TAPE 1, SIDE 1
MR. MARTIN COLLINS: In advance of the interview, I had sent you several documents from the l960's that related to organization of NASA [National Aeronautics and Space Administration] at that time. I thought we might just look at one paper entitled "Adapting NASA's Organization and Management to Future Challenges: Staff Paper Prepared by the Office of Administration, October, 1963." It's essentially a descriptive document that gives substance to the organizational change that took place in November of 1963. I wondered whether you had any background comments on the document.
DR. GEORGE MUELLER: Let's see, I started to work on this in June, I guess, of '63, before I joined NASA, and was working with the people in Washington and the centers, formulating a structure that I thought would be capable of carrying out the program. In particular it was apparent to me that it was necessary for the management, the program management and the center management, to be under one head if we were going to be able to get the coordination and the singleness of purpose that was essential for a successful Apollo program. And so we set to work to provide the rationale and the understanding of the roles of the various participants in the management within NASA, in July and August of that year, and then had a whole series of meetings trying to be sure that the center directors understood the direction in which we were going, and that Bob Seamans and Jim Webb and Hugh Dryden understood what we were trying to accomplish, and that culminated eventually in this staff paper.
COLLINS: Was the notion of a matrix management, as you were suggesting for the centers, in which the program elements within the centers had dual reporting to center directors and to yourself, was that an idea that had occurred to others? Was this a novel idea?
MUELLER: The original structure that existed when I came in was a super-matrix management, if you will, because the center directors were reporting to Bob Seamans, and the program offices--Brainerd Holmes, for example--were reporting to Bob Seamans, and the theory was that they were able to work through Bob with the centers to carry out these major programs. Essentially the psychology of NASA centers doesn't lend itself to that working very well, because one person controlled the funds, to some extent, the program directors; the center directors controlled the resources. You had Bob having to adjudicate all decisions, and furthermore most of the decisions never got escalated to where they could be made. For matrix management to work, it has to be at a level where the participants in it can directly communicate and be sure that they're all working on the same problem.
This reorganization, where we set up with the center directors reporting to the associate administrators with a matrix within that structure, provided that level of communications and that level of coordination that's essential to successful program completion. Obviously that depends on people as well, so it was necessary to get the right set of players in place. You can't just do it with organization. It's more nearly a problem of getting the right people working on the right problems, and so that was an essential part, but unless you've got a structure that you can work with, you can't get there no matter what people you bring in.
COLLINS: How did the center directors see their role in this new situation, or how did you see their role?
MUELLER: Well, it was interesting. They were very sensitive about their prerogatives, and they were concerned that they would not have a role. It was essential to build a strong management council where the three center directors and I presided each month over a review of the programs, and have the program directors and the project directors at the centers report in each month to this management council, so that everybody knew what everybody was doing, and so that the center directors realized that they still had an important role to play, which was not only to provide the resources for carrying out the programs but also be responsible for the quality and the quantity of the work being done.
COLLINS: Since you had your own kind of reliability offices, is that something different from what you're talking about here? As part of your structure for the program, you had various offices at the associate administrator's level, and then those were duplicated within the center, one of which was reliability.
COLLINS: How does that relate to the center directors, for example?
MUELLER: Each of them had their own reliability and quality assurance office, just as we did, and we had two levels of it at headquarters, one reporting to me directly and the other reporting to the program manager or the program director. So we had a structure that carried it through all the way to the prime contractors, who also had reliability and quality control groups. I would add that we did insist on the carry through of that matrix structure all the way into the prime contractor's organization, so that in a sense we had counterparts that we could talk to in the prime contractors as well as at the centers.
COLLINS: So this tier of five offices with various responsibilities was duplicated in the contractors as well as in the NASA centers.
MUELLER: At least to the extent that it made sense to duplicate them, so that you had independent lines of communication. The important thing was to have threads of communications that went all the way up, not authority but communications, so that you had a counterpart who could follow what was going on in the contractor. Now that kind of a structure isn't feasible for small programs, and you really don't need it then. But when you have a large program like a space shuttle or an Apollo program, you've got to have duplicate lines of communications to be sure that there aren't gaps or major failures, and that worked very well. In retrospect, I think that that was probably the best thing that we did, in terms of a management approach.
COLLINS: I'm still a little unclear on how the center directors went about getting information about the program activity.
MUELLER: Every month before they would come up here, they would go over with their project directors a pre-review, if you will, of what they were going to say up at headquarters, so they had very good communications with the projects in their area and the problems that they were having, not only internally or with their prime contractors but intercenter-wise. So it was a forcing function, to keep everybody informed as to what the problem structure was at the moment.
COLLINS: It might be useful, as a basis for further discussion, to lay out the areas of responsibility of these five groups--program control, systems engineering, test, reliability, quality and flight operations--because I think it's an interesting question to look at how you parse up what are the key responsibilities as you tackle this large kind of job.
MUELLER: Well, let's see. If you think of program control, it is an active element. It isn't just simply keeping track of how the money is being spent and how the schedule is going, but it is also trying to forecast where problems are going to occur from looking at trends of progress in the programs and looking at the trends in budget allocations and trying to keep track of what was actually happening over in engineering and the changes that are coming through. Systems engineering is the other side of that. They're the two key elements, if you will, in program control because systems engineering, of course, is influencing what's actually being done in the program itself, and it's responsible for making sure the designs are adequate and that all of the contingencies have been examined and trade-offs made to get an optimum program. A lot of trade-offs have to be made with respect to schedule versus the kinds of changes one would like to put into the program and make sure that those are properly evaluated and understood.
It's interesting that it goes all the way from geodesics to really understanding what the base of the guidance system is working from, and it was interesting as an aside to find that Marshall had one set of geodesic standards, and Houston had a different set. One was based essentially on the earth coordinate system, and the other was a space coordinate system, and they were slightly different, so one had to be sure that they meshed when the transition took place in going from earth to orbit.
COLLINS: Let me just ask here, what was the principal area of interaction between these two groups?
MUELLER: You mean systems engineering and program? Actually they worked together when you're thinking of interaction. In all of the design reviews, both program control and systems engineering were there. They represented different basic skills, but they worked together to really understand the program and its implications. All of our program control people were engineers, and all of our system engineers were engineers or physicists or something, and so they were complementary to one another and worked very effectively in that regard. The reliability and quality assurance group was set up to both assure that we were doing everything we could to have reliable equipment--and that meant setting up standards for the acceptance of components, all the way from the acceptance of components to the final approval of designs--and then through into the review and assurance of the testing.
Now, we set up a separate test group because so often the engineers who design the systems don't pay an adequate amount of attention to its testability and to the testing that needs to be done to be assured that their design actually would perform as it was needed. That was a most important group, but it acted more as the counterpoint to the systems engineering designers because they were always interested in how are you going to test this when you got it built and what have you built in in order to make it testable, and so on.
So we had two groups, the reliability and quality and the test groups, who were working to validate the design concepts and make sure that they were reasonable in going into the field, and then the operations folks, of course, were worried how that design was going to be operationally useful, and so they came later in the process. Probably we never implemented that as well as we should have in terms of getting them involved in the early design process, partly because it took longer to set that office up and partly because of the nature of the operations folks, but they played a very key role, at Houston particularly under Chris Kraft, in influencing the design. The headquarters group had much less influence in this instance than the field group, and in particular Chris was working both with Marshall and with the Cape, as well as with Houston, to assure the operations, and when I say Chris, I mean not just Chris but the astronaut office. We had the astronauts involved early on in all of the design decisions.
COLLINS: Was this in the same manner in which pilots in aircraft can make sort of seat-of-the-pants judgments about what will work if they're going to have to operate a craft? Was it similar to that approach?
MUELLER: You know, in that sense it's similar, but it was different hopefully because they had a whole host of other people standing around explaining how they could do it, so it was again an interchange and a set of, to some extent, compromises, but a coordinated approach is what I'd call it, rather than an independent approach. It's not just seat-of-the-pants. They had to have some reason for what they wanted to do and be able to defend it. Of course, the astronauts were also all engineers at that time.
COLLINS: I guess another way to phrase that would be, there wasn't a lot of experience at being an astronaut. There wasn't a body of knowledge there.
MUELLER: That's right.
COLLINS: That you could rely on to say this probably works better when you're up there in space than something else. So what would be the manner of their input then?
