GAM: Even if we have machines that are so clearly superior to the run of the mill?

BM: No, I don't think so and I think it's to our disadvantage too.

GAM: Yes, I think that's just true.

BM: Because, I think what you do is slow down the whole process and then often you get what is sometimes second best.

GAM: Indeed! Well, in the normal day-to-day things, I'm sorry to be jumping around like this, but these thoughts just occur to me. Did you guy's use Sid and others around you to meet once a week, or every day, just to discuss strategy or whatever else might have needed to be discussed?

BM: We, of course, had regular staff meetings. But, many of the issues were just day-to-day operational issues. Particularly in the early days, when a startling new machine was evolving, there was kind of a single purposeness to it. There were few objections to these new toys, and so there wasn't a great deal of debate. Most of the efforts, or contests, occurred shortly before delivery. After the manufacturer was well along in the development of it and we started testing it and developing operating systems for it, that kind of thing, then the actual contests or the discussions went on between the technical people; as to direction, course of action, that kind of thing. Those were generally technical arguments or issues that I wasn't directly involved in that much. I was really more responsible for the operational activities.

GAM: Well, apropos of that, can you describe the budget process? How did you get to where you said that's the budget?

BM: Early on, again, it was a relatively simple process. We told the Weapons Program how much it was going to cost to run the machines that they wanted and they put up the money. We didn't have to solicit support from any other organization because they put up all the money. When I first came to the Division, my first budget was 7.1 million dollars. When I left, it was something like 61 million dollars, just to run one center. That's not the two or three centers we were responsible for, that was just the weapons center. As the user base became more complex and more diversified, we really got to the point where we were negotiating with the Weapons Program to make room for other people. That was a process that was very slow and very painful. The Weapons Program really didn't want to give up very much of the machine time. But they also wanted to get it for nothing and so they really couldn't have it both ways. The cost of the operation was reaching a point where they literally had to have other participants and there were people clamoring for that involvement. Of course the restrictions on the closed machine, that is, the classification issues that we had to deal with, precluded many people from using the machines. That, of course, opened the door for many of the small supporting systems that evolved throughout the entire Laboratory. There were many groups that couldn't get into the computer in a direct fashion, they had to do it in a very indirect fashion because of the lack of all the networking capabilities that evolved after 1964 when we got the 6600's and started developing the Octopus network and what not. From a budget point of view, we shared the resource always, of course, deferring to the Weapons Program up until very recently. The largest portion of the resource went to the Weapons Program in one fashion or another. Of course, today it's considerably less than it was back in the 60's and 70's and early 80's when it was really the only game in town and there was a difference.

GAM: Well, I remember everyone had to go the Weapons people to get time, even though they were supporting the Weapons Programs themselves. Engineers, Chemists and so forth.

BM: Well, the Weapons Program took advantage of the fact that it was their money and so they did not really allow us to sell time to their own Program. They allocated for themselves what was going to go to those Programs and that gave them control. If we could have treated Chemistry and Test Program and others as individual customers, we would have been able to deal with them in a different fashion. But, the Weapons Program took a different course and they basically said if they are going to support our Program they are going to have to come through us to get the time and we'll allocate the time to them. And in that fashion they did control the allocations of time.

GAM: It was a fairly relaxed thing though. I developed a whole bunch of diagnostic programs for their testing of their devices and I merely told them that I was going to do that and they said, "Swell, do it."

BM: Yes, and I think that is why that system worked for as long as it did. They were not generous with their time, but whenever it was of interest to them they would readily give up the time. They had no reservations.

GAM: It's interesting in all these interviews, there's a great flavor of nostalgia.

BM: Yes, the good ol days, it always comes through.

GAM: Well, that's good, really good. It's just amazing.

BM: I spent so much time doing basically the same kind of thing, I was a Physicist and then I was assistant to the next Associate Director and then the next Associate Director and the next Associate Director that had Computation and I saw this change in the attitude from the AEC to ERDA to DOE and, God, I'm glad I'm out of that place.

