TAPE 1, SIDE 1
MR. MARTIN COLLINS: Last time we left off our discussion with the various positions that you held in the Space Technology Laboratory. To get started this time, one thing that might have been unclear about it is the organizational relationship between STL and TRW.
DR. GEORGE MUELLER: Well, let's see. When I first went to work for Si and Dean, it was Ramo-Wooldridge. Then in about 1959, I guess, Thompson acquired all of the stock of Ramo-Wooldridge. They'd been the principal financial backer for Si and Dean. And in an exchange of stock, it became TRW. But at that same time, because of the involvement with the Air Force and the ballistic missile program, they spun off what was then Space Technology Laboratories and set it up as a separate corporation with the idea that it would continue to operate as essentially an FCRC, Federal Contract Center, but for profit. And there was a very large problem at that point in time with the feeling that that was unfair, that STL would have an unfair advantage in competitions, and that the Air Force finally decided that they needed a non-profit organization to do that, and at that time they formed the Aerospace Corporation. STL was originally set up to perform the function that Aerospace eventually took over. And so STL was a full operating company. It had Jimmy Doolittle as the chairman, and I believe Dunn was the president. Rube Mettler was the executive vice-president. Actually, when it was first set up, Rube was the vice-president and an old friend of mine was vice-president and I was a vice-president. Then Rube was made executive vice-president, and I took over as the vice-president in charge of, I guess they called it, research and development. But with the advent of Aerospace, we were forced to go off and get some new business. So I also became the first marketeer for STL.
COLLINS: That's how you got into that world. Why did the formation of Aerospace cause you potential economic problems? In other words, was some of the work that the Air Force previously assigned to STL given to Aerospace?
MUELLER: Yes. In fact, all of it was originally intended to go over to Aerospace, and so we were going to be left with nothing. The Air Force decided that they would continue using STL for their support of their basic ballistic missile program, and they moved the space activities, that were a fairly substantial part of STL, over to Aerospace. That was the first split. And new programs were all supposed to be managed out of Aerospace, and that is essentially the way it went over time.
COLLINS: What did that leave STL with then?
MUELLER: Well, the base program which they're still doing is the systems engineering technical direction of the ballistic missile program. That's the Atlas, Titan 4, Minuteman complex. And Aerospace has a supervisory role in that, but not really in any depth operation. And it left STL with the opportunity to go out and get more business, which it turned out we did.
COLLINS: Where did you think that you could find new business?
MUELLER: Well, we became a major purveyor of space activities to NASA, as well, and primarily to NASA. We also competed fairly effectively for space activities in the Air Force.
COLLINS: Was this systems engineering?
MUELLER: Actually, developing hardware. My first set of tasks at Ramo-Wooldridge was to build the first of the Pioneer spacecraft, and we parlayed that into a whole series of programs with NASA. We also took on the task of supporting the Air Force in essentially competing, so we built some of the first communication satellites for the Air Force, the synchronous satellites. We were a major competitor of Hughes at that time. And so that was one general thrust that we undertook and, I think, succeeded in creating clearly a viable organization that's grown over time.
COLLINS: Now, when you first joined Ramo Wooldridge, it was a for profit corporation, is that right?
COLLINS: And STL when it was formed was intended to be a for profit corporation.
MUELLER: Yes, and actually made a profit, too.
COLLINS: Why did the Air Force feel that a non-profit could better undertake the systems engineering that STL had intended to do?
MUELLER: That was primarily Congressional pressure that said, "Look, you can't make a profit on advising the Air Force." So people like Proxmire and company said, "Look, this is immoral," and so we had the choice of either going not-for-profit or forming Aerospace, and Louis Dunn and Jimmy Doolittle were quite opposed to the idea of going not-for-profit, as were most of us in the organization. We finally volunteered some people who went over to form Aerospace. Ivan Getting was brought in. But a fair number of people from STL were transferred over to form the nucleus of Aerospace Corporation. Al Donovan was the key person.
COLLINS: Do you recall what the major elements of the argument were for remaining a profit corporation?
MUELLER: We had a number of examples of not-for-profit corporations that were supporting the Air Force and Army and so on, and they were uniformly ineffective, over a period of time. They always started out with a great deal of vigor and energy, and within a period of ten years, they gradually became bureaucracies whose only real purpose in life was their existence, in my view, our view at the time. And to some large extent, that's true. I would say today it is true. Somehow or other, there's a lack of drive and motivation amongst the FCRC's. There's no measure of success, if you will, other than existence, and that's hardly enough for most of the very competent people.
COLLINS: I'm interested. It seems like obviously Jimmy Doolittle would have a very insightful appraisal of this kind of issue. Do you remember either hearing him discuss it or discussing it with him?
MUELLER: I don't--well, I remember some of the discussions. But it was just his feeling that that was not the way to go. You know, to some extent that gets to be an emotional issue, when all is said and done. And reason sometimes fails.
COLLINS: Well, when you first came to Ramo-Wooldridge on a full time basis in 1957, did you stop to consider what the role of a corporation like that was? This idea of a specialized corporation to handle the systems engineering and technical direction was a novel enterprise.
MUELLER: You mean standing back and saying, well, now, is this a good a thing for the nation? Or is this the right way to do things? Well, of course, remember, I was coming from a university background, and almost anything is better organized than a university, and so I felt that this was indeed one of the most constructive things that I had seen since it brought, it enabled one to bring together the talents of a large number of corporations and yet provide some overall direction to that effort. And clearly there is no way of bringing co-equal organizations together without having someone in charge of it. And if you have a prime contractor, it turns out there's always an inherent conflict between the prime and the subs and an inability on the part of the government contracting agency to find out what's really going on. Now, if everything works well, that's not a problem. But most of these very large programs that are complex, nothing ever works completely well, and so there needs to be built in a series of checks and balances, an ability to penetrate in a technical sense down to the work that's being done. And at that point in time, Ramo-Wooldridge was by far the most effective group I know of in the country for really understanding in depth what was going on within the technical progress of the programs. And that, in turn, probably is why it succeeded as well as it did in creating the ballistic missile program.
COLLINS: That raises very interesting issues. I don't have a very good sense of how it worked in detail, the relationships between Ramo-Wooldridge and the contractors and sub-contractors and the federal government. It might be interesting to kind of walk through the role of a place like Ramo-Wooldridge from the concept of a weapons system like one of the ballistic missiles, the Minuteman, and then the succeeding responsibilities as the program progresses. What was, for example, Ramo-Wooldridge's role in helping to initially define and conceptualize a weapons system like Minuteman?
MUELLER: Well, Minuteman, of course, was the third or fourth weapons system that they were involved in creating, and by that time there was a fair body of knowledge about how one goes about doing that. The Atlas was the first one, and that was a learning experience all around. But the Minuteman was actually the product of Paul Garabedian's work, and there's another guy, whose name I've forgotten, who was really the progenitor of it, an Air Force colonel. Now, what started as an idea which said, "Gee, we need a steerable configuration." You know, the original Titan was a liquid rocket, but its storage capabilities were limited, and so solid rockets had been under investigation for some time. But the ISP was low enough so that you needed to find some way of having a more efficient propellant, and there'd been work and technology developments that were providing that. So there was a basis, a technical basis, for a new concept to be inaugurated. And usually, this was carried out by bringing together some key technical people from industry, from the Air Force, its technical laboratories, and beginning to see what parameters one could establish for a weapons system, a new weapons system. At that time, Ramo-Wooldridge took the lead in bringing them together, finding the experts around the country both at the universities and in industry, and really forming a task group to develop this basic idea of a solid rocket, which was originally invented by a team at Ramo-Wooldridge who were looking at alternatives.
COLLINS: Did they then go to the Air Force and say, "We think we can do this weapons system. This is something that we know we have a need for?"
MUELLER: It didn't work that way because the Air Force people were in the same offices with us, so we were working side by side. And some of the Air Force people were quite capable technically, and so it was a joint effort of formulating this, if you will, this vision of new weapons system. And then it was a problem of selling up the chain of command, that this was the right way to go, and this was the thing that needed to be done. Of course, the circumstances were great at that time because General Schriever had a direct line to the Secretary of Defense. And all of the bureaucracies within the DOD were circumvented, and the pressure was very great because it was clear the Russians were busily building an arsenal of these things, and we were trying--and of course this was, at that time, a quite secret operation, so it wasn't known in general that there was any such work going on. And that included within the Air Force. This was one of the segregated projects.
COLLINS: In the case of selling this idea up the chain of command, would this be the responsibility of the Air Force officers who were sitting at TRW or would the TRW people participate in this?
MUELLER: Well, actually it was almost always the TRW or the Ramo-Wooldridge folks who did the selling. We prepared the briefings. We drew up the arguments and so on. Mostly the Air Force was in a supporting role because most of the technical, now I'm talking about the technical part of it, the funding and that sort of thing, contracting, was handled by the Air Force completely.
COLLINS: Once the idea was sold, I'm looking at the case of Minuteman, how did things proceed then?
MUELLER: Well, let's see. To start, we first had to convince at that time the head of the ballistic missile division, which was Rube Mettler, that this was a good idea. And then you had to convince Will Duke, who was the propulsion side of the house, that it was a good idea. And then you had to convince at that time Jim Fletcher and Frank Leehan that the guidance and control was possible. So we had to get those three people all together. Then the next step was to convince Si Ramo and Ben Schriever that this was the way to solve a whole host of problems that had been surfacing as we deployed the Thor and the Titan vehicles. And these involved logistic support, command and control, and that whole spectrum of things that needed to be addressed, that were not working very well. When you talk about putting a Titan in a hole and leaving it there for years on end with all those liquids sitting there eating away at things, why, it was more of a challenge than using solid rockets was.