MUELLER: Well, of course we were drawing very heavily on both the Mercury and the Gemini experience, so those were the principal contributors in terms of what the impact of their feelings were on the design. And yes, it was intuition, but it was intuition on the part of people like Chuck Berry and some of the other flight medics and intuition on the part of the engineers as well as the astronauts because they were all in a different regime and doing their best to understand what the trade-offs were.
COLLINS: At these monthly review meetings, would the heads of, say, the center offices, the program control, systems engineering, etc. participate? Is that how these things would work?
MUELLER: We had each of the program offices with their five key people at each of these monthly meetings. Now that takes a lot of time. We went around to the various centers to hold our management council meetings, so everyone didn't have to travel every month. But, in fact, later on in the program, that was one of the contributions that Boeing TIE [Technical Integration & Evaluation] brought in, and that was a setup where we could hold telephone with chart transmitted electronically, for our meetings, so that everyone didn't have to come to the same meeting place. That worked after a fashion, but it never quite takes the place of getting together as a group. Teleconferencing is never a completely satisfactory substitute for actually being together as a group and looking at a problem jointly.
COLLINS: Was the management review mechanism your primary means of keeping abreast of things? The influx of data about the progress of the program must have been immense. I mean, you had a fairly elaborate structure for providing information. How were you able to keep on top of it all?
MUELLER: Essentially I spent about three-fourth's of my time traveling around to the various centers and the various contractors just seeing what was going on. Just being there and being able to walk through and see what the real progress was and getting to the problem areas personally gave me some feeling of confidence that I understood what was really going on. In so doing, of course, the center director involved and his key people also went and saw what the problems were, and that got the prime contractor's attention, and so the problems got worked on. Most of my time was devoted to identifying weak spots in the program and going there and making sure sufficient attention was being paid to them to get them solved. That's true of our program directors as well, and our five offices were the sensors that kept us abreast of where the problems or likely problems were going to be.
COLLINS: The five offices in headquarters.
MUELLER: And the corresponding groups at the centers and to some extent the prime contractors. We had independent sources of information about what was going on.
COLLINS: What was the relative value--I'm not sure that's the way to phrase it--of the reporting systems, like PERT [Program Evaluation & Review Technique] and other things you had to keep track of costs versus these personal tours that you did?
MUELLER: The only way you ever got PERT really implemented was to go around and ask the guy who was supposed to be doing it where he stood on his PERT program, and you could usually find that he had his own program in his desk drawer. PERT was the thing he was talking to you about, but whether it actually meshed with what was going on in the plant was in some instances coincidental. So we did a great deal of time in terms of educating people on the importance of reporting properly on schedules and so on. PERT only is as good as the people that are using it, and I've never been overboard on any reporting system. The important thing is for it to reflect what's going on in a way that the local people can work with.
So we actually did not insist on our particular scheduling approach within the prime contractors. They had to translate their systems, though, into something we could report back in a uniform basis, so we could understand what was going on in our system. But you know, a GANT chart is just as good as a PERT chart, and companies have cults or cultures that are immutable with respect to reporting systems. Most of them, however, are reasonably good. If they aren't good, then we went to work and insisted on a system, and while they were putting in a new system, they might as well put in ours since it would simplify the whole translation process.
COLLINS: I'm still kind of feeling my way around this question. If you had a well-working reporting system, would there be the same need to do the extensive travel and sort of personal oversight of the program?
MUELLER: Reporting systems don't do you any good at all unless you do something about them. And so if you have a well-working reporting system, it only tells you what the problems are. We had a fairly well-working reporting system. We knew what the problems were, in large measure, and so then we concentrated our resources on getting those problems solved. But reporting systems just don't do anything for you other than tell you what has happened, and the important thing is to forecast what's going to happen and fix it before it gets to be a major problem.
COLLINS: So the balance that one would like to work out in this kind of situation is having an effective reporting system so that you can devote your attention to problems and look ahead to the future.
COLLINS: Okay. Going back to the functional offices, during the period after you came in, there was a very strong move to switch all the contracts over to incentive contracts.
COLLINS: What role did, say, the reliability and quality people or the test people play in evaluating contractor performance to meet the particular incentive criteria?
MUELLER: One of the problems with incentive contracts is the incentive has to be simple enough so that you can measure it and precise enough so that it incentivises what it is you really want to have done. So in our case, what we really were seeking was cost control and schedule control, and those were what we incentivized on. Reliability and quality, test, were gates that people had to go through in order to achieve the schedule and achieve the costs, but they weren't part of the evaluation of the incentive. Clearly they affected the incentive program because if you couldn't get the thing approved, the component or the system approved, you didn't make your schedule, and you didn't make your costs.
COLLINS: So the performance criteria were not part of the contracts usually.
MUELLER: That's right, for a very good reason. That is, they tend to be somewhat subjective, and you want to have the reliability, the quality people, and the test people independent of the incentive pressures. You want them to be the gates or the hurdles that you have to get over in order to get home.
COLLINS: With this set of reporting systems that were set up, did the people in the centers feel like your staff at headquarters really knew what the problems were and were really sensitive to the kinds of pressures and things that they were grappling with? Was there a sense that you understood what was being communicated to you?
MUELLER: Oh, I think so. Put it another way, it took a long while to get the rapport going between headquarters and the centers and, I guess more particularly, not so much between the project offices in the centers and the program offices at headquarters but the mass of people there--the engineers and the support folks and the center directors and the administration there understanding what that rapport was and how well it was working. So it took maybe a year before we had the system working properly or perfectly, and I guess perfection is something else again.
COLLINS: Resuming after a brief pause. When we talked at our last interview, you indicated one of the problems you had to grapple with when you came in, initially, was to better define what was expected of the contractors on the various projects they were carrying out. I was a little unclear what the process involved was in better defining what the contractors ought to be doing. How did that work?
MUELLER: The first thing we had to do was decide what it is we wanted to do, and that required us to go back and review our basic program structure. That's when we decided on all-up testing when we eliminated the Nova rocket and when were structured the Saturn I program to be just a precursor to Saturn V and concentrated on the Saturn V vehicle and the Apollo. We made the decision on the Lunar Excursion Module [LEM], and so that whole series of decisions we firmed up at that time. They'd been floating around, but there hadn't been an agreed-to and well understood program plan. That was the first thing.
Then having accomplished that, we found that our contracts were a hodgepodge of different kinds of contracts, everywhere from fixed-price to purely cost type contracts, and it was apparent that we needed to get some rationality into the contract structure to reflect our new programs. Most of those contracts also badly underestimated the cost of the program simply because no one had clearly defined what needed to be done, on the one hand, and on the other hand, there's a great tendency on the part of contractors, particularly on cost type contracts, to bid them in at less than they expected to cost with the expectation that there'll be sufficient changes so that they will eventually survive at least. This was a time when it was possible to restructure contracts and to do it on a reasonably rational basis. By that I mean that it was politically feasible to restructure these contracts. Today to accomplish the same thing, it would probably be far more difficult because of the adversarial relationship that's grown up between government on the one hand and the contractors on the other hand.
And so we set out to restructure all of our contracts to an incentive type contract. The concept of incentive contracts had just come into being at that time. It was a new concept and one that we embraced because it gave us an opportunity both to restructure the contracts and get them on a more rational basis and also to provide the right set of incentives so that the contractor and the government interests coincided.
COLLINS: I guess the next point is, did NASA have enough sense of what it wanted? You've sort of described the clarification of priorities, but it still seems that a greater specification of the engineering details needed to be made to do these kinds of contracts.
MUELLER: Indeed, we did work over a period of six months to work out the specifications and the details of these contracts. But again, the incentives were basically on schedule and cost, and those are things that can be set to some extent while you're setting your specifications up so that you can agree on some overall figures and recognize, as always, that particularly in a program of that magnitude, there will be changes. So we did provide for changes in the contracts, and that's an important ingredient. You can't make a contract at the beginning when you aren't sure what kind of insulation you're going to use in the second stage that defines exactly what that's going to cost and how you're going to go about doing it. So you need to preserve flexibility as well.
COLLINS: How do you balance flexibility and the incentive elements of the contract?
MUELLER: With a lot of good engineers. That's essentially what you do. You've got to have a close working relationship and a really good set of engineers who can make the right sort of value judgments.
COLLINS: Engineers on both the NASA side and the contractor side.
MUELLER: Exactly, and if you don't have good engineers on one side, you aren't going to have them on the other is what it amounts to, so it's essential that the government have good engineers and that the contractors have good engineers. We were fortunate, of course, in being able to draw upon Bellcom and eventually on Boeing, which provided us with engineering resources, in addition to those that were quite strong at that time in the government.