GAM: Yes, I think that is what lot's of people are suggesting. Well, let's get to some of these superlatives. Was there a particular procurement that stands out as the most interesting or the most difficult or most pleasing? Something like that?

BM: Well, I think the whole series of early acquisitions with Control Data, that is, the 6600's and 7600's were exciting times. We had so many of them and they were almost spilling over at times. Each one had it's own character. The first 6600 was a real adventure, but the ones after that kind of cascaded so quickly that we seemed to be doing only that. There was finally enough money to buy them or, later on, lease with purchase options and the other kinds of financial restructuring, or financial shenanigans that we went through to acquire them were very entertaining. But it was a period of very concentrated activity. I mean we seemed to be doing a procurement constantly. The Test Program was starting to be more and more controlled. They went to underground testing, of course, then the amount of computational needs were increasing dramatically and it was an exciting time. Later on it became very challenging to get the procurements through the process, they were much more complex. The documentation you had to put together was much more broad. You had to deal with many, many other issues than you did initially. Early on our justification was primarily based upon a classified description of a particular problem. It might even be an event, a test, a particular test that had some unique complexity. So you wrote a classified technical justification and then wrapped it in an envelope and, in affect, dealt with the administrative and contractual issues. You sent it off to a limited number of people, because only so many people had access to that kind of information and you were finished. Then, later on, these things were reviewed by so many facets of DOE and then by so many committees within Congress that they were not classified documents, they were general documents. They had added so many facets to the justification process that it read like a book as opposed to a nice, neat, little tight argument that focused on one issue. You kind of had to solve the problems of the world in one document as opposed to just one simple event.

GAM: In the middle 60's, we got the first 7090 or 7094, whatever it is, and another one that was coming up at Christmas time from Los Angeles. The truck went over a cliff or something like that? Can you tell us about that?

BM: I just remember, very vaguely, the failure on a transportation problem. The machine was damaged, but it was replaced shortly.

GAM: Yes, that incurred for us the undying hatred of the University of Minnesota.

BM: Yes, because they were scheduled to get the next available machine. Back in those days, there was a priority system that allocated scarce resources. One of the scarce resources in the United States were major computer systems. We invoked the priority system whenever there was a danger that a computer delivery would be too late to support the Lab's test program. We had the privilege of imposing a DX (Defense priority) rating on a given computer procurement. That meant that we could move to the head of the line and that's what was done in this particular case, and the next person normally scheduled to get the machine, lost it. The unfortunate thing is that after we had done it for a few times, other people decided that they had comparable needs. Because everybody wanted to impose them we had to make special pleas to AEC. AEC carried a great deal of respect within the Congressional structure in those days. It was a very small organization, not the bureaucracy it is today, and they were considered to be a highly technical, scientifically oriented organization. Their level of respect was really held in high regard. And so our DX rating was more DX than others.

GAM: Well, the real adventures I sort of enjoyed was the procurement of the IBM 1360 - the photo digital store. That was a pretty interesting thing, it was something Norman and I had schemed about for some time and it was a very elegant piece of machinery. You were involved in that.

BM: Look at how long it lasted! More than 10 years! I mean can you imagine anyother mass storage system that was conceived, built, and installed, and then worked for more than 10 years, and at its departure was still at the near forefront of it's technology?

(Note added: The Photostore was delivered in 1967, went into full operation in 1969 and was still fully operable when it was shut down in 1981 because spare parts were no longer available.)

GAM: Well, it's still, in many respects, ahead of its time. Very few of the systems that are available today have all of the features that the photo store had. But, they are certainly getting bigger now, much bigger. I always liked to tell the people who visited the Lab that when we put that thing on-line it doubled the memory capacity of the planet earth. That was quite an achievement in 1967, but now, of course it's much different.

BM: Of course, not only that, but the reluctance of the people at the Laboratory to get off of it. I mean, can you imagine, after 20 years of test results the people said "we need em, we need em." It's capacity was just incredible, particularly when you consider its off-line capacity.