So then we had to solve the--as another courtesy then you went by the Air Force head of the Air Force, Secretary of the Air Force, and then you went on up to the Secretary of Defense and told them this is what we needed to do and got approval. Once we had the approval, then it was a simple matter to move out and enter into a competition, brief sharp competition, between various people that looked as though they were capable of contributing this as an industry.
The selecting process was done very carefully, technical evaluation of the whole thing, but much more expedited than we do today. That whole time took about six months from the concept to the letting of contracts, actual letting of contracts, and starting to build the vehicles. Today it would take six years.
COLLINS: What was Ramo-Wooldridge's role in the selection process?
MUELLER: They did the evaluation, the technical evaluation of the various proposals. The final selection was made by the Air Force.
COLLINS: So Ramo-Wooldridge did some rating of the value of the technical components of the --
COLLINS: And then the Air Force weighed the other considerations that went into selection.
MUELLER: Yes. Si Ramo and Dean Wooldridge in particular, however, played a fair role in working with Schriever and his colonels in making that decision. And, as far as I know, their decisions were a lot crisper and had a lot less controversy than the ones we have today.
Once the program was under contract, then we would have people--there's a program office within Ramo-Wooldridge--that would be responsible for schedules, program plans, the technical documents that described the program. That was one of the real contributions that Ramo-Wooldridge made, which was to provide a framework within which you carried out these programs, where you had a program control function that really hadn't existed in any real sense before, where you had a systems engineering function, where you did the trade-offs between, on a continuing basis, as the program was going because you had to make changes, and sometimes dramatic changes.
In most of these programs, we had parallel paths. So we had two companies working on engines and we had two companies working on guidance and we had two companies working on the ground support equipment and so on because we were trying to move very rapidly and we knew exactly how it should have been done at the end. And so parallel approaches allowed you to make decisions later than you otherwise could have made and, therefore, gave you a better chance of success. And, of course, the internal competition between the two approaches also led to much greater progress and much less cost in the long run. I firmly believe that having two parallel programs reduces the cost of the program rather than increases it.
COLLINS: When you say internal competition do you mean internal competition within Ramo-Wooldridge?
MUELLER: No, no, internal competition between the two industrial partners.
COLLINS: Would the same person in Ramo-Wooldridge oversee the parallel developments in a given area?
MUELLER: Yes. Right. Although there would be other people that were working directly with each contractor.
MUELLER: And it was fairly open communications because everybody knew what everybody else was doing. We tried to keep that program as a whole visible to all the competitors.
COLLINS: Yes. Now, the funding for each of the contractors went directly from the Air Force to the contractor.
COLLINS: What leverage did Ramo-Wooldridge have to suggest changes or re-direct an effort once they thought problems were occurring or changes needed to be made?
MUELLER: The Air Force contracting officer was in a supportive role, so when you gave technical direction, he went along and he had to sign off that yes, this was within cost and budget or this was the budget impact. But it was primarily Ramo-Wooldridge that did the direction, with support from the Air Force. Now, if it was going to double the cost, why, obviously that went shooting right back up to Schriever and Ramo.
COLLINS: That's on the Air Force side, but how did you, how were you able to, say, get a contractor to change course if they thought perhaps the course they were already on was the appropriate one?
MUELLER: Oh, well--
COLLINS: I'm sure there must have been different opinions on it.
MUELLER: Well, technical directives. First of all, you'd have to consider it. Then you argued it out and finally you reached a conclusion and more often than not everybody would agree. In the event that there was a direct conflict, why, we just simply directed them to do it and, of course, the Air Force gave them contractual direction. So they did it.
COLLINS: So essentially you had the Air Force as your leverage.
COLLINS: But the thought behind that was that Ramo-Wooldridge, in a sense, had responsibility for the quality, reliability, of the final system, and therefore they had to have this kind of control over the direction of the program.
MUELLER: Yes, right. Of course, that was a bone of contention, at least initially, because there was absolutely no feeling on the part of industry that they needed any supervision from a technically competent organization. They were used to dealing with the Air Force, who really in large measure worked on the contractual side of it and not on the technical side of it. That's not completely true, but if you look at various programs, you'll find that the ballistic missile program did accomplish more in less time than almost any other program in the history of our defense. And that really was true of the Navy side as well.
COLLINS: Is that flexibility and ability to respond quickly partly a function of having a corporation like Ramo-Wooldridge in the kind of structure of the aerospace industry? Or would that have happened if the old system, simply of the Air Force working with prime contractors, had continued?
MUELLER: Well, you've got some good examples. You had a whole set of programs that were developed under the old method of the Air Force, the Navajo, and there were three or four missile systems that did not work, never would have worked. But part of the technology was used in the ballistic missile program, so the technological work was transferred. But I think beyond a doubt that this was the most effective way of doing the job, and it would not have worked very well without having tilted the mechanism rather dramatically. Now, you ask, how would I do a new program of some major magnitude? I would want to have a completely different chain of command than the existing one within the DOD and really again put it in a segregated area which is kept secret from everybody and really move out and do it. You know, the old Kelly Johnson Skunkworks approach is one that works remarkably well if you have to do something in a short period of time. The present set of review processes, where you spend years, literally years, from the concept to the time when you can get a person on contract, by that time it's already obsolete, and you shouldn't start it anyhow.
COLLINS: Just to proceed on that line a little bit, the heart of this whole Phase A, Phase B, Phase C/D kind of method for doing technical work rests on this idea that you need a very careful review and understanding of what you're attempting to do. How does that balance out with the approach that you're suggesting might work better?
MUELLER: Well, you know, it's interesting because, to some extent, Ramo-Wooldridge and, for that matter, I introduced that concept at NASA when I went back there. You would do it in that fashion, but you just wouldn't spend several years working on Phase A and four or five years working on Phase B and four or five years on Phase C. The concept of phased approach is a good one and particularly if you have a high degree of parallelism between competing teams in Phase A and do a selection in Phase B and then do a further selection for Phase C. The most effective one is where you also have competition in Phase C. That's prior to the time you actually commit to production.
COLLINS: I guess the question might be cast as, how much review is enough? Is there some sense in which you can kind of draw a line?
MUELLER: Sure. You ought not to spend more than three months on Phase A.
COLLINS: Okay, you've got a chronological line.
COLLINS: But I mean, in a technical sense, how do you know how much review is enough?
MUELLER: There is no limit to how much review one can do, as we've demonstrated. Now, what is enough? You need something more than zero, but you need not more than five people reviewing it, five of the right people.
COLLINS: All right, so what you're saying is, it's assembling the key group of individuals who have a deep understanding of the issues to make the judgment in a short period of time.
MUELLER: Exactly. And it isn't going to get any better if you wait and you spend eight months looking at it, really, or ten. Now, all of these things involve judgment, so you have to make the right guess. You know, the intuitive result is important in any of these major systems decisions. There's a substantial amount of logic, but when all is said and done, when you're choosing between modes of transportation going to the moon, for example, do you do earth orbit rendezvous, lunar orbit rendezvous, you know you can do any one of them, and now the trade-offs become subjective.
COLLINS: How do you balance this notion of a few knowledgeable people making the review with this great urge on the part of Congress for accountability? How do you kind of put this notion of what's best for research and development in relation to the sense of public accountability?
MUELLER: Well, you know, one of the forces in accountability is that you get more and more people involved so no one's accountable, so from the standpoint of accountability, Congress is working on the wrong end of the problem. They ought to be insisting that one guy be responsible and accountable, instead of requiring these hoards of people. It's like the medical profession today where you can't get an opinion from a doctor. You've got to get three opinions, so that when the lawyer comes in, he's got several people to sue instead of one. And most of that accountability stuff comes from the legal background of most of the people in Congress. If we had engineers in Congress instead of lawyers, you would have a quite different approach to the whole business of government. One of the fascinating things is, if you look at the government of Russia, you'll find that almost all the key people are engineers. And they have very few lawyers, and they're always in second roles.
COLLINS: Did Ramo-Wooldridge and later STL, when they still had some of this activity, have individuals actually sitting in the contractor's plants working side by side with the people, or was it more on a visit basis?
MUELLER: Oh, we always, both the Air Force and Ramo-Wooldridge, had local representation, but most of the technical direction was given by folks in headquarters. The local people were following up and making sure that things happened and reporting back what was going on, but the tendency was for major decisions to be made, or decisions to be made back in headquarters, because basically what we were doing was balancing the needs of the several parts of the program. We had three or four contractors building pieces that went together to build the ballistic missile, and you had to provide the interfaces. In fact, that was the principal way of controlling the program. There were two principal ways. One was that you have a carefully drawn set of interface documents which you couldn't change without full scale review because otherwise the pieces wouldn't fit together when you got finished with it.
COLLINS: You're talking about the systems themselves.
MUELLER: Right. And the other way was that we maintained control of the guidance equations. And the guidance equations are a crucial part of a ballistic missile because they take into account all facets of how the thing is working, built, and so on, and that forced us to really know what all of the pieces of the system were. So we did the guidance and control, which is software programs, but at that time, it was esoteric. In fact, it's amazing what we were able to do with such a very small amount of computational capacity.
COLLINS: That, I think, is what you were directly involved with. Is that?
MUELLER: Well, not really. At that time I was more in the total system end, and I didn't get into the bits and bytes except when they got into trouble, and then I ended up knowing more than I ever wanted to about bits and bytes and organizing to find out where they went and what they were doing. Originally I started out on the radio guidance system and did some of the early analysis for that, and then I got involved in guidance and control. In an error finding mode, what are the errors that contribute to the final CEP? And did some fair studies in that arena. Then I spent a lot of time trying to figure out how to build a system that would do certain things like get to the moon and come back, for example, or get to the moon, period.