COLLINS: Was the kind of engineering that needed to be accomplished in Apollo of such nature that most of these contracts could be changed over to incentive contracts? Were there any instances of research or development activity that simply weren't amenable to incentive contract?
MUELLER: Well, when you're thinking of research and development contracts not being amenable, I would argue, that's not really the case. If you've got pure research, there's no point in having or trying to have an incentive contract. So much of the work in advanced research and technology isn't particularly amenable to anything other than the kind of incentives that are based upon an evaluation, and those are very effective, but in the case of Apollo, we were doing a fair part of advanced development. At least the engines were all new. The stages had a different structure and so on. But yes, you can incentivize those contracts, and it isn't all that difficult to do, but you do have to recognize that you may have to change. But at least if you change, it's from a baseline of understanding. Most of our first six months or nine months were set in developing a baseline, both at the highest level and then at every level underneath that, so that we had some basis for making our incentive contract changes. And those changes took place over a period of a year.
COLLINS: In making the initial contract awards, NASA went through a process of evaluating companies' technical capability as well as their management capability. When you went to change these contracts over to incentive contracts, did those kinds of issues come up again, or was it simply a question of contract detail rather than a re-evaluation of what the company might be able to do?
MUELLER: No, it was not going back and deciding whether or not the company was going to be able to do the job. Most of the aerospace companies had sufficient resources, at least those involved in Apollo and in the shuttle, so that almost any of them could do the job, and the original selection process is dependent on who they promise to put on the work rather than whether or not they could do the job, in most instances. So that wasn't a real problem. Not part of the incentive contract structure but in part of our management structure, we did have a fair amount of influence on who was actually working on our programs.
COLLINS: How did that work?
MUELLER: We reasoned with the management of the prime contractors. Very important that as we got into the program further and ran into difficulties, we found that certain of the management structure in our prime contractors was not as strong as it should be, and we then worked with the senior management of the companies to get the right sort of talents on the program.
TAPE 1, SIDE 2
COLLINS: Most of the major contracts had been let by the time you came to NASA, but did the Lunar Excursion Module contract come after you were there?
MUELLER: Well, Grumman had been selected, but the Lunar Excursion Module came afterwards. No, I guess Grumman hadn't been selected. I guess that was the only major selection I ever played a part in, in terms of the original contract structure. My timing is off a little bit on that one, Martin, but most of the major ground support equipment was awarded after I was on board. But the major aerospace components, with the exception of the Lunar Excursion Module--and even there, the Lunar Module had been given to Grumman at the time before I got on board if my time sequence is correct. It hadn't been described quite properly, but they had been awarded the contract. And that was because while I was at TRW I was busily trying to sell them on an engine that TRW had developed for the lunar module, so that helps me keep my sequence of things straight.
COLLINS: Okay. I guess what I was getting at was whether there were any cases we might examine to look a little earlier into the process. We've been talking about--
COLLINS: Restructuring rather than being able to, at the initial stage of discussions with the corporation, make determinations about the corporate resources that would be devoted to a project. Are there any instances of that that we can talk about? You mentioned the ground support contracts. Would those be a good case to look at for a few minutes?
MUELLER: Not particularly, because for one thing my recollections aren't all that great, and two, in the ground support area they're more mundane, and you don't have the same kind of questions about capabilities that you would have otherwise. The only thing I recall vividly is the crawlers, and that was more because they were monstrous things, and there was some question about could you really do this, and so we did some feasibility studies before we awarded a contract to be sure that, in fact, one could build such huge vehicles and have them work properly.
COLLINS: This is for transporting the launch vehicle from the assembly building to the pad.
COLLINS: Let me phrase it another way. Do you have any observations on the procurement process and the source evaluation process that we might discuss?
MUELLER: Well, I think that Jim Webb was a master at the final evaluation process. He insisted on having Hugh Dryden, Bob Seamans, and himself sign off on every procurement, and in turn, they had to reach agreement as to what contractor would be awarded. To my knowledge, the recommendations from the Source Evaluation Board [SEB] were almost uniformly upheld by that trio, but on occasion the Source Evaluation Board really didn't do the job that needed to be done, and so they got sent back to do their homework, and that actually occurred at several levels. One, I usually reviewed our SEB's before they went to Seamans, and we would want to be sure that all of the factors had really been sufficiently considered before we went forward, and that introduced a more rigorous review, I think, than had been characteristic earlier in the process. Then again, Webb, Seamans, and Dryden might ask some questions that had not been thoroughly addressed and would then be evaluated. As I say, I don't know of any instances, however, where there was a reversal of the Source Evaluation Board's recommendations. Normally they didn't recommend a particular contractor as such but evaluated them and ranked them.
COLLINS: One document that I thought might be useful for us to take a look at is something I passed along to you. I don't know whether you've had a chance to review it.
MUELLER: Oh, Bill [William] Rieke's comparison. Eventually the DOD [U.S. Department of Defense] changed over to a system much closer to that of NASA's. This was back in 1965, wasn't it, that Bill did that?
COLLINS: Yes. We're referring to a memorandum from William Rieke, dated December 21, 1965, under the title, "Significant Differences Between NASA and DOD Project Planning Systems."
MUELLER: To some extent DOD and NASA were leap frogging each other because when I went in, we didn't have a very well structured proposal route, you know, from concept to final article, and we put in place this four-phase program. Bill Rieke came in from Lockheed originally as my deputy and then took over as head of administration in headquarters or head of industrial affairs or whatever they called it at that time. But in any event, we were trying to get a structure for our procurement cycle that would avoid some of the problems we were running into with going too rapidly from an idea to try to buy a finished article. So we tried to introduce a set of steps with milestones that you had to pass in order to make the next step.
At the same time, the DOD was beginning to introduce more structure into its procurement policies because [Robert] McNamara at that time was trying to get a handle on what was happening within the services, and so they introduced a three-base system. We introduced the four-phase system, and eventually the DOD went to a four-phase system because it just happens to be, if you look back there you'll see that the DOD structure, right there, that little graph shows that the DOD concept stage covers three steps or two and a half steps of NASA's procurement at that point in time, and so it blurred the early phases, which are probably the most important phases of one of these projects.
COLLINS: Does this phasing introduce an unwieldy amount of bureaucracy into proceeding with a project?
MUELLER: It didn't originally, but it sure does now because each one of those phases now has taken on its own life, and getting from one to the next requires an act of Congress almost, and that never happens. You know, any of these things in concept are perfectly good. Then it's the practice of them, and as time passes there are more and more requirements placed on each one of these phases to get through the next wicket on account of because somebody made a mistake once, and therefore you're going to plug that loophole, and the first thing you know those four phases used to take something like two years to complete, at least to the last one, now take five or six, and it's just simply because people get enamored of the phases. If I were going back to NASA today, I'd start all over. I'd change the whole system, at least the names, and look at a new way of doing business simply because you get too much structure and not enough function. It's all form and no function as time passes.
COLLINS: What were some of the reasons for the development of phased project planning? I think we've already alluded to some of them here, but it would be useful to describe them.
MUELLER: Oh, tremendous cost overruns, schedule, and buying into things that didn't make any sense when you finally got around to paying for them or what was coming out of them. So the phased project thing was to be sure that we carried out the proper set of evaluations at each stage along the way, development from a concept to a finished article, and that was more characteristic in the case of DOD initially than it was in NASA's case. But nevertheless within NASA we had a very wide range of projects going on at that time, and it was difficult for our management to really understand what the implications were of these various projects. First thing you know, they'd have a little project that they had approved over here that suddenly took the entire agency's budget if they were going to finish it. Then they had to cancel it. So it was an attempt to make sure you understood what it was you were getting yourself into before you finally got caught up in the final commitments.
Interestingly enough, as an aside, I was talking to some of the people at a recognition dinner for Si Ramo, and it's always amazed me that we were able to take and develop the Thor program in something like fifteen months from the time we first thought we needed an intermediate-range ballistic missile to the delivery of the first article. Today, you probably could never do it. It would be fifteen years, and you couldn't get there from here. It's like Midgetman. That's been on the books now for fifteen years, and it still isn't a program.
So the phased approach is a very useful one, properly applied, but if it's used to kill a program, you're better off killing it to start with and not going through this charade. If you really are interested in doing a program, then it's a good tool, but you can't let it drag on for fifteen years. You've got to decide. You've got to set a schedule and make clear decision points and hold to that schedule and be willing to cancel if that's what you finally decide to do, but don't let it drag on in Phase C for five years or ten years.