GAM: Yes, you could put stuff on the shelf. Well, a great deal of that is credited to IBM, but there is also great credit to Garrett Boer, in particular, and John Fletcher and Clarence Badger, in general. They did a really good job putting that thing on-line and making it trustworthy, so to say.

BM: It was so trustworthy that people didn't want to get rid of it.

GAM: Well, it met the need. It clearly can't today, but it certainly lasted longer than anyone had a right to expect.

BM: Well, I think that those are the kinds of things, like the mass store that we did because of a real need, and they were very important contributions to the emerging computer technology.

GAM: Well, the TMDS was another thing that was a great success that is still used 30 years later.

BM: Yes, those, I think, are really the major technical contributions that Computation made to the industry as a whole. Those things never would have happened if the Laboratory hadn't driven their development.

GAM: That's true, that's true.

BM: I think that the supercomputers would have been driven. What I mean by that is without the operating system effort that we put in the 6600's, they would still have been populated across the United States as quickly they had been. Because the need for fast computation was felt all over the country and others quickly built parts of simple operating systems, whereas our time sharing system was an envelope, or umbrella, that made it very easy for CDC to just increase the 6600's popularity very quickly. On the other hand, from a technical point of view, I think the things like the mass store would never been accomplished if it hadn't been for the driving force of the Laboratory. There were other areas like that where we really made major contributions too. We made major contributions in the supercomputer area too, obviously. I don't think many organizations would have taken the chance that we took with brand new products. That was something that gave stimulus to the Control Data's and the CRAY's, and others, to do things because we were willing to take some risks.

GAM: One of the things I think is a little sad, though maybe I'm over exaggerating, is the lack of recognition that Seymour Cray has generally. He's really a national treasure. In our case, since 1962, he's been the principal supplier of computers to support the Weapons Program, not only at the Laboratory, but all over the Country.

BM: Well, actually, he's been the major contributor to supercomputing period. I mean one single individual.

GAM: He's an amazing guy.

BM: I think there is that appreciation from the older hands at the Laboratory, but I'm not sure that today many of the new computer scientists even know who Seymour Cray is.

GAM: Well, you may be right. We have strayed from the path in many respects. I'm not sure if there's other things that we might talk about or anything like that.

BM: I'm not sure there is really anything that we haven't talked about. There are many new events that have occurred, but I'm not sure they have any meaning to anyone other than you or me. Part of the history of Computations is it's recent history and that's something I don't think we want to ignore. There were a number of people who came later on who have contributed to the evolution of Computation, Gus Durough is one. Gus, when he came in, had to be exposed to computational activity. But, once exposed. he really made an effort to carry on many of the ideas that Sid had fostered very strongly. For example, Gus was very supportive of the creation of NERSC. (originally called MFECC), that could have fallen by the wayside. There was a lot of resistance from the Laboratory, and outside of the Laboratory, to bring that facility into Livermore. And, Sid played an enormous role in achieving that, in fact he was the primary mover of that. But, with the new Associate Director, that momentum could have been lost.

GAM: It's been suggested that putting it at Livermore and taking the stuff from inside and converting it so it would run at MFECC saved them 5 years.

BM: Yes, I think that is very possible. So, if the momentum hadn't been maintained then, I think, that would have had an effect on that large customer base that is now represented by MFECC.

GAM: Apropos of that, I am reminded of another thing that Sid did which is rather ironically interesting. He got those primitives at Los Alamos to start using time sharing. He got our time sharing system down there and they took to it like a duck to water.

BM: After great reluctance, and frequent failures on the part of their own activities.

GAM: Yes, I understand, but now they are avid. In fact, it must be admitted, they did a much better job of documentation than we did. Which is fine.

BM: They have actually done that in a number of instances. In their own history, they maintained a well-defined history of their computational activities where we haven't done that. We maintained millions and millions of lines of records, but they are basically meaningless from a historical point of view unless you go through and analyze them, and try and build all of the framework for which those documents had some meaning. I know that is what your trying to do now.