COLLINS: In the case of the various ballistic missiles, systems engineering or the idea of systems engineering had been there a long time. What things needed to be added to this kind of concept? That, I guess especially, was fairly well worked out in the electrical engineering area, but what needed to be added to this notion of systems engineering?
MUELLER: Well, first of all one had to get the concept that a system included the total system because although you say, well, systems engineering was well known, it was really applied to what we would think of as a sub-system. It's a motor that got systems engineered, but not really the power plant particularly. If you think back as to how ships were built, for example, one would think that there would be a systems engineering approach to building a ship. Quite the contrary. First you build the hull and then you decide where the bulkheads are going to be and then you buy from various contractors the pieces that you put into the ship. And there's a general view of what all is going to be there, but there's no real interface control and no real mechanism for tracking and making sure that those interfaces are met. No one has really thought through all of those interfaces in the case of a ship and, as far as I know, that's true today. The first time these things are ever brought together, you find that, well, we have to cut a hole here because a cable has to go through there and it wasn't designed that way. And the roll moment of the ship is such that it interacts with the control system on the antenna and it resonates and it hunts wildly back and forth so we have to redesign the filters and get them so that they are stable. All of these things are done after they're all assembled. What we were able to do at Ramo-Wooldridge and later at NASA is to have a clearly enough defined set of interfaces so that we had a minimum of these idiosyncracies when they arrive at the Cape to be put together.
TAPE 1, SIDE 2
COLLINS: You gave the example of how systems approach was not applied in building of ships. Why did they come to prevail in the development of ballistic missiles then? It wasn't a generally accepted idea about constructing a large device?
MUELLER: Well, probably the genius of Si and Dean and the assembly of people who had been through this kind of a problem. Louis Dunn, for example, was running JPL, who had built some of the first of the Army short range missiles, and so they had a fair experience in what worked and what didn't work. And it was clear that if we were going to succeed in getting this thing through in a reasonably short period of time, we had to start all of the components at the same time, and a little thought showed that unless they all were designed to fit together and work constructively, it wouldn't work. So it was sort of a necessity to do it this way. And the Navy later picked up on the same necessity in their structure for the Polaris program.
COLLINS: You mentioned the interface documents. Were these in essence prepared, at least the first origin, prepared in advance of any development activity? How did that work?
MUELLER: Well, what you do is, you define the interfaces before you begin, before you really go out for contracts. And so during the first phase of this, one of the important things is to define where you're going to cut this thing up and what needs to go across that interface, what is the connection. And that's a crucial part of it because you want to simplify that to the maximum extent possible. But at the same time you want to have reasonable packages to build so that you can get a large enough package for someone to build it effectively and then, as you go to the next stage, which is the actual design, you go from the concept to the design. Then one of the key things is the definition of what it is you're going to supply, what are its characteristics, what are the mechanical and electrical interfaces, and just exactly what is going to go through and hold them together, and what are the strengths and all that, all of the details that make up a successful mating of two systems.
And those interfaces then become control documents, which require the concurrence of all of the contractors--because we're now on a contract basis--before you make a change in them because interfaces that change affect everybody in that whole system. And so you've got to have a review process that makes sure you understand what the impact is of even a minor change somewhere in that interface. Now, the beauty of this is that within the package itself you have a fair freedom for changing things, as long as the interface remains the same. So you're not hampering the design, you're not specifying the complete design, but you're just defining what it has to supply to the other systems in the system.
COLLINS: So initially the definition of the subsystems is a very important exercise.
COLLINS: Does the sense of what it would be profitable to produce enter into this notion of definition, apart from technical requirements? In other words, I assume you don't want to define a subsystem that is not attractive to a contractor to manufacture.
MUELLER: Well, anything is attractive to a contractor if there's enough money. What you don't want to do is define a subsystem that requires more development or more research before development in order to build it. So the trick is to know well enough what technologies are available and how much you can extend them before and then define those subsystems so that they are in fact doable within a reasonable length of time, and reasonable may be fairly short compared to our modern standards. When you say economical for a contractor to do, that isn't part of the program. It's whether it's economical for the government to contract that's important. So you have to worry about the end cost of what you're doing. That's true. As a matter of fact, much of the present emphasis on life cycle costs came in because that was started in the ballistic missile program. That was a new thought at that time that became important because we were looking at having these things deployed for years on end, and the cost of maintaining them was an important part of the total cost of the program.
COLLINS: Let me put the question of economy in a slightly different way. In defining the subsystems, I assume one of the objectives is to produce something for the most economical amount of money, in other words, to do the job within a budget.
COLLINS: How does that affect the exercise of definition of the subsystems?
MUELLER: Well, you know, it's built into the whole process. You have to worry about--and when you talk about costs, generally it is the cost of the number of people involved in working on it that determines in the end what this thing is going to cost. So you end up finding the simplest and most straightforward way of accomplishing what needs to be done. Now, that sometimes requires some sophistication, but more often than not it's finding the simplest solution that is the most economical, and that's why solid rockets were chosen because they are inherently simpler, not necessarily more reliable.
COLLINS: Resuming after a brief pause. One of the key elements, as you suggested in defining the subsystems for a particular weapons system, is knowledge of the state of the art, how far you could extend the art in a particular area, or whether you could take something off the shelf. How did the Ramo-Wooldridge staff keep current on this stuff? They were not making hardware themselves. They were not intimately involved in working with the state of the art by the nature of the Ramo-Wooldridge business as I understand it. How did they keep current?
MUELLER: Oh, most of the people in the organization had backgrounds in research and development. You could ask the same question about how a university professor stays current with the state of the art, and clearly he doesn't do it all himself, but he keeps track of what's going on through his colleagues. And we did the same thing at Ramo-Wooldridge. We were in many instances forcing the state of the art. But basically, the people involved had all worked for many years at the forefront of technology. That's how they got selected. So it will only succeed when you have a very competent group who are more competent, at least as competent as anybody they're working with in the various companies that are working on the project and the various governmental organizations that are in the review process.
COLLINS: Is it an individual's responsibility to keep current, or did Ramo-Wooldridge recognize the importance of this and provide certain mechanisms for keeping up to date, or did it just evolve as a natural process during the work?
MUELLER: The people were the leaders in their fields, in large measure, so we didn't have to create a process to do it. They were doing it. That was part of their nature. That's how they lived. They would have been deprived if they hadn't been at the forefront of knowledge in their particular areas of expertise.
COLLINS: I don't know whether you can generalize, but how did you make a decision about when it was appropriate to extend the state of the art and how much to attempt to extend it?
MUELLER: Well, basically, we tried to analyze, in a qualitative way, what the existing technology would do for you, and if that did it, if that accomplished your objective, you used that. If it didn't, then you tried to figure out what you could do to improve it. Guidance accuracy, for one thing, was a major activity because obviously you wanted to get the CEP as low as you could at the target areas. So we set up a fair cross-section of teams that looked at each of the sources of error and then figured out what we had to do in order to reduce those errors. And they went all the way from a very extensive measurement of gravitational field of the earth, which it turns out has a reasonable effect on the trajectory, to very careful measurements of the atmospheric variations with time because again, as you're coming back through the atmosphere, that will affect your accuracy, and you've got to reduce the effects of that, and that goes back into the shape of the carrier, the vehicle, re-entry vehicle. And you had to worry about survivability of the command and control system and all that that implies and how that affected the accuracy of the system as a whole because that had to put in the coordinates of the target. You therefore had to know where the target was in a set of coordinates that it understood, and it required time to get this together fairly accurately, which really hadn't been done up until that point in time. So the ramifications of asking that question were considerable. And that led in turn to specifying some developments that were needed.
COLLINS: Such as doing research into how the gravitational field would affect the path of the missile?
MUELLER: And the same for the re-entry bodies, for example, and should it be ablative or what will give you the best accuracy at the end?
COLLINS: In this case or some other case, would Ramo-Wooldridge do some of the basic research necessary to answer these questions or suggest improvements in re-entry design and that sort of thing? Or would that be up to a contractor to do that?
MUELLER: Whenever we had the individuals who could do it, we put them to work doing it. We drew heavily on the university structure as well in some of these areas, like the gravitational anomaly.
COLLINS: Would that work in the same way--for example, you served as a consultant for a while. Would it be on a consultant basis, or would it be a formal research contract to a university?
MUELLER: It just depended. Both. But usually it was a consulting basis because that was quicker. It takes time to enter into a contract, so it's usually easier just to hire a guy as a consultant for a while. We would co-opt them as well, if they were smart enough.
COLLINS: Co-opt them in what sense? Get them interested in working for Ramo-Wooldridge?
COLLINS: In this early period, did Ramo-Wooldridge, besides doing ballistic work, they did take on a little bit of work for NASA?
MUELLER: Yes. Well, when it started, of course that was an Air Force project to start with, but then when NASA was formed, it was transferred to NASA, but we did a fair amount of NASA, as it was being formed. We also worked for the Air Force in their open--well, really and in their covert activities as well-- so we did provide them with support, as Aerospace was growing, for a number of their programs. Then we went off and began to compete directly for work once the Aerospace was formed and we were turned loose to compete.
COLLINS: In this earier period, what were the differences between doing work with NASA and doing work with the Air Force?
MUELLER: You've got to remember, NASA was just being formed at that time, and they had the basic problem that they had Wernher [von Braun] down at--they had the old NACA group, they had Wernher and his group at Huntsville. They had a bunch of internal conficts because the NASA centers have never worked well together, you know, as a total. They work well individually, but they jealously guard their independence in thought and structure. And they really had not been ever set up to do major programs, and so their ability to manage a very large program structure was quite limited. Whereas the Air Force had been doing major programs. Remember, the NACA used to depend upon the Air Force to do their contracting for them and to get the systems built. And Marshall, Wernher's group, was used to building their own things, so they didn't have a great deal of use in the development process for external contractors. They used them to build things after they'd been designged and the first several built by the folks at Marshall. And so that was a different culture than NACA or than the Air Force culture. And what's ahead in the case of the Air Force, when we were working with the Air Force, was a set of dedicated people on the government and on Ramo-Wooldridge's side, dedicated to getting the job done. In the case of dealing with NASA, they didn't have a structure they were able to interact with that effectively, to begin with.