COLLINS: As originally conceived, was this to be only an internal engineering tool?
COLLINS: Or was it to be an engineering tool and something to be presented to Congress to sort of provide something understandable for them in terms of a project's progress?
MUELLER: Well, it started out as a management tool, not really an engineering tool but a management tool, and Congress immediately picked up on it because they could see that this was something that they could understand. So yes, it immediately became to some extent part of the political process of describing what programs were, where they stood in their development process, and so on.
COLLINS: One point of contrast here that strikes me as interesting is the statement that NASA has more decision points involving top management. Was this a function of your approach as well as Seaman's and Webb's approach to overseeing the program? How did that come about that top management took an active interest in approving each step of the process?
MUELLER: I've always felt that one ought to make reasoned decisions as you go along in the development of a project, and this phased approach was one way of being able to do that in a reasonably clear and understandable fashion. But Webb particularly and Seamans to a large degree also liked it because it gave them an opportunity to see where the whole organization was going, and what tradeoffs they had to make with respect to which programs they could undertake and which ones they couldn't. So you know, I think it's a very useful management tool properly used. The number of decision points here are more than is even true today in DOD, but they come at logical points. And the problem with having too few decision points is you can have gone too far, and then when you cut it off, you've got a real political problem, as well as an internal credibility problem. That's what DOD keeps running into. They wait too long to make decisions in many instances.
COLLINS: One aspect of this, I wonder whether you might comment on because I'm not sure of all of its implications, comes under the heading of competition and contractor selection. The document here indicates that NASA restricts detailed project definition contracts, phase B, to major prime contractors capable of total project implementation and does the same thing with phase C, whereas DOD at this time did not impose such a restriction.
MUELLER: From a practical point of view, DOD did too. Right now I'm sure they would no longer say those words because of the political implications of favoritism or whatever. But in practice on large projects, it takes a major contractor to be able to even bid. I mean as a practical matter, there's no selection per se. The DOD had the same practical limitations, and I don't think Bill [Rieke] was right in saying that there was a great difference between the two.
COLLINS: Okay. But from someone in your position at the time as an associate administrator responsible for a large program, what was the value in restricting the field this way? In other words, not allowing new entrants at phase B?
MUELLER: We didn't restrict it as such. Anybody who really insisted could bid. Even then we didn't have any firm restrictions. But we were quite willing to explain to people that unless they could do the job, they weren't going to win. I mean that's true.
COLLINS: I guess what I'm trying to work at is the assumption behind this kind of approach, and is that basically it, that the assumption is that only the people who have prepared detailed phase B proposals are really capable of doing the job.
MUELLER: Right. Oh, I see what you mean. I missed what you were saying there. Is this the one that says that only the phase B winners will be competitive for the phase C? Is that what you're saying?
COLLINS: Well, it says at both stages here. Yes, right, that's the second.
MUELLER: Actually the idea there was to have the broadest possible participation in phase A, to select from that group of contractors those who were most qualified to do phase B, and then the final competition was between the two or three phase B competitors for phase C, the actual implementation. So it was phase A where you really did your first selection process, and then you selected from that in phase B. That was the process we used. Now, that did not preclude somebody--in some instances they in fact insisted--on bidding against the winners in phase B from outside. But as a practical matter, I don't know of any case where they ever won that competition because the phase B people had the advantage of working with NASA for a year perhaps in defining exactly what was going to be in phase C. But this was not an attempt to limit the competition but to structure it in a way that you got the best people for this particular job and could select over a period of time. The phase A selection process was really the wicket that you had to get through if you were going to get down to phase C in the end.
COLLINS: Again this whole process is kind of a testament to the complexity and technological uncertainty of most of these projects.
MUELLER: Yes, and one of the real reasons for the phased approach or basic reason why we went to it was to avoid having somebody who hadn't worked on a program at all, didn't understand it, coming in in phase C competition and bidding something that they couldn't deliver on because they didn't understand. You can bid it at a low price and then you get into, well now, I bid half of what this other guy did. How can you possibly award it to him? But you didn't know what it was you were going to build, and therefore you were going to overrun by two or three or four times by the time you got finished, and that was the reason, one of the basic reasons for that structure.
COLLINS: Another element of the NASA approach as laid out here by Bill Rieke is the early involvement of the major prime contractors in the total project conception, from a very early period.
COLLINS: And that strikes me as an interesting angle in handling these projects.
MUELLER: Yes, and it's very useful if you can get it, and it works quite well actually. Your basic problem comes when you have to make an early selection, and it gets more complicated when those four phases last over ten years instead of two because by that time you've lost some momentum that you otherwise would have in carrying out a program. The original participants in phase A have retired or gone to other companies, and you've lost most of the reason for that particular approach.
COLLINS: Was there a semblance of this formal system in the earlier TRW management activity?
MUELLER: No, that was much more a concurrent process. Bennie [Schriever] and Si [Ramo] and Dean [Wooldridge] evolved concurrency, where you did things in parallel in order to get things done rapidly, and that's another alternative. This is one that says you're going to end up with a single contractor at the end. If you could afford to have parallel contracts so you have two contractors or three contractors working on the same program, then you can short-circuit a lot of these steps because you've got alternates then you can choose from at the end.
That was the technique Dave Packard used or advocated, for example, and was the one that the ballistic missile program pioneered. You had Atlas and Titan going in parallel. You had Thor as a third element. Then you brought in the Minuteman. These were all competitive programs for arriving at the same end product. That's a very efficient way of spending your money. Probably the most cost effective way I know of getting a program done. It's just that our Congress and budget folks don't really understand the problem well enough to recognize that.
COLLINS: Was this the case in the sixties as this phased project planning approach came into being? Because this is really pretty explicit, that you're going to go with one thing. This phased approach is directed towards making a single choice.
MUELLER: Right, whereas in the late fifties, there was the approach of parallel programs where you had several contractors producing an end product that had the same purpose, maybe different--Titan and Atlas were quite different approaches but, nevertheless, delivered warheads quite effectively. And so you had a choice at the end as to which one you actually produced in any quantity. That was what happened with the F-l6 and F-l7 programs where Dave Packard had two contractors producing the same aircraft, and when they got to the final procurement, they selected the one that they wanted and liked the best.
COLLINS: Did NASA entertain the notion of concurrent approaches for any of its activities?
MUELLER: We did, for example, in the Saturn program. We had the two approaches but never carried it through to an end point. We never had the conviction that we could do that. We were dealing with very large sums of money at that point in time and trying to build both the Nova and the Saturn, which was one of the original approaches. We just couldn't see our way clear to getting enough funds to do it.
COLLINS: So the idea of a single approach and doing it in a phased manner was in part a recognition of the political realities.
MUELLER: Right. Unfortunately, yes. If we had been able to, we would have carried out both lunar-orbit rendezvous and earth-orbit rendezvous and carried out parallel approaches to that same, which would have ended up in the same way with the lunar landing, but we would have had alternatives with which to do that.
COLLINS: I think we may have mentioned this in the previous discussion, but there were instances for components of these larger systems that you were able to do concurrent production.
MUELLER: Yes, and in many instances we had concurrency particularly in the electronics arena, where our costs were not so tremendous.
COLLINS: Let's change gears a little bit. I have never had a good clear sense of the role that Bellcomm played in the NASA organization. I wonder whether we might discuss that.
MUELLER: Yes. Bellcomm was in place when I arrived there and had a somewhat ambiguous role, partly because people didn't really understand system engineering as such and partly because they were an outsider, brought in primarily by Jim Webb to provide some technical capability that he perceived was not present in NASA. Maybe it was a combination of Bob Seamans and Jim Webb. In any event, when I got there, they were sort of floating in what they were supposed to do. They were trying to define a role, and various centers and other parts of headquarters were trying to define a role.
So I elected to make them our systems engineers in that box at headquarters and give them an in-line role because at that time they were simply a staff function that didn't have any real responsibility. That gave them a real responsibility, but it also created a little bit of difficulty in the center understandings because they were not used to having a contractor working on the overall system design. But we finally got that concept understood, and it worked very well.
Now the one thing that Bellcomm did do, which they did superlatively, was to manage to cause system engineering to get done. Not necessarily to do it, although they did a great deal, but they caused the centers to develop their own system engineering talent and to be sure that it was getting done and the proper set of trade-offs made and the proper understanding of what the design consequences were. Bellcomm was very useful in terms of providing some very bright people who could go out to the centers and catalyze that kind of an understanding, so they also played a training role, if you will. They also served as our principal interface with the scientific community, and they were well regarded amongst the scientific groups, and they were probably as competent as anybody in the rest of NASA in the area of space sciences. So they played a liaison role with the other organizations, OART and OSSA, during that period so that what we were doing in Apollo and Gemini could support the activities in the other organizations and vice versa.