GAM: Well, really, I've found out that I'm an amateur at the process. I'm just trying to save some verbal history from you guys. Some of the guys are already dead. Bob Pexton is dead, Chet Kenrich from the UNIVAC days is dead.

BM: One could take each one of those stages in the evolution of the computational activities and base it on machine acquisitions, or whatever you wanted to do, or the restructure of the department and it's evolution. You could do a great deal of work on each one of those aspects. All of those records exist. There are stacks of those pieces of paper.

GAM: Carolyn Gousey was trying to do that.

BM: Yes, but that process died. I think it's unfortunate that the IBM antitrust suit occurred when it did. Because, there was a whole series of actions that were literally boxed up and set aside by the Justice Department edicts and the legal actions taken. Because we had to preserve those records, those records were treated as kind of sacred and we did not do anything, even though at that time Los Alamos was trying to build a relatively large group that was actually writing the history of their computational activities. We didn't and we should have, because the task would have been easier for you now if we had done that.

GAM: I suppose, you know you could just make the excuse that it's much more fun to do than it is to recall, you know. Apropos again, of a remark you made earlier, Sid didn't pay too much attention to the administrative aspects of computation or getting administrative computational support. Did you guys have any computer assistance for the things you were doing inside the Department?

BM: No, not really.

GAM: I remember I would see people running around with secret plots of salary curves and stuff like that.

BM: Yes, we did many, many, things by hand. We really did, compared to today. Today, there are so many readily available tools, but the financial management was done basically with pencil and eraser. Never in ink.

GAM: That's amazing, Are you going to say the operation was simple enough so you could do that, right?

BM: Well, yes, right, in part it was. We actually did the salary calculations by hand. Today, of course, you wouldn't even consider that, but we really did all financial management by hand. For salaries we would make the calculations on percentage of increase and, what not, by hand and on an individual basis. So we didn't make any effort to computerize many of our own operations. Our focus was on scientific computing. Now that's what Sid's focus was, he was not intolerant of it, I'm sure he would have given me as much latitude as I had asked for, but I guess I became indoctrinated too. It was sacred ground that you didn't tread on. What I was referring to earlier, I was really talking about the administrative activities at the Laboratory. Computation had an opportunity to assume responsibility for administrative computing at the Laboratory at one time and Sid said, "No." In effect he said, "I don't think we ought to do that; it's a meaningless activity and it would be a diversion of funds." I think that slowed that process across the Laboratory and it was to the Laboratory's disadvantage and to Computation's disadvantage. I think we would have been better off just making that an extension of the computational activities and letting people develop the expertise in administrative computational skills as well as formulated scientific skills. It had a major impact later on, when there was such a broad diversion of resources to personal computers and whatnot. Those could have all be under the control of Computation at one time, as opposed to being a part of Engineering, and a part of the Administration.

GAM: Bulkanized, as we say.

BM: Yes, that's right. If Sid had faults, this was one of them. He was very single-mindedly focused on scientific computing. That's a fault only from the broadest point-of-view. It's not a fault from the scientific point-of-view, because that's who really benefited from it.

GAM: Well, you know, when we had our memorial symposium for him, one of the things that came through pretty clearly was that he was extremely highly regarded outside the Laboratory. Much more than inside, to be honest. People from all over the world still recognize him as a pretty interesting pioneer.

BM: Yes, he certainly was that. He was an incredibly interesting man to work with.

GAM: I want to interview Sid's wife, Leona Fernbach. She worked on my original computation back in 1953 and 54. It's an amazing thing, it was a tight little group.

BM: Actually, it was a pretty small organization for a long, long, time and then it grew to be a very large organization. Then it was split up and now I hardly even recognize it.

GAM: It seems that no one wants to look at the lessons of history, at least with the idea wanting to learn from them. That biggness is not good, etc. and certainly administration ought to be focused a lot more than it is. Well, does anything else occur to you?

BM: No, not that I can think of.

GAM: Well, I'd like to thank you for coming by and doing this interview.




For information about this page, contact us at: comment@computer-history.info