COLLINS: When NASA took over the Pioneer program, how did that work? How was TRW able to work with them? What was the character of the relationships?
MUELLER: Well, of course initially, we were essentially just -- well, Harry Goett was running Goddard at the time, and he set up a couple or three guys and sent them out to work with us, to look at what we were doing, and essentially and in principle be the manager. But, of course, the Air Force was still managing in the sense that they managed because the funds were coming from NASA through the Air Force to us, and so we were pretty much running full out on our own. And about all we did was get criticized because we weren't as successful as we should have been. So there was not very much in the way of constructive direction from NASA on the Pioneer program. And that's because they hit it at midstream. And then later, they began to develop some process for managing programs, but it never got as well developed as the Air Force Ramo-Wooldridge approach and still isn't, for that matter. We, I think Apollo was by far the best managed program, but it was really forced on the NASA organization. It wasn't something that they had built up over time and, that is, the basic approach they now use was formed during the course of the Apollo program. But like most such structures, unless you have people that really understand why these things were done that way and what needs to be accomplished, it tends to drift into being form but losing its functionality.
COLLINS: That's a very interesting observation. You mentioned the Ramo-Wooldridge approach to project management as very sophisticated and developed and, in some sense, of a higher order than even NASA's enterprise under Apollo.
MUELLER: I didn't say under Apollo because we picked up all that was good of Ramo-Wooldridge technique and put it into Apollo.
COLLINS: I see. Okay.
MUELLER: But you asked about earlier on, and the structure within NASA didn't really exist, not too surprisingly. You had research centers and one development center, Marshall being the development center.
COLLINS: This is perhaps a bit of reiteration, but what were the key elements of the Ramo-Wooldridge approach? You mentioned the definition of systems, interface documents. What else would characterize the Ramo-Wooldridge approach?
MUELLER: Well, design specs. You had a systems spec, then you had subsystems specifications, part of which was the interface and part was the design of the subsystem itself, and those were kept current. Very important. You didn't just have a systems spec at the beginning and leave it there. Every time you made a change somewhere, it had to iterate through that entire specification structure.
There were test programs, test documents, so that the test programs were defined, and the results were defined in advance of this work you were testing and what you expected the results to be. And if they differed, then you did something about it. There were reliability criteria set up and a reliability budget developed which forced people to take a look at their designs with respect to would they be reliable over a period of time, would they work.
And we had an operations group, a set of operations planning, plans for operations which required the designers to understand what the operations were going to be and what the requirements of the operators were going to be on the design. And that whole structure then was created as part of the process of carrying out the program, and that was essentially the paperwork that provided the basis for, against which you could measure progress. It was a baseline set of documents, which were living documents. They changed as you learned more about the system and knew how you had to change it. You kept changing them.
Then you had assigned very capable technical people to be responsible for this thing, and you had people who--there was one I forgot to mention, and that's program control, which included not just interfaces but in-depth scheduling of every piece of the development. And we developed the mechanisms for measuring progress against the schedule so that you really were in budget, so you really had a program control organization that was forecasting problems before they occurred and not just a set of bean counters that kept track of how much you'd spent, but some very creative people there to keep you out of trouble and to keep you within budget so that you could apply resources early on to solve problems that otherwise would keep the whole program from moving forward. People don't realize that a little problem down in one of the subsystems can drive the cost of the program up tremendously because everybody else is continuing to work but not producing as much as they could.
COLLINS: I guess a brief example of that is the guidance systems that Northrup is doing on the MX.
MUELLER: Yes, sure. And you can afford to put any money you can think of into that to get it solved because the rest of the costs are going on anyhow. You've got that marching army that's just using up money at a great rate, and you have to keep them there with this thing, can't turn them off and turn them on again. Does that answer your question?
COLLINS: Yes, it did. It suggested another question about the representatives you had at the contractors. It sounds as if these people had to be skilled not only in technical areas but in understanding the management requirements of the program, as you've outlined, especially in terms of the program control function.
MUELLER: Yes. They did. They did, but one of the characteristics of the Ramo-Wooldridge structure was that you had somebody back in LA who was responsible for a subsystem, and he spent about half his time out at the contractor's and traveling around making sure that the other contractors were, in fact, meeting specs so that this contractor could work. And usually on a major subsystem you had five or ten people working that problem in LA, plus one or two people at the contractor's plant. The people at the contractor's plant were more eyes and ears than they were trying to direct the contractor. The direction came from LA, not from the local.
COLLINS: I guess my point was, they were part of the feedback mechanism.
MUELLER: Right, exactly, and a very important part.
COLLINS: One thing that distinguishes NASA from the Air Force is this strong interest in developing an in-house capability to evaluate contractors. How did the Air Force know that Ramo-Wooldridge was doing a good job? To what extent were they able to interpret and evaluate the overall effort?
MUELLER: They didn't have to. One of the problems we have today is that at least our Congress and a lot of people in the government think that they have to be as expert as the people they're hiring to do the job. And the Air Force at that time didn't have any such illusions. They knew that they had a set of competence and they knew that Ramo-Wooldridge had a set of competence and there was mutual trust and understanding. A rare and marvelous time in today's climate in Washington, but it sure improves progress no end. Why did they need to? That was what they hired Ramo-Wooldridge for, and why would the Air Force have to be able to independently evaluate the guys who were hired to do the evaluation for them?
COLLINS: How would you then compare the NASA approach, where there was this strong interest in being able to essentially do the technical direction themselves?
MUELLER: Well, that's fine because NASA had the technical competence, at least initially, to do the technical part of the thing quite well. Their problem was they had zero competence in managing a program. The very strength that the Air Force left to the Ramo-Wooldridge team was lacking in NASA.
COLLINS: What was your reaction and the reaction around Ramo-Wooldridge when Sputnik happened?
MUELLER: Well, since we were very aware of some of the work the Russians were doing, we weren't all that surprised. I guess if there was any surprise, it was the reaction of the nation to Sputnik. We had been working, as had Marshall, on our first of the satellites for some little while. It was at the thought process stage. So we knew it was possible and had been trying hard to be the first ones up. Probably better that it worked out the way it did because it sure got the nation stirred up and created an opportunity to do something very extraordinary. But once you had a ballistic missile, you knew you could do something in the way of orbiting satellites. But, of course, at that time Ramo-Wooldridge was in the black area, so they weren't talking about what they could do, and then a whole host of things all sort of churning around, to play catch-up.
COLLINS: Was Ramo-Wooldridge involved in any of the planning or technical direction for the early reconnaissance satellites?
MUELLER: Yes. Yes, that was part of their operation.
COLLINS: Discoverer and --
COLLINS: One other thing that stirred the public mind in this period right after Sputnik was the question of effective management in the nation's space and missile programs. There was the Gaither Report regarding the so-called missile gap. There was this very strong concern that somehow things weren't getting done properly. Did any of that public discussion have any impact in the TRW community?
MUELLER: Well, of course we had people working with Gaither on his report, so we didn't have any direct impact. The principal thing it did was to force the creation of NASA and develop a civilian program. Then there was a whole set of discussions going on back in Washington as to whether we shouldn't just turn the Air Force loose and let them go do this job, as they were quite capable of doing, versus creating this civilian organization, which didn't look like they would be capable of doing very much, at least in the near term. And the Navy was vying for being the space-faring arm of the DOD, and they initially won, only, small problem, they didn't create something that would fly. Other than that they'd be the leaders in space activities.
COLLINS: But this general hubbub in Congress and the press regarding management didn't affect TRW very much.
MUELLER: Oh, yes, we were charging around saying, "You've got to have some management. You ought to use us."
COLLINS: But that was essentially your business.
MUELLER: Sure. And some people thought it was a good idea to use us, and some people didn't. But ultimately NASA decided that they were quite technically competent to do the management themselves and proceeded to do so.
COLLINS: When Space Technology Laboratories got spun off, you essentially changed your business from primarily systems engineering technical production to hardware development, is that correct?
MUELLER: Right. Hard to do. Had to hire...
COLLINS: How was that change effected? How was that ability to do hardware development?
MUELLER: Well, you know, we had people whose background was in hardware, the propulsion group. Our laboratories were quite capable of doing hardware. We had built some capability into it as part of our support activities for the managing programs, so we had at least a rudimentary kind of hardware capability, and we were developing it fairly rapidly. And primarily, in terms of being able to assemble systems and put them together, we could buy the pieces and get them built for us. And so we organized a marketing group and began to go around and compete, find out where people needed something and put ourselves in a position to compete for those activities. And we learned the hard way to win competitions, actually got pretty good at it, probably one of the better proposal mills in existence.
COLLINS: That was one of your key areas of responsibility. How did you go about deciding what areas would be best for you to go into?
MUELLER: We did a fairly comprehensive method survey and we did look at our strengths and we tried to create a plan, a long range plan for our activities. And then we chose the places where we were strong and we could compete effectively and went out and tried to--the best way to do it, of course, was to create a need and then be ready to fulfill it. So we did a lot of work in just pointing out opportunities to potential customers and trying to help them invent their needs, which in this business is the only way you really succeed when all is said and done.
COLLINS: What did you feel, did the top management of STL feel, were the strengths of the organization?