COLLINS: So Bellcomm had this liaison role within NASA, not to the external scientific community.
MUELLER: That's right, although they also played a role with the external scientific community as representatives of our Apollo program. They were our engineering arm at headquarters and, therefore, were interfaced naturally with the National Academy of Sciences and so on.
COLLINS: So originally Bellcomm was staffed to what part of the organization?
MUELLER: To Brainerd Holmes. They were staffed to me, too, so they had an element which was systems engineering in the Apollo program, but they also had an element of supporting the advanced planning operation that we had. They were a key player in the development of our long-range plan, the one that we put in place before I left which included the space shuttle, the space station, lunar colony, and the Mars colony.
COLLINS: I think that's a whole separate discussion that we probably can save.
MUELLER: That was a part of the Bellcomm role, was supporting that activity. They also, for example, did the--I mentioned earlier--geodesic study. They set up the first set of complete specifications of the environment we were going to operate in for the program, that is describing the lunar surface, the set of trajectories we would use, and so forth. A baseline set of specifications for the environment that the program had to meet, live with.
COLLINS: That seems like a function that the centers could have undertaken.
MUELLER: Well, and of course they participated in it. But if you're thinking of what the lunar surface is like, we didn't have any great set of knowledge in that case at that point in time, and so sure, everybody participated, but Bellcomm took the lead and actually wrote the documents.
COLLINS: You mentioned they had a role in inculcating a systems engineering approach in the organization. How did they work with your systems engineering functional offices in headquarters?
MUELLER: Well, you recognize they were our systems engineering functional office in headquarters.
COLLINS: Okay, that's where they come in as a line organization.
MUELLER: Right. So, they were it, and then their counterparts in the field they worked with in making sure that they were doing the systems engineering that needed to be done at that level.
COLLINS: Okay. That clarifies that substantially. I guess it would be a good point to explore your relationship and the manner in which you reported your activities to Seamans and Dryden and Webb. How did that relationship work? What were the mechanisms for you to keep them informed of major events?
MUELLER: Well, let's see. There was the formal set of reporting, which was a monthly review that they held for all of the program offices. Then there's the informal reporting, which involved any time there was a problem, calling them, either Seamans or Webb--depending on the kind of problems--or both, and keeping them informed on what was going on in the program, where they were particularly sensitive to anything that would hit the newspapers, and so we all developed a sensitivity to that. Usually we were ahead of the newspapers but not always.
Then there were special meetings for things that needed to be worked out, particularly between program offices. Things like in Skylab, the question of the telescope, who was going to build it, the X-ray telescope, were things that became somewhat sensitive. I remember having a series of discussions with Homer Newell and eventually with Bob Seamans where the question was finally decided as to which center was going to build this telescope and who was going to be responsible for it, and that worked out reasonably well. Somehow or another Homer never felt it worked as well as it should. But then that's life, I guess.
COLLINS: Well, that was always a question in balancing out the manned program, of how the science that might be done on the manned craft would be coordinated. What were your feelings about how that interaction should work?
MUELLER: I think if you think about it, one of the basic problems was that the centers, particularly Houston, was an engineering center and regarded science as being an unnecessary burden that had to be put up with but reluctantly, and so there was that end of the spectrum. Then there were the space scientists, who felt that anything spent on the manned program was a great waste of money and should have been devoted really to doing the fundamental space research that needed to be done. I tended to be a supporter of the scientific work and really forced, particularly on Apollo, the science into a prominent position in carrying out that work. There wasn't much point in landing on the moon unless you could do a reasonable scientific investigation in the process. In that sense I was supporting Homer to the maximum extent that I could.
TAPE 2, SIDE 1
COLLINS: We've jumped around in our discussion somewhat today, and I think we might as well take one more jump and then tie up some of our earlier points of discussion. We've examined some of the mechanics of the government-industry relationship during your period at TRW and later at NASA. I thought I'd ask you to make some summarizing comments about what stands out as the characteristics of a good working relationship.
MUELLER: Well, as we were discussing at lunch, there's a cross-section of possible working relationships, and if you describe a good working relationship, it is clearly one where the interests of the parties concerned coincide, and that is a general statement. More particularly, if you can bring to bear the effective competition for producing an end result, no matter what it is, then you can certainly reach a better working of the competitive process, which is fundamental to all human endeavors when all is said and done. In the case of government and industry, one of the first things that leads to good working relations or to greater productivity, I'll call it, or greater efficiency is to define clearly the role that each is going to play in the development of the end product and to make sure that the incentives, whatever they may be, are the same for the contractor and for the government personnel.
I think to make that more specific, if you look at Ramo-Wooldridge and its relationship with the associate contractors and with the Air Force, you have to think of the Air Force as being the government, aided by Ramo-Wooldridge as a support arm, and the prime contractor teams as being industry. So you have a spectrum going from the procurement officers in the government through the systems engineering people in Ramo-Wooldridge and the prime contractors who were actually producing the missiles. Now in this instance, because of the availability of resources, the Air Force was able to fund competitive approaches to intercontinental ballistic missiles. In fact, initially there were some four or five contractors who did some preliminary studies, and these were then sorted down to the Atlas, Titan, and eventually the Minuteman, and somewhere in the middle, the development of the Thor intermediate distance vehicle. That Thor was in competition with the Army's Redstone Arsenal, whatever it was. What did they call it anyhow?
COLLINS: Are you talking about the Redstone?
MUELLER: No, I'm talking about the Jupiter.
COLLINS: Oh, the Jupiter, right.
MUELLER: And so there you had a case of parallel programs or concurrent programs where a lot of the management had to do with essentially getting the resources and balancing the exchange of information and not so much directing the contractors because they were highly motivated to succeed in competition with the other prime contractors, in the one case the Jupiter versus Thor, in the other case the Atlas versus Titan. So that's one mechanism for getting the management structure in place, and of course, the Air Force was able to provide a common objective. It was get those ballistic missiles out as quickly as possible, and that was something that industry could understand, and together they managed to accomplish that. Not only were there parallel approaches in the vehicles but also in the guidance systems. There were four different guidance systems developed in parallel, and there were parallel developments of engines for these vehicles, both solid and liquid, so there's a case where one can accomplish a miracle, if you will, by simply making available the necessary resources and standing out of the way and allowing the competitive forces to take charge.
But in no case was there an adversarial relationship built up between the prime contractors, the associate contractors, the Air Force or the systems engineering contractor, and one of the keys to that was the effective communication between the top management of the Air Force and the top management of Ramo-Wooldridge and the top management of the prime contractors, a technique that I borrowed when I went back to NASA. I think that we were able to achieve in a different way a very good set of working relationships between the government and the prime contractors on Apollo, and that's also true on Gemini. We had begun to develop an adversarial relationship when I got into NASA because all of the contracts were overrun, and there was a tendency on the part of our NASA procurement people to say, well, that's the contractor's fault. In a situation like that it's clearly a joint problem that needs to be attacked jointly, first of all by defining our program clearly, then by renegotiating contracts and by establishing the proper set of communications between the prime contractor's top management and our management within NASA. We were able to build that credibility in each other that made it possible to do many of the things that were accomplished in the Apollo program without its being one of those things where it's "who struck John" and that sort of a negative relationship.
So I can't emphasize more clearly the need for a clearly defined program, and I think the use of incentive contracts is a great aid in being able to establish the proper communications between the contractor and the government, and the need to be sure that there's adequate understanding on the part of the management, both in the government and in the contractor, of what really the aim is and what's expected of the contractor, and what in turn the government has to do to support the contractor in carrying out this thing. The present tendency to avoid that communication, that really is what is causing a problem in many of the contracts. The inability to have a frank and open discussion about problems because of the conflict-of-interest laws that have been built up and how they're interpreted within the government hierarchy makes it very, very difficult to have an effective understanding of the other's point of view and what needs to be accomplished.
COLLINS: I guess I'm not clear on how conflict-of-interest laws apply when you've got a program person in NASA or the Air Force sitting down and talking to his counterpart in the company.