MUELLER: Well, first of all, technically, we had probably the broadest set of skills anywhere in the country when you think of the totality of any one of these space systems, and we had a stronger understanding and a stronger set of people to deal with those sets of problems than existed elsewhere. And so that's where we decided to exploit it. Now, generally, we also tried to team with somebody that had a major hardware capability because it was clear we didn't have the credibility to do that at that point, and so, for example, one of our major activities was to team with GE on bidding for the, what turned out to be the Apollo program. That was later on. We were the systems engineers on that program, at least to some extent, but for satellites, we would propose directly. We were strongly competitive for most of the early NASA satellites and continued to be.
COLLINS: Well, I'm sure there were quite a number of satellites now.
MUELLER: Yes, and we built a number of them, the Able series.
COLLINS: And did you do that in conjunction with another corporation, or did you develop the manufacturing facilities to do this?
MUELLER: Yes, we did a lot. On some occasions, for example, on communications satellites, where we had never been in the field of ? , we teamed with another company, in one case RCA, in another case Motorola. But as we were able to win contracts, we began to build the hardware capability until today we probably have as good a hardware capability as exists in the industry.
TAPE 2, SIDE 1
COLLINS: So you began to develop a hardware capability. I'm a little unclear, what did TRW see itself doing after STL was spun off and after the Aerospace Corporation was formed?
MUELLER: Well, TRW went off and, you know, there was what became Bunker-Ramo, was one of the elements that Si and Dean had set up as a separate entity to develop electronic, I guess, communications capabilities. They eventually ended up building the Stock Market remote displays that became so prevalent in the Stock Market industry. And, of course, the reason Thompson Products wanted Ramo-Wooldridge was to get some of the technology fed back into their primary businesses and try to get a leap forward, great leap forward in competing in the automotive and other areas in which they were involved. So there was a strong desire on the part of the Thompson management to bring the Ramo-Wooldridge technology over into their basic business, and that worked to some extent, not as well as they had hoped but not as poorly as happened with other companies that tried that because the technology is transferrable only to the extent that you have people who are flexible enough to be able to understand the new business and apply the technology they have constructively to that new business. Just straight technology is hard to absorb. You've got to have somebody there to really understand both sides of that equation.
COLLINS: STL was not just a subsidiary but a fully independent corporation, at this point.
MUELLER: Yes, but of course the major stockholding was withRamo-Wooldridge or TRW. They owned 80 percent of the stock. Twenty percent was auctioned off or given or sold to the employees.
COLLINS: But it sounds to me like the greatest nucleus of technical capability was the people who were spun off into STL?
COLLINS: What then did Thompson get when he bought out Ramo-Wooldridge, in terms of a capability, if most of the good people had gone with STL?
MUELLER: Well, he got Si Ramo and Dean Wooldridge and this group out in the Valley, and I think that's about all.
COLLINS: Did they maintain any kind of a capability in systems engineering? Was that one of the areas they continued to cultivate?
MUELLER: Not particularly. Of course, they had Milt Moore and he would have regarded himself as a systems engineer, and really was, and he was running their Valley facility and that was a substantial one. But what they really were getting, Thompson was getting, was the genius of Si Ramo and Dean Wooldridge and his group of people who were the administrative heads of the Ramo-Wooldridge Corporation. So they didn't get a whole lot, except that's probably the most important thing they got in any event, because the people in STL were not particularly well adapted to working in Cleveland and addressing the automotive business per se. You might better ask, why did they want to put it back in when they spun off Aerospace, and I think the answer was, well, there was this reservoir of talent that was likely to lead to a future business and so, therefore, best they keep it. Otherwise they would have eventually spun it off completely.
COLLINS: I'm lost on the connections there.
MUELLER: Well, you know, if Aerospace had not been formed, then the plan was to spin STL off completely from Thompson-Ramo-Wooldridge and eventually to offer the stock to the public and have it a publicly traded corporation, but with the decision to form the Aerospace Corporation and spin off that government-related business, then Thompson decided that they would take STL and merge it back into the corporation, which they did.
COLLINS: Okay. That bit of history was confusing. Do you know when that reworking of STL occurred?
MUELLER: Well, essentially it must have occurred before I went back to NASA, so it was in the '61, '62 time frame. Actually, it was happening just about the end of '62.
COLLINS: You mentioned earlier about some of the approaches to making STL a reliable corporation and the interest in identifying potential needs of customers, creating needs. Can you cite a few examples of ideas or concepts or products that STL cooked up that were meant to satisfy potential needs of clients?
MUELLER: Well, for example, the Able series of satellites were at least partially created by the work we were doing in defining possibilities for the use of space and working with the folks at Goddard primarily to develop a program approach, which they eventually then let out on a contract. And these were in the form of unsolicited proposals and things of that sort. I had for some time thought that communication satellites (a) ought to be in geosynchronous orbit and (b) ought to be attitude stabilized, not spun as the Hughes approach was. And so we convinced--let's see, the Army was given the nod to build the communications satellite, rather than the Air Force, for some reason which escapes me, but that's one of the things that happens in the DOD. We were able to bid and convince the Army that they ought to build an attitude stabilized satellite, so that we got started. Now, unfortunately, the Air Force was given the task of building the satellite and the Army was given the task of building the communications equipment, and that resulted in the program essentially being cancelled.
COLLINS: Is this the Advent?
MUELLER: Yes. So there was a case where we invented the system but couldn't get it managed. That's one of the problems of being a contractor instead of being in systems engineering and technical direction role.
COLLINS: Was it possible, from contemporary launch technology, to put something in geosynchronous orbit at that time you were proposing this?
MUELLER: Yes. Could do. It was done, as a matter of fact. The Hughes approach for NASA was carried out. We were involved, and we actually bid on that but lost it.
COLLINS: This is what eventually was Syncom?
MUELLER: Yes. Of course, this was a period of considerable turmoil, both within STL and Ramo-Wooldridge and the industry as a whole, so we were working very hard trying to get a broad base of programs going.
COLLINS: What was the source of the turmoil primarily?
MUELLER: Oh, well, we were in transition from being fully an Air Force contractor into being an independent contractor. Sputnik had gone up, and NASA was flailing around trying to get itself in position. The Air Force was fighting for a share of that, and the Army and the Navy were all trying also to become space-- everybody wanted to be in space. We were trying hard to survive in the midst of all of these conflicting desires.
COLLINS: And the whole industry was experiencing this kind of shaking out, if you will.
MUELLER: Shake up. Because they weren't shaking anybody out, they were trying to bring them in, and there wasn't a real broad base of industry talent at that point in time.
COLLINS: Did STL continue to do some systems engineering for the Air Force?
MUELLER: Yes. Yes, in fact they continued to do all of the systems engineering on the Minuteman program and supporting most of the systems engineering on the ballistic missile part of the total Air Force program. They passed the space portion of this, the support of the space activities from the ballistic missiles, over to Aerospace Corporation. And I mentioned Al Donovan. He was the leader of the space activities in Ramo-Wooldridge and went over to head up that activity in Aerospace Corporation.
COLLINS: So you are referring here primarily to the Air Force and satellite program.
COLLINS: I think that covers most of the questions I had about the TRW STL period. Is there anything that you feel we should add to the record here?
MUELLER: Well, as it turns out, there were a couple of things that had some influence later on. One of them was we launched an Atlas down at the Cape which was carrying a probe, and that probe--I'm drawing a blank as to where the probe was going, but I remember very well that we had developed a set of guidance equations for carrying out this launch. I happened to be down at the Cape when it went off, because I quite often went down there to watch and participate in the final countdown sequence, and we had assured the Air Force that, yes, these equations would work. We'd patched together two sets of equations, one for the launch phase and the other for the oribtal phase. You know, you had a drift and then a second firing of the system, and that second firing put it into orbit. And in patching the equations together, it turned out that--which we didn't know at the time-- it turned out that we had used one symbol in one equation for a smoothing function, and in the other equation it had been simply a symbol and didn't have any smoothing attached to it. And when we fired the vehicle, the one that didn't have the smoothing function in it was the dominant one, and the thing began to oscillate wildly and eventually went out of control. And I happened to be down there and couldn't quite--it was clear what happened, but why it happened was a little bit of a problem, and I got tagged with trying to find out what happened, and that's one of my first experiences in trouble shooting software. And eventually it turned out to be just simply a matter of a symbol meaning one thing in one part of the set of equations and another thing in another part and the computer not being smart enough to know the difference.
COLLINS: We talked earlier about Ramo-Wooldridge and STL in the systems engineering activity, assessing the state of the art and where appropriate attempting to move it forward. To what degree did you do that in the area of computers?
MUELLER: Oh, we were probably the first utilizer, we were way in the state of the art of computers, later in understanding and in promulgating different computer designs. We made the decision, for example, to go solid state computers fairly early on, even though they weren't proven at that time. It's hard to believe, but in 1960, 1955, when I first looked at it, we were still using vacuum tube computers because that was the more reliable mechanism. By 1956, the first of the solid state computers were being built by a group up in Milwaukee. We went up there and looked at it and said, well, that's the way to go, and so we funded the first of the solid state computers. It eventually became Burroughs. These are the ground guidance computers, and then we also started the work on inertial guidance systems with-- of course, those were solid state--and built the first of those or contracted for the first of those. We also developed some alternative sources to MIT for inertial components.
COLLINS: Are you referring to Stark Draper's laboratory?
MUELLER: Right. So we had a good working relationship with Stark but also an arm's length kind of thing which he never, we didn't completely relinquish our designs and capability to him.
COLLINS: Why was that? Corporate proprietary concerns?
MUELLER: No. Not that at all. It was just that we felt that he was more conservative, and we needed, for our space activities, we needed a more creative approach, so we introduced some additional competition.
COLLINS: Who else?
MUELLER: Well, at that time, General Motors AC division, L.C. Delco, were brought in.
COLLINS: So this was part of your approach, pursuing parallel paths.
COLLINS: With technology.
MUELLER: Very important to make real progress is to have several teams working the problem. It's remarkably creative; innovative thought comes out of that.
COLLINS: Did you have a core group within Ramo-Wooldridge who were actively researching problems in inertial guidance?
MUELLER: Yes. Bob Burnet and Bob Bennett were the two leaders in that activity. I guess so was--well, there were a number of guys. But we had a very substantial group because, you know, although our first guidance systems were radio, it soon became apparent that if you look at the long term suvivability of the system, an inertial guidance system would be a heck of a lot better. If you have a radio guidance system, it's one target that you can take out, and then that gets all those missiles that were controlled by it out. If you've got an inertial guidance system on every missile, you can't take it out that well. You can't home on it very well either.
COLLINS: In this particular case, how did it work? You wanted to work with the Draper Laboratory. You worked with Delco. You had your own in-house capability. But you didn't fully--
MUELLER: We never developed a capability for building gyros, for example. That just was not something that we could justify.
COLLINS: But you developed the theoretical.
MUELLER: Oh, yes.
COLLINS: But you didn't fully share your theoretical understanding with Draper.
MUELLER: Oh, sure we did. Yes, heavens, yes. We weren't in a proprietary segment at all. Never were while I was there. But that didn't say we didn't do some independent thinking and argue with the folks back at Draper Lab.
COLLINS: Okay. I guess I got a little confused by your remark about arm's length with Stark Draper.
MUELLER: Well, I shouldn't have said arm's length, but independent. You recognize that when you're dealing with Stark, you're either on his team or you're not. And that, we weren't on his team, and so therefore we were not, as far as he was concerned. But we didn't hold anything back, and we worked closely with them, as close as we could.
COLLINS: But you had a sense that perhaps he was a tad conservative, and you needed to develop some independent theoretical capability.
MUELLER: Yes. Well, we needed a parallel approach, because look, no one had ever built an inertial guidance system for a ballistic missile at that time, and there were lots of different thoughts about how one might go about doing that. We looked at a whole set of substitutes for mechanical gyroscopes, for example, and began some of the work that's now current, supported some of the work that's now current.
COLLINS: I asked you earlier whether there were any other things you wanted to add to this, and you mentioned your experience down at the Cape and problems with the navigation equations. What was the significance of that experience?
MUELLER: Well, it taught me some of the problems associated with software, in terms of the complex set of programs that one needed to carry out a ballistic missile, and also taught me a good deal about how to find out what the problems were and what one had to do in order to assure that there wouldn't be a repetition of said problem. It was a very intensive learning experience.
The other arena was in getting into the building of the telemetry system for the Pioneer program where, I guess, there are actually two in that system, the telemetry system and the other one was the problems associated with the second stage firing, both of which were mysteries and, in fact, that second stage firing thing persisted for a very long time. We lost three vehicles in that process and, eventually, we did a ground test. Dick DeLauer finally did a ground test. He was responsible for the second stage and found out what the problem was, which was shrapnel from the explosive bolts that separated the stages shooting out through and wrecking the engine consistently. And that taught me that, boy, if it's possible to test something on the ground, you test it on the ground. You don't test it out in space if you can avoid it. The corollary to that is, though, you don't want to be testing piece-wise in space. You want to test the entire system because who knows which one's going to fail, and you'd better have it all together so that whatever fails, you have a reasonable chance of finding the real failure mode, not just the one you were looking for. Lesson learned.
The other one was with the digital telemetry which was, I guess, really one of the first digital telemetry systems set up, probably second or third. But, anyway, building that and getting it to work taught me a good deal about how difficult it was to get contractors to do what they should do--you know, subcontractors--and introduced a much more rigorous set of controls for controlling the basic design and manufacturing process than we had had before, which we eventually carried over into the Apollo program.
COLLINS: This question about subcontractor performance came up while you were trying to develop hardware under STL?
COLLINS: Is that a problem that you had picked up on when you were part of the systems engineering effort at Ramo-Wooldridge, as a problem that the prime contractors had with their subcontractors?
MUELLER: It didn't really come out as clearly as when we were doing it ourselves. There's something about first hand experience. Yes, it's been a problem all the time, trying to find out what a subcontractor or a prime is really doing, because usually the prime doesn't know either or didn't used to. They waited for the thing to get delivered. When it didn't come, then they panicked. Well, we tried to find out where they were in the process early on. And that increases the administrative burden, but it avoids an awful lot of problems later on when the pieces come together and they don't fit.
COLLINS: Going back to the Ramo-Wooldridge systems engineering activity, you mentioned that headquarters staff would go out and travel to the various companies. Did this include subcontractors?
MUELLER: Yes. Sure. I remember going up to Marquardt, for example, and talking to them about their attitude control engines and what they were doing and why. Yes. Now, initially the prime contractors were very reluctant to let us go anywhere near their subs, so it took a fair amount of reasoning by us and the Air Force to get us access. And eventually we were trusted enough to be able to do that, but it took a few examples of things that didn't work.
COLLINS: This was a concern on the part of the prime contractors that it might reflect on their management?
MUELLER: Well, abrogating their authority or circumventing their authority, and they were afraid that we would be directing their sub and running it and doing something that they wouldn't agree with.
COLLINS: One question, what relation did the mechanisms that Ramo-Wooldridge set in place for their systems engineering effort, interface documents, program control function, that sort of thing, have to such project management approaches as PERT? Is there any relationship there?
MUELLER: Well, you have to ask yourself what PERT [Program Evaluation & Review Technique] is. We had, of course, in place schedule tracking and critical path analysis about at the same time as, well, probably prior to the time the Navy invented the name PERT. But the basic problem with any of these things is getting the people doing the work to use it and contribute to it. And I have been to many contractors, and they have this marvelous PERT diagram up there with all those boxes and things. So I finally get the guy aside who's really doing the work, and I say, "Now, how do you really keep track of it?" And he has a desk drawer he pulls out, and there's his schedule. And it may or may not have any relationship to this master schedule that they show the customers. The PERT isn't of any use, nor is any other system, unless the people that are actually doing the work use it.
COLLINS: I'm sure that applies about equally to NASA as it was carrying out its activities.
MUELLER: Oh, yes, indeed. And that's the trouble. You get into this form, and people like a PERT diagram. Well, that's good, but there are lots of other ways of looking at progress and finding critical paths. That's as good as any I know of, but it gets unrelated to what's really going on in the program. And the trick is to keep what's going on in the program reported, not a form up there that keeps people off your back.
COLLINS: In the Ramo-Wooldridge situation, you accomplished this through your representatives at the plant and through these frequent visits to contractors.
MUELLER: Yes. You can actually see what's happening, then, and know. That is, if you have the schedules, you know whether or not they're meeting their milestones.
COLLINS: Well, let's talk a little bit, then, about how it is that you became aware of the associate administrator position at NASA and were recruited for it.
MUELLER: Well, Dave Wright was a friend of Jim Webb's, and when Jim got into trouble with Brainerd Holmes, or Brainerd Holmes got in trouble with Jim, I guess would be a better way of putting it. You know that the problem occurred there that Brainerd decided that he was more competent to run NASA than Jim Webb was and went to Kennedy and said, "Look, you've got to make me the head of NASA," and Jim was there and he said--I only hear this third hand, as it were--he said, "Well, you know, Mr. President, you appointed me and you can not appoint me and that's up to you, but you do have to make a decision, and obviously I can't continue to work with Brainerd Holmes." So the President obviously said--well, since Lyndon Johnson was a backer of Jim Webb, and Brainerd didn't have any backers to speak of, why, the President said, "Well, Jim, you're in charge, you decide," and Brainerd said, "I resign." So that left Jim with the problem of getting the Apollo program carried out because that was their prime objective at the time.
COLLINS: Were you aware of the circumstances before you came to NASA?
MUELLER: Oh, yes. Sure. Because there was no secret about it. Brainerd got himself fired and was fairly vocal about it. But at the same time, the Air Force had been making a big push to take over the Apollo program because, after all, they had a competent management system, and it was fairly clear that NASA didn't. And so that created a problem for Jim, in addition to the Brainerd Holmes problem.
The combination then of being in a position where Congress is threatening to reduce the budget, the program being in disarray because of its management structure, and the Air Force pointing out these things and saying, "Well, don't you think we ought to take it over, make sure it comes out right?" led Jim to talk to Dave Wright. I assume he talked to a lot of other people, but Dave was the guy that he finally picked on. And he asked Dave to volunteer some people who would be acceptable to the Air Force to take over the Apollo program and, at the same time, who could manage it and run it. So Dave asked Louis Dunn who should go back, and I guess the first guy they asked was Rube Mettler, but Rube said he couldn't do it, and I'm not sure they really asked him anyhow. But they asked me and they asked Ed Dahl to go back. Ed Dahl, he was running the, I guess, the ballistic missile portion of program management at the time. I was the chief technical guy, vice president for R&D, and he was the vice president for ballistic missile programs. So they asked Ed to go back. They asked me to go back and talk to Jim Webb and see whether we were interested in taking on that. And we both spent a few days back there, chatting and seeing what the lay of the land was. It turned out I had a number of friends back there, actually, that I'd gotten to know in the course of my work with ballistic missiles, so I knew a lot of people inside the Apollo program and, of course, I'd been in charge of the bid for the GE effort, so I got to know a fair amount about the Apollo program as well. And I guess Ed, before Jim Webb had to make any decision, Ed decided that he wasn't interested in moving back to Washington. So Jim got me aside and took me out to his house for dinner and discussed what I thought could be done and would I be willing to undertake it. And I said, well, if I could get this organization straightened around, if we can change the organization so it will work, I'd be willing to undertake the task. And so over the next month, why, I worked with Bob Seamans and company to restructure NASA, to make it a workable system, and that's when they shifted the three centers over to report to me directly, as well as the local group at Headquarters.