MUELLER: Well, the way it applies is that there's an implication of impropriety if, for example, you go out to dinner with a contractor. There's a conflict of interest there. Or if you ride in his airplane or he rides in your airplane, there's a conflict of interest. Yet it's in informal exchanges that you gain most of the credibility and the understanding of each other's points of view that are so essential to carrying out a contract effectively. If you're like in Russia, where you have a KGB agent on one side of you while you're talking to your counterpart over in the U.S., why, it does inhibit conversation or the exchange of information of any sort. Essentially the government has built itself into that position, where you have a reporter on one side and a lawyer on the other side while you're doing your discussions on the contract, on what the problem is that he's facing out there. It doesn't lead to very much progress or very real exchange of information.
COLLINS: So in a sense what you're suggesting is there needs to be some sense of confidentiality and trust.
MUELLER: Yes. You know, the present view is, as near as I can tell, that no person in government can be trusted to have a conversation with a contractor without his taking advantage of it, being bribed or subverted or something, and that's ridiculous. Ninety-nine point nine percent of the people in government are honest, straightforward, and doing their darnedest to do something useful. And in fact, the whole era of distrust that's been built up is conducive to the kind of subversive--not subversive, but the kind of accusations that float around government all the time. If you're seen going out with a contractor, why, you immediately are suspicious that some idle information is being transferred over to that contractor that otherwise he wouldn't do, or some other foolish thing like that.
COLLINS: There are a lot of outside pressures that can affect this relationship between government and industry. What is the role of corporate lobbyists in Congress, where in a sense it seems like they're trying to take a path outside of the relationship, say between the agency that's letting the contract or overseeing a contract, being perhaps an end run around that relationship to get what they want? I mean that's become a very prominent part of the Washington scene.
MUELLER: Always has been. If you remember back in the Apollo days, who was this guy in the Senate who was selling privileges for contractors? I think that lobbyists are highly overrated as to their effectiveness. There are exceptions, but primarily their usefulness lies in the gathering of information, trying to forecast what Congress is likely to do, and providing information to the corporate management, and for that matter to the government managers, what the needs of various Congressmen are. You know, as a practical matter they have certain requirements that they have to meet as well, and it is understanding those and making sure that you accommodate to the extent possible those needs that lets you get through Congress. The lobbyists that claim, or the purveyors of influence that claim they can influence the selection of a contract, usually offer their services to all of the participants, and whoever wins gets to pay them, and they have almost zero effect on the end result.
Now there's a new phenomenon, though, that's come into being, not particularly new but new within NASA, where the NASA center directors have been using their lobbying power to prevent changes in their situation. That's a new phenomenon and one that is very dangerous because center directors who lobby for their own particular center's interests will end up causing more problems, not only for NASA but for themselves, than they can imagine as time passes. But that's a symptom of the times, that people think they can influence the decision-making process, and in fact, that exacerbates the problem, and about the only thing it does is reduce the credibility of the agency as a whole in the eyes of not only Congress but the media and elsewhere.
That's been a continuing problem at NASA in recent years, is the lack of credibility in their relationships with the Congress and with the press. You know, government works to an amazing degree on people's belief in other people and trust in the judgments of the people. And you can understand why because in the case of Congress, they're dealing with subjects for which they have relatively little background, and they're making decisions, very far-reaching decisions, based more on their faith in the individuals that are presenting the problems than in any real understanding of the problem itself, and that's also true everywhere else in the government. When all is said and done, it is much more a matter of people-to-people relationships than it is in solid facts that can be defended on a factual basis. Not that you don't need the facts, but the decisions are made on things other than long-range plans or logical bases.
COLLINS: This is kind of a value question. I don't know how well you can respond to it. But is it your sense that by and large, corporations in the aerospace industry see it in their best interest to proceed along the approach that you outlined as the more productive or efficient one?
MUELLER: You mean working closely with government?
MUELLER: I think they do, but they're hamstrung by the set of laws that Congress has enacted. You know, let me just take a case that's just been resolved, Jim Beggs, for example. Here he was in government and doing his best, and then he was indicted on a series of charges for some actions that took place at General Dynamics some years before. Then those charges were dropped when it was really discovered that the government had not only known about them but had approved the very actions that he was being indicted for. Now that kind of thing has created a credibility gap between industry and government so that, you know, people think three or four times before they begin to communicate with their government counterparts, and government, on the other hand, thinks seven times before it talks to anybody in industry about anything of any substance whatsoever. So until you change the attitude that all people in industry are out to screw the government, and the government's attitude and the industry's attitude that, boy, if we do anything with these people in government, talk to them or otherwise, why, we'll suddenly find ourselves in court being indicted over improper influence.
COLLINS: But how do you handle these questions where there are genuine screw-ups that involve substantial amounts of money? How do you handle those?
MUELLER: Treat them on an individual basis and don't make a general law that proscribes everybody from doing anything constructive. You know. That whole body of laws came about because of one or two examples, and so therefore, they're going to cure the whole problem. And boy, when you cure the whole problem, you create more problems than you ever thought of because for every one problem you've created now a thousand new problems. Lack of communications. It's amazing how important a real understanding is in carrying out any contract of any size and how hard it is to make that real understanding, that communication, continue over a period of time, and it doesn't take much of any exterior force to really make that very difficult indeed.
COLLINS: What makes it difficult to maintain that?
MUELLER: Lack of trust in the other person. Afraid that if you tell him that, well, you know that he's got a problem, and it's going to cost the government a lot of money, that he will report that to somebody else who will report it to somebody else and somebody will come back and say, gee, you're unfairly helping this guy, or something like that, when really what you're trying to do is get the job done as expeditiously and as within a budget as you can. You have to ask yourself, now, what do you do about all that, and one thing you can do is take the SR-7l program. You can get yourself out of that whole complex of bureaucratic stuff and go into a black program, and of course, that's how the ballistic missile program was run. It was run completely under wraps. It wasn't known what we were building or what we were trying to do, and very highly classified material, and that's about the only way you can get something done today in the government, is to start something new. Shut NASA down and start a new space program and do it totally under cover, bury it. Then you could get something done.
COLLINS: In the Apollo situation, obviously you had a tremendous amount of public scrutiny.
MUELLER: Oh my, did we!
COLLINS: Is it your sense that that inhibited your ability to build an effective relationship with industry in that situation?
MUELLER: No, not in that case because at least everyone knew we were striving for what at that time was an almost impossible goal, namely getting men to the moon by the end of the decade, so with all the public scrutiny there wasn't this business of, gee, did this guy take unfair advantage? Was there some information given to this contractor that would give him an unfair advantage? Or whatever. None of that came down. To be sure, at the time of the fire, we had a fair amount of scrutiny by all parts of the media and the government, but we finally were able to describe in some detail what really happened and why it happened and convince people, at least the key people, that it was one of those things that was the result of a basic design decision that was wrong, and that we'd fixed that, and therefore we could go forward with a good conscience and with real understanding. It took us a year, but that was probably necessary in order to get the system rebuilt properly.
COLLINS: In this period of the Apollo program, you talked about earlier the need for continual communication as a basis for maintaining the trust, and that that trust, even under those circumstances, was fragile. I guess I want to get a sense of why, even in the best of circumstances, that connection is difficult to maintain.
MUELLER: I don't think it was fragile, but it was difficult to maintain because people, when they get into a problem arena, would rather solve the problem themselves than allow anybody else to help them, and that's very understandable. Yet in a program like Apollo we had to have visibility into the problems so that we could marshall the resources to solve them before they got to be showstoppers, and that's an inherent conflict at the basic human nature of people, and so you have to keep those lines of communication going freely. That was one of the principal reasons for these five functional boxes, was to get more lines of communication so there wasn't a possibility of a single line breaking down for some reason, whatever, personality or what. When you're dealing with 200,000 people, personalities can block communications very easily, and so you need to be able to work around them and make sure that people realize that communications are going to happen, so they go out and communicate quickly rather than slowly.
COLLINS: It would seem also that incentive contracts encouraged the contractor to bring problems to the agency's attention because obviously quick resolution means being able to meet your incentive criteria.
MUELLER: Exactly. Those do help no end and particularly the schedule because once you have an incentive on schedule people get real interested in making sure they meet their target dates.
COLLINS: Do you have any more comments on the industry-government relationship?
MUELLER: No. I think that they're key, however, to success. I would also say that, you know, that phased procurement we talked about, when it is overused or stretched out, is counterproductive rather than productive. There have been more programs in recent years that have been studied to death and literally have gone nowhere. Although there's been a substantial amount of money put into them, nothing has been produced because of this restudy and restudy and restudy. I guess a good example is the space station. There's a strong move on the part of parts of Congress to restudy what the space station should be. Goodness knows that space station has been studied now for twenty years or more, and there's hardly anything that you can ask about it that hasn't been studied at one time or another in some depth.