COLLINS: This was before you were actually hired on that you undertook this study?
MUELLER: No, it was after I'd agreed to go back but before I went on board. There was a month and a half there when I was closing things up at STL and working my way out from underneath that.
COLLINS: But you had not officially been assigned a position in NASA yet.
MUELLER: Not yet, because we created the position. In addition, that's when the associate administrators were created. And so I did. Not without some trepidation, but it was a challenge, and it looked like it was something that needed to be done. Not the least of the challenges was that it cut my salary in half. I'm not sure I would, under today's circumstances, I'm not sure I'd want to take on that task.
TAPE 2, SIDE 2
COLLINS: What was the character of your initial discussions with Mr. Webb? You said you were in Washington for several days. Did you meet with him before you went to dinner at his home?
MUELLER: Oh yes, I'd met him and chatted with him just briefly, and he explained what his view of the problems was and what he would like to accomplish. And Joe Shea was one person whom I knew quite well from his work at Bell Laboratories, and he is a most impressive individual, one of the brightest people that I've run across in life and one of the most competent. So through him and through other people, I got a fair understanding of what the management structure was like and what the problems were.
COLLINS: This is kind of an informal survey on your part.
MUELLER: Right. And, of course, I had some ideas about how one ought to organize to carry out a program of this sort, and so I set about figuring out what structure could make it possible. And it was clear immediately that you had to have the people that were doing the work report to you or else you just weren't going to be able to get there from here. And that's what Brainerd discovered but really wasn't able to bring out. The other thing was, it was clear we had to get some focus on the program as different from the institutional structure, and they had been all mixed up before. And as long as you had the center directors worrying about the institution, it took precedence over the programs, and you needed both. So that was one of the things that I set up immediately, which is to set up a very strong program management chain, in parallel with the administration or institutional chain of command, and make sure there's communication up and down both chains and in depth.
COLLINS: I think we can explore the mechanics of that in a little more detail as we go on.
MUELLER: Now, with respect to Webb, I had to explain to him that I needed to have that authority to deal directly with the center directors if I was going to be able to do the work. And he agreed, and he was trying very hard to figure out how to get it done. That was his major objective. And so, we went forward with that restriction. I remember one of our conversations before I went back there. He called me and said, "Where do you think we ought to put the control center? Should we put it at the Cape, or should we put it down at Houston?" They'd made the decision to set up the Manned Spaceflight Center at Houston by that time. And I said, well--It was tentatively scheduled. It was scheduled to go to Houston--I didn't know enough to make a decision, but I didn't see that it was crucial that it be one place or the other. So the decision was left that they were going to move it down to Houston.
Then there was at that time, also, a lot of discussion of the lunar orbit rendezvous, earth orbit rendezvous, and so on. And he wanted my opinion on that. I said, well, I hadn't been in it deeply enough to offer an opinion, but as much as I knew, any one of them would work. And so it was a matter of some in-depth analysis as to which one had the highest probability of success, which was the simplest one to carry out. And I still think that was the case, that any one of the three or four approaches that were thought about would have worked, including the direct flight and return. But the lunar orbit rendezvous required less hardware than any of the others, that is, the direct flight would have required a larger booster, and so, eventually we settled on that one. I guess, in retrospect, I wish we had done the earth orbit rendezvous and the lunar orbit rendezvous because, although it would have taken somewhat longer and been more expensive, we would, nevertheless, have had a space station in being back in '68, which would have been a constructive thing to have done, but that was a substantially greater investment, and we were working hard to stay within our budget at that time.
COLLINS: Just to backtrack a little bit, when you were still at Space Technology Laboratories, how did you keep abreast of the political shifting and developments in relation to the space program or the defense program? These were volatile political subjects, and the political positions taken on particular programs, the particular character of the programs, had important impact on the corporations. How did Space Technology Laboratory, Ramo-Wooldridge, keep abreast of what was going on in Washington?
MUELLER: Well, Si, of course, and Jimmy Doolittle were both heavily involved in Washington politics and continued to be, so that was one way of doing it. Then, of course, I spent about as much time in Washington as I did in Los Angeles during that time because I was involved with NASA in its organizational setup. I was still doing some work for the Air Force and its operations over in the Pentagon, so I was a bridge between all of those things, to some extent, so I got to know the people and got to know what their problems were fairly thoroughly.
COLLINS: You developed your own kind of informal network for understanding it.
MUELLER: The one thing I didn't have any real contact with was with Congress, and until I got moved into this job, I really hadn't had any appreciable contact with Congress. So that was a new experience for me. But I had had contact with PSAC and knew the people there, so I knew the executive branch reasonably well.
COLLINS: I'm curious, at your initial meetings with Mr. Webb, what it was that he told you the NASA organization needed and what role he thought you would fill for them.
MUELLER: My recollections are fairly limited in that regard because, you know, Jim is a very charming person and a very voluble person, and he generally talks about things in a--he covers a very wide range of subjects, and only occasionally touches on the thing that you are there to talk about. I don't know whether you've noticed that or not.
COLLINS: I've met with him a good many times. Yes.
MUELLER: And, as a consequence, I knew pretty well what he needed, and I think he was really trying to decide whether he had the confidence in me to do what he needed. And I guess he decided that he did.
COLLINS: In these initial meetings, did you go in with a set of requirements for taking on the job? You mentioned this concern about the people who did the work reporting directly to you. Was that a requirement that you had at this very early part in your initial--?
MUELLER: Oh, yes. That was, in fact, the only one. Jim was concerned. He wanted me to understand that I didn't get a car and driver to take me back and forth to work and that, you know, he was riding around in a Checker Cab because he wanted it to appear as though he were being very parsimonious in his work. And he wanted me to know what my salary would be. And he was concerned that there weren't very many perks that go with being a government employee, and he wanted me to understand what those were and what the limitations were. And he did offer, if I needed help, to lend me money. A very generous kind of a guy in that regard. But I don't like borrowing money from my bosses. That's not my idea of the best way to survive in the world. But it was kind of him to offer. And it was a financial strain, too, so I had to really consider whether we could afford it. Ultimately we decided, well, we can do what we need to do for the country.
COLLINS: Were you put in a position of divesting any stockholdings at TRW?
MUELLER: All of it. Well, we actually put it in a blind trust but administered by a bank, but I must say that blind trusts administered by banks are disasters. Not a satisfactory way of doing things. I had all of that kind of thing. Had to meet with some of the Congressmen, not nearly as much as they do nowadays, but I remember vividly my first meeting with Tiger Teague, who was a supporter of Brainerd Holmes, and Tiger explained to me that his view of life was that the authority goes with the position and not with the person, and--you know, typical Army background, the general is always right. But he made it clear that if I wanted to maintain relations with him, I would have to level with him at every instance. His favorite expression was "You can"-- he didn't say it this way, but "You can betray me once but not twice." But, as it worked out, we were compatible and worked very well together.
COLLINS: Was TRW encouraging you to take this position? Did they feel strongly about it one way or the other? Did they want to retain you? How did they feel about it?
MUELLER: Well, they were neutral at best. They didn't encourage me to go, but they didn't discourage my going. I guess they felt that there was a real need for somebody to go back to Washington, or they wouldn't have volunteered me in the first place. But since Dave had volunteered me, Dave and Si really, they weren't going to discourage my going. But there was no financial encouragement to going. They didn't--I guess the one thing they did do was transfer me back to Washington so that they could pay the costs of the move, once I made the decision to go. But that's about the only financial encouragement that we got.
COLLINS: So you had this initial few days of meetings with Mr. Webb and other NASA staff, your own kind of informal discussions with people about what the job would be like, what the problems were, and it was on this initial visit that you did indeed make up your mind to accept the position.
MUELLER: Right. Well, I'd gone back knowing that that was a possibility, so I really felt that if I could get the proper set of conditions, that I would take it.
COLLINS: And those conditions were really just one, the question of --
COLLINS: -- direct control over the people that worked with you on the program.
MUELLER: Right. Very important, however.
COLLINS: So what followed after that? How did things work? You mentioned that you began in concert with Dr. Seamans to begin to work out in detail some of the organizational changes that would be put into effect.
MUELLER: Right. And while those were moving along in parallel, I also contacted the center directors and got them to come out and visit me in Los Angeles to tell me about their view of what the problems were and what needed to be done. And so I had a meeting with each of them before I was actually appointed.
COLLINS: This is a joint meeting, or they came out individually?
MUELLER: They came out individually. It's hard to get them to a joint meeting at that stage. And I also talked to some of the key players, George Low and Joe Shea, and got their views of what the problems were.
COLLINS: What were you hearing from the center directors when they came out to talk with you?
MUELLER: Well, that Headquarters was really interfering with progress, and they needed more money and less direction. And these other center directors were great people, but they really didn't understand the program well enough. And there needed to be some better coordination, mechanism for coordinating activities. It was sort of a mixed theme. They knew something of their own problems, but they really didn't have any idea about how to solve the total problem, and they were frustrated. That's about the best way of saying it. Bob Gilruth, of course, was up to his eyebrows in trying to create that center down in Houston, at that point in time. He had a feud going on with Harry Goett, and that had not been completely resolved because at one time Harry Goett was in charge of all manned spaceflight down at Goddard, and nominally Gilruth reported to him. But Bob is an individual who wants to do his own thing.
COLLINS: So there was a sense from the center directors, to generalize, that there was interference from Headquarters.
MUELLER: Headquarters was all screwed up.
COLLINS: And difficulties in working with the other center directors.
MUELLER: Right. Well, of course, they reported to Bob Seamans, and their direction was coming from Brainerd Holmes, and their money was coming from someone else. And Headquarters represented an opportunity to divide and conquer, but it wasn't getting the program going along very well. And also, each of them had developed over time a constituency in Congress, and there was a tendency, then and now, to at least keep their Congressional local people informed about what they were doing and what their problems were. The latter creates no end of difficulty on occasion. And so one of the first things we did was to try to get everybody to talk about the same set of problems anyhow, and not individualize it particularly, and then to fix the problems as rapidly as we could.
COLLINS: What did these meetings do for you, in terms of your thinking about the reorganization of the NASA Apollo effort?
MUELLER: Well, primarily it was just to get a feeling for how difficult it would be to reorganize, and to begin. You know, when you're dealing with, for example, Bob Gilruth, who's running a center, and you're just coming in from industry, he looks at you with a fair amount of skepticism about what it is that you're going to be doing to him and what can you do for him, and so these were sparring matches, more or less. I happened to have known Wernher better because we'd been competing in the space arena. And so I had a considerable respect for him, and I guess he had some for me, although again, he was trying to figure out how much control he could continue to exercise from his vantage point as center director. The most straightfoward one was Kurt Debus, who said, "Look, we've got a real problem here. These center directors aren't talking to each other nor to me, and I'm supposed to fly these things when they get here. You've got to do something so, at least, we're going to build something that's flyable." He had--you know, that was a newly created center that Wernher had spun off really from Marshall to bring together the activities at the Cape, so that when we became operational, you had a place to operate from that would work. And yet as soon as Kurt got down there, why, there began to be some parochial differences between the folks back at Marshall and the folks down there. The folks down there had a certain set of requirements, and the folks at Marshall thought they knew better. And over a period of about a year, why, they were beginning to drift apart. And, of course, the folks down at Houston wouldn't even talk to them, so you had a little bit of a "not invented here" in spades throughout that organization.
COLLINS: You're beginning to get a flavor of this from your intial meetings. How did you feel about walking into such a situation?
MUELLER: Well, I knew about it from the beginning because, you know, when you're selling into these things, why, you've got to know what the proclivities are of the people. So I thought we needed to do something drastic to get their attention, and one of the first steps was to get them to report to me, and the second step was to introduce the idea of a program office chain of command that was separate from the institutional chain of command. And don't think that wasn't a traumatic experience for a center director to have somebody tell him, "Well, your program office is going to report to me, not to you." "And to you" is what I said, and the way it really works, but he's got two bosses, not one. And that had a significant impact and took about a year to get to work.
COLLINS: You mentioned, during this period before you actually came to NASA, you met with the Center directors, and you also worked with Dr. Seamans in establishing the reorganization plan. How did you go about educating yourself further about the NASA organization so that you could properly reorganize?
MUELLER: Well, I knew how I wanted to organize when I started out, so what I was really doing was trying to find out who the key players were. So that's why I spent some time chatting with people, largely by telephone. And then I had some support inside of STL. I had some of my friends working on organizational schemes and tried out the one I finally chose on Sala Loom (?). And Ed Dahl had spent some time back there. I worked some with him on it. They all told me I couldn't do it, the scheme I finally put in place, that it wouldn't work because the center directors wouldn't agree to it, and they would sabotage it. But I didn't think there was any other way of making it work, so I decided to go ahead with it anyhow.
COLLINS: Were there any models that you drew on in that sense from your previous experience with the ballistic missile program? Were there any instances of this kind of dual reporting structure?
MUELLER: In a real sense, the whole ballistic missile program was set up that way, with the Air Force being the contractual arm and Ramo-Wooldridge being the systems arm. And in addition to that, we had the program office structure inside Ramo-Wooldridge and then the laboratory structure, the technical expertise, at least the broad technical expertise, over in the laboratories. And we were working together to solve the problems that we ran into, so in a real sense, that's how we organized NASA. We had a program office structure and main technical competence in the basic center structure. And we had an operational group in Ramo-Wooldridge which was being created at NASA, as well. So yes, we used everything that we could think of. I used everything that I could think of that was good about Ramo-Wooldridge structure and moved it over, modified for the circumstances, to NASA.
COLLINS: When you were referring to an operational group, you were referring to what sort of activities?
MUELLER: At Ramo-Wooldridge there was a whole group that worried about the logistics, the support, stuff like that. How do you actually fly these beasts?
COLLINS: You mentioned that you had talked to other key people besides the center directors. Can you give some examples of individuals you talked to?
MUELLER: Sure. Seamans. Somewhere in that time frame I spent some time talking to Homer Newell and to Jim Beggs, who were eventually associate administrators. When the reorganization came out, they became the two other associate administrators. At that time there were only three associate administrators. And I talked to the--I'm trying to think now, who in PSAC I was talking to, but several of the PSAC people, who had a different view of what was going on. And a fair number of the people at Headquarters.
COLLINS: When you say PSAC, you mean you were talking with Jerry Wiesner or...?
MUELLER: No, it was some of the people on the PSAC. I had known Jerry Wiesner, though. But I didn't talk to him at that time. The day I arrived in Washington, why, he collared me and told me about the fact that we were going about it the wrong way. He was not in favor of lunar orbit rendezvousing. He was in favor of the earth orbit rendezvous approach.
COLLINS: PSAC generally in this period was relatively critical of NASA.
MUELLER: Yes. With some reason.
COLLINS: What did you hope to incorporate, from their point of view, as you were thinking about these issues?
MUELLER: Well, what I was hoping to do was to get them to become part of the problem, instead of kibbitzing from outside. To get them involved in a constructive way. And, to some extent, we succeeded. My view is that you want to get help from whoever and wherever you can find it, in one of these things. So I did my best to get as many of these folks involved in helping work out the problems as I could and trying to keep them informed about what the problems were that we were running into so that they at least understood where we were coming from and why we were doing some of the things we were. And by and large that succeeded, but it takes an awful lot of time, energy, and effort to interface with all of these constituencies. It's not a trivial matter at all. And my observation is that it hasn't been done all that well in recent years at NASA.
COLLINS: As you finalized your ideas about the NASA reorganization, how did you present this to Dr. Seamans and how did you go about the business of arguing for its appropriateness, or did you have to?
MUELLER: I didn't have to. I simply briefed Bob Gilruth and Jim Webb and Hugh Dryden about what I felt was the way we needed to go, but I didn't do that until after I had been there for several weeks. The first thing was to get myself appointed and the centers reporting. The second thing was to go meet the Congressional people involved and the PSAC people and then to fly around to the centers and meet each of the center directors and their key people and sit down and spend a day with each one. So I spent three days running to the three centers. And going back to an earlier question, I had several rather informative discussions with Harry Goett, whom I'd gotten to know fairly well, figuring out what needed to be done. But other than listening on these first rounds, I really didn't begin the reorganization, internal reorganization, until I'd completed that first set of rounds and listened to everybody tell me how they felt we ought to organize and what was wrong with the system, outside of their center, that was keeping them from doing what they needed to do.
COLLINS: What were Harry Goett's observations?
MUELLER: Well, I think his view was that Bob Gilruth would never become a team player. And, in fact, he wasn't sure how you could go about getting this thing pulled together as a system as long as you didn't have a team. He was not particularly optimistic about the ability to make this thing go. Too many diverse agendas and not enough direction or ability to direct and control. And then, we had to restructure the program, too, because we had three different launch vehicles we were building and no real understanding of what it took to land on the moon yet or what really was going to have to be done to carry out lunar orbital rendezvous. Relatively nebulous set of schedules, which had the great propensity of changing every time anybody thought about it a little more. So one of the first things we did was try to put together a baseline that we could live with, at that time introducing the basic concept of limiting the things we were going to do, rather than proliferating them, and introducing the concept of all-up testing. With a program of the magnitude of Apollo, which had been in place now for a couple of years, it really wasn't possible to put in a parallel effort within any conceivable budget that we could get. So where we did parallel developments was at the subsystem level, where we had critical design problems that we knew about. But we weren't able to do what would have been nice to do, have two complete approaches. You know, in the best of all possible worlds, we would have had an earth orbit rendezvous mode being developed by one set of contractors and a lunar orbit rendezvous mode by another and perhaps even a third with the direct flight mode, and at least carried those through the design stage.
COLLINS: Did you at any point argue for this kind of approach?
MUELLER: We were cast in concrete by the time I got there, in that sense. But we were still making the decision between earth orbit rendezvous and lunar orbit rendezvous, and a lot of argument about that, for example, and the Nova had not yet been cancelled, although essentially it had been.
COLLINS: During your experience at TRW, you had developed a very sophisticated approach to systems engineering, technical direction. Did you find an equal level of appreciation of these concepts in NASA when you got there?
MUELLER: Well, no, not in general. But, you know, that was one of the catch words of the time, so everybody talked about systems engineering. But George Low and Joe Shea, particularly Joe Shea is one of the best systems engineers I've ever known, and so he understood and was doing a set of basic trade-off studies. He was the one who latched onto the lunar orbit rendezvous as being the way to go and developed the rationale for it and the technical justification for it. And so, we had a small but very competent group at Headquarters who were systems engineers, and they, in turn, began to locate people out in the centers who had that kind of a thinking process. To some extent, people are born as systems engineers, and you find it hard to educate them into being one. But, fortunately, each of the centers had some quite good people in them, and we managed to get them involved and begin the work. George Low was invaluable with respect to Houston because he really knew the folks there. And Wernher, once we got past the first little jockeying for position, was very cooperative in putting the right people from his organization into this little task group we set up to take a look at what we needed to do.
COLLINS: How are we getting along on time?
MUELLER: I think we ought to quit.
COLLINS: This is the end of our discussion. Thank you.