COLLINS: When we talked earlier about phased project planning, we talked about it from the government point of view. What benefits or detriments does industry see in this approach? Would they rather see the old primitive style system where a concept is simply put forward and then they run off and try to make the hardware? Or do they prefer this more rigorous approach to the production of a piece of hardware?
MUELLER: Oh, I think industry would much prefer a clear decision. You know, if you want the optimum way from an industrial point of view to conduct one of these programs, it is you set up two or three parallel programs, describe what your end objective is, don't describe how you're going to get there from here, and let a competitive contract to two or more companies to go reach that end goal. That is the least expensive way of carrying out a program. From industry's point of view this phased procurement results in first of all, you're spending a tremendous amount of money on phase A to win one of the three or four or five contracts. Then if you win, you have the opportunity of spending a lot of your own money on phase B because the government always underfunds phase B, so that you can have a chance to win phase C. Of course, if you win phase C, you usually then can recoup all of that money you spent before, but if you lose phase C, you've just lost all that money you spent on getting ready to compete for it. So from industry's point of view, that process is very costly, and since there's only going to be one finalist anyhow, it's roulette with a lower chance of return than playing roulette in Las Vegas.
So it's convenient from the government's point of view. It gets a lot of free work done. From industry's point of view, it's very expensive and not very rewarding, and from the standpoint of the nation it probably in the long run costs more than would a parallel contract for the end product given to two companies to compete. I think the F-l6, F-l7 competition resulted in two very capable airplanes that cost the government a fraction of what they would have cost under a phased procurement approach leading to a single entity, and a fraction of the time.
COLLINS: Has anyone to your knowledge ever done a comparative study of the two approaches to examine this very question?
MUELLER: I think that there were some economic studies made of that, but there's never been a truly comparable--you know, there are always differences, so I don't think there's been a reasonably good study made.
COLLINS: I'd like to return to looking at some of the organizational questions. We talked earlier about the role that Bellcomm played, and I wanted to ask a similar question about the role of General Electric [GE], and what that contract was specifically for, what it did, and where it fit into your organization.
MUELLER: Let's see. The General Electric contract is an interesting one. While I was at Ramo-Wooldridge and after the initial set of contracts had been let, we were on the team. We were the systems engineering element of the team that General Electric put together to bid for the Apollo spacecraft. Looking at what was going on, I came up with the idea that we should propose to NASA the concept of an integration contractor, somebody to take all of these various components that were being procured by the various centers--the Saturn vehicles, the spacecraft, the guidance system--these were all pieces that were put together, and somewhere there needed to be somebody who was going to pull all this together.
I proposed, we put together an unsolicited proposal and brought it in, and the most sympathetic place that I could find, and probably the place that needed it as much as anybody, was Marshall. So we pitched that to Marshall, and Wernher von Braun was enthusiastic about the concept, so we took our proposal back to Washington and proposed it to the management council and really to Webb and Seamans, and the story I got was Webb and Seamans thought that was a good idea and sounded great, and then they said, "Well, but we're going to give it to General Electric," and so GE got this integration contract.
Now the only small problem was that GE had very little experience at integrating large systems, and so when they started out to do it, they put together a team and went around to each and sent that group to the centers and said, "Now we're the integration contractors. What do we do?" That created a small amount of difficulty, and the first thing the centers did was to isolate them, get them away from whatever it was that they were doing because they didn't want these inept folks blundering around in their business. GE really never recovered from that. They went at it in a fairly arrogant sort of way, and the centers rather promptly put them in their place, and Marshall was not very sympathetic because they'd recommended us, and the powers that be had chosen someone else, and so they didn't do too well at Marshall for a while.
COLLINS: When you came to NASA then, how did you utilize this resource?
MUELLER: We tried to find a use for them, and we eventually got them involved in the test program as being something where they could be useful in terms of providing overall test planning and so on, and that was, I think, the principal use that was made ofthem.
COLLINS: How was this role then of an overall systems integrator fulfilled?
MUELLER: In the case of Apollo?
MUELLER: Of course, we had the Bellcomm doing systems engineering. We had our own program. The program office really was doing that overall integration of systems. Now when we brought Boeing in, they helped very considerably, particularly in following up on the paperwork that needed to be done, on making sure that we had in fact done all of the testing and keeping a record of all the tests, making sure that all of the errors found had been closed out. There's a tremendous amount of just details that need to be followed in a program like that and need to be followed across the board because there are a lot of interfaces that need to be sure that the open items are closed on and so on. Boeing did a tremendous job in doing that, in making sure it was getting done, not doing it so much but just providing the mechanisms for accomplishing that. They took up where Bellcomm and GE left off and provided an overview of that.
COLLINS: I'm still a little unclear on how the systems integrations responsibility was carried out.
MUELLER: It was done through a series of task groups, actually, when you think about it. We clearly set up interface groups, and that's where you really are carrying out your system integration, is in the definition of the interfaces, and making sure that what's happening on both sides meet the end objectives that you've set out for yourself. These are made up of personnel from systems engineering. In many instances we had program control there as well, but primarily systems engineering groups both at headquarters and the centers, and the design engineers from the various disciplines, and of course, the contractors. So these interface groups then were responsible for the definition of and change control on those interfaces, so they met regularly to examine what changes were required in the interface in order to continue.
TAPE 2, SIDE 1
COLLINS: What role did these systems integration task groups perform when you got down to the physical assembly of the craft and the launch vehicle?
MUELLER: Now you're thinking about the interface groups, control groups, and they really had very little to do down at the Cape until and unless there was a discrepancy found when they were putting them together or if there was some problem that surfaced once you had the assembly that required changes in the design or to go back through and do some quick fixing, because the groups at the Cape consisted of people from the prime contractor, who went with the equipment down there, plus the Cape people who were doing the actual testing. The other interface groups had done their work primarily, so it was just a matter of assembly and completing the final testing. The integration really took place far earlier with the establishment of the interfaces. That's more complex than the mechanical interfaces because you also had to define all of the signals that went back and forth, all of the cabling, and also the transfer loads that were going through, and try to identify, and that wasn't just from stage to stage but also between all of the components, like the engine interface and so on had to be defined since those in turn were controlled in order to get the various contractors to have the right interfaces when they came together.
COLLINS: Would the contractors meet independently to work out interface problems? Was that part of the mechanism?
MUELLER: Not usually. Usually it was set up by NASA, and the contractors met with the cognizant NASA people to work it out since all of these interfaces have implications beyond just that interface.
COLLINS: So just to return to the initial question about General Electric, they played no role in this larger activity for which they had been proposed initially but took on the different role of testing, checkout, that sort of thing.
MUELLER: Yes. It varied from center to center because in some instances they were used for part of the support for the interface descriptions and so on. In other centers they weren't. They essentially became support contractors for the centers and ceased to have an overall integration role.
COLLINS: Just before we stopped for lunch, we began to explore some questions relating to the relationship between the Office of Manned Spaceflight [OMSF] and other elements of the agency. What was the relationship between the interaction between the OMSF office and this Office of Industry Affairs?
MUELLER: You mean Bill Rieke's office?
COLLINS: Yes. On the organizational chart it's shown as a staff element of the associate administrator.
COLLINS: Separate from OMSF. So I wondered what the relationshipwas between them?
MUELLER: Bill came in as my deputy, and we began to get involved in union problems. We had, I don't know, ten or twelve contractors working down at the Cape, some union, some non-union, some quasi-union. With a lot of different unions involved, you had the potential of stopping all work on a continuing basis. So Bill took on the job of working that interface, and then it looked like it would be a better role at headquarters, and also we were trying to build some further relationships in the whole industrial community. So Jim asked him to come up and head up that Office of Industrial Affairs and he did and he was very effective at it. His background was one that particularly fitted him for that role.
COLLINS: Yes. How did his role complement the working relationship between your various levels of the project offices?
MUELLER: In a sense he didn't really have anything to do with our program office structure.
COLLINS: Well, in the sense that each of these levels had connections with their counterparts in industry, did he play any role in those relationships?
MUELLER: Not overtly. When we had a problem, why, he could help, but his main function was external to the program, getting the external interfaces, keeping the union bosses from calling a strike, keeping DOD informed of what we were doing, and keeping them from doing something that would be harmful to our progress, and so on. So his role was primarily for the outside world and keeping things going smoothly there. But as far as the programmatic aspects go, in that role he had almost nothing to do with it.
COLLINS: Resuming after a brief pause. We were talking about the connection between the Office of Industry Affairs and the OMSF activities.
MUELLER: Yes, and we had finished that.
COLLINS: I guess. There were various staff offices associated with the administrative secretariat and the associate administrator that set agency-wide guidelines or policies for the kinds of data that ought to be reported and that sort of thing. To what extent did these guidelines help or hinder what you needed to do in the Manned Space Flight Program or alternatively, how did you work with them to make sure that the reporting structure set up fit in with the way you organized your activities?
MUELLER: Most of that was set up as a result of interaction between our administration, primarily Bill Lilly, and our program--my staff level of program control, and the headquarters --Bob Seamans' man. That working relationship was relatively good, and so almost all of the reporting was developed jointly, or in many instances, we originated it and they adopted it is about what it amounted to. I think in that regard we were taking the lead in the reporting procedures, trying to develop an effective reporting arrangement within the agency as a whole because it was essential for Webb and Seamans and Dryden to know what was going on and understand what the implications were.
COLLINS: Did their information needs require you to institute additional channels of information or did the channels that you had already established serve the needs?
MUELLER: They had a subset of our reporting is what it amounted to. You know, we had more information than they could use, so we simply selected out of that, that which would be useful for them. On the other hand, the other two associate administrators had a more difficult problem because for one thing, they had a much larger diversity of small projects to report on, and so they had to develop some schemes for getting meaningful reporting on those projects. But no, that was cooperative and quite constructive, I think. Now, there's another area that you didn't mention, which is the Congressional liaison arena.
COLLINS: That was my next question.
MUELLER: And the external media question are two other areas where we worked with the administrator's office. We had in general very good working relationships with the Congressional liaison groups, and we had our own small group that supported them, and they supported us and worked quite effectively, I believe. Congressional committees at that time were divided essentially like our three associate administrators were. There's a science guy, a manned spaceflight guy, and a technology guy. So the Congress mirror-imaged our organizational setup is what it amounted to, and when we got reorganized, they reorganized, under Tiger Teague to match us, and that was on the Congressional side. On the Senate side they essentially worked with Jim Webb in large measure, and he also worked with his Congressional liaison with the appropriations committees, whereas I primarily worked with the authorization committees. But that work was again a joint effort, and we tried very hard to keep them informed about all of our contacts up there and, in turn, used their resources whenever we needed them. By the end of our time there, we were actually the chief contact with the authorization committees in the House and to some extent in the Senate as well. Because we were the principal spenders of money, we had the largest burden of convincing people we were doing it well and creditably.
In the media relations area, we had a less cooperative set of affairs. Julian Scheer was far more prone to maintain his personal control over the external relations, and that was more difficult over time than was the Congressional arena, and partly because dealing with the media is always difficult and partly because Julian managed to get himself at cross-purposes with almost everybody in the media at one time or another. So you had a buildup of an adversarial relationship between NASA and the media, and that has continued more or less to this day. We've never quite--maybe it isn't possible--built the kind of cooperative relationship with the media that would be most productive. Not that I think you can ever influence the press or the radio or TV to not find something at fault but really so that you can at least have an opportunity to correct the inaccuracies and fix them up before they get to be published.
COLLINS: On dealing with Congress, this was a relatively new experience for you when you came?
MUELLER: Oh, yes. Yes, indeed. I remember vividly my first occasion, and that was just after I got there. Walt Williams had written a document describing in some detail the horrors of the Mercury program and published it, and it had just gotten to Congress, and Tiger Teague called me up and said, "Would you come up here and explain to the committee what was really wrong with Mercury and what this saga of the errors and commissions means?" So I went up there along with Walt and said, gee, this was just an internal document that we'd written to make sure we didn't repeat any errors in the future, and of course, we didn't write about all of the good things that happened and all of the things that were right. All we wrote about were the things that were wrong, and besides we didn't ever mean it to ever get out anywhere anyhow. That was my first experience with Congress and Walt Williams.
COLLINS: How did Congress greet that argument? Were they sympathetic to it?
MUELLER: Oh, they were not very sympathetic, but after appropriate discussion of the shortcomings of management of NASA and so on, why they went on to other things.
COLLINS: How did you learn to work effectively with Congress? How did you go about educating yourself in how to work with the Congress?
MUELLER: I just spent some time talking to the various people up there and finding out what their problems were and early on decided that the one thing they needed was to have as much information as they could before a newspaper got it. So we set up something that I think was fairly innovative and that was we invited the authorization committee members down every month to have a review of our program--what we'd done that last month and what our problems were and what our prospects were.
COLLINS: This would be at headquarters or the centers?
MUELLER: It would be at headquarters, primarily. Tiger did a constructive thing and that was he invited his committee to go with him once a year or twice a year before the authorization hearings, to go around to one, two, three centers and contractors and just see what was going on, and we provided airplanes and things for this. That was constructive because then they got out to see the hardware and began to understand something of the complexity involved in bringing this thing through. From the first more or less adversarial relationship that we had when I went in, to the end of the program, we were really working very well together. Again, it's a business of building credibility and spending enough time and energy to establish real communications, and that just takes time and cultivation.
COLLINS: Were you encouraged by Mr. Webb and Bob Seamans to have this direct and frequent contact with Congress?
MUELLER: The original view of Jim was that I ought to stick to my last and run the program, and he would take care of the Congressional relationships. But after the first series of hearings on the budget authorization, why, we did a fair amount of PR in the process, capitalized on really having a defined program and defining this clearly enough so that Congress could understand where we were and what we were going to do. We gradually worked our way around to gaining Jim's confidence that we could deal with Congress, and we carried a fair part of the burden after the first couple of years. His reluctance in the first case was because of Brainerd Holmes and Brainerd's end run through Congress and the White House around him, so he was a little bit sensitive, quite a little bit sensitive to anybody estalishing external relations, as it were.
COLLINS: Did you see this experience as beneficial in the sense that it encouraged you to take a broader view of the political elements of what you were doing, or did you already have a good appreciation of that?
MUELLER: Oh, I had some appreciation from earlier. I thought it was essential in order to get the program done because you had to get the support of Congress if you were going to get the money you needed to carry forward the programs. It was equally important to get the support of the White House, so I spent a fair amount of time working on that side of the equation, not nearly as effectively as I wish I could have. The Bureau of the Budget, or the Office of Management and Budget, is one of the more difficult ones to deal with. Jim was far more knowledgeable in that arena than anybody else.
COLLINS: Resuming after a brief pause. You were talking about our connections with the Executive Branch and talking about the Office of Management and Budget.
MUELLER: Yes. Well, I established some fairly good working relationships with Ed Welch, who was at that time head of the Office of Space Policy, I guess they called it, and so I got him involved and got him working with us in terms of understanding what we had to do in order to get to the moon. He was dedicated to getting us there, and so that helped in various budget discussions going on internally. But when the impact of the Vietnam War was beginning to be felt in '68, why, we began to get under real pressures on the budget. It usually took Jim Webb going down with President [Lyndon B.] Johnson, down to the ranch or some place, to resolve the question of what was the final mark for NASA, and usually it was lower than it should have been but not appreciably during that phase of the program.
But it was essential to build the liaison not only with the Office of Space Policy but also with the PSAC [President's Scientific Advisory Committee] at that time, and they were key to much of the support within the administration for the lunar program, and particularly were they key to maintaining the support of the scientific community. It was one of those areas where Charlie Townes was particularly helpful. For example, the lunar camera we used on the moon, Din [Edwin] Land got himself involved in the selection of what needed to be done, the description of what needed to be done, and of course, he was a key player over in the PSAC. So there were a fair number of leads into the White House that needed to be cultivated in order to keep them supportive.
COLLINS: Right. But there were some negative elements there too, as well. I mean Jerry [Jerome] Wiesner was not fully supportive of the NASA program.
MUELLER: Oh, Jerry. Yes, I had quite a lot of contact with Jerry. He wasn't supportive of the NASA program. He thought we'd made the wrong decision, and so he kept asking, "Are you sure you're doing this thing right?" I kept assuring him that we were doing it as right as one could, but no, he was supportive. It wasn't that he was negative, but he was properly critical. I think the important thing was we had to choose a way of doing it, and there were three different ways that one could have done it and a matter of taste, to some extent. If I had it all to do over again, I would do an earth-orbit rendezvous, a lunar-orbit rendezvous, and a lunar landing, basically because that would have given us more technology.
COLLINS: Perhaps that would be a good note to end our discussion on today. Thank you.