Robert Everett
Founder of the MITRE Corporation
80 Rollingwood Lane
Age 74
Interviewed December 12, 1995
Concord Oral History Program
Renee Garrelick, Interviewer
Science & Technology Segment.
During World War II the relationship between technology and our nation's defense was firmly established.
I was trained as an electrical engineer at Duke University. I went to work in the Servomechanisms Laboratory at MIT in the summer of 1942. I was in graduate school at the time, and the war was on. I went to work for Jay Forrester, who was developing a hydraulic servomechanism for a stabilized radar mount for ships. The Servomechanism Laboratory was directed at that time by Gordon Brown, and there were quite a few servomechanism projects going on.
We worked on that for several years, and when that was going into production, we started a new project called an airplane stability control analyzer. This was a project to build an aircraft simulator with a general purpose computer attached so that wind tunnel data, flight data, design data could be used to set up the computer and a simulator would then behave somewhat like the actual airplane without having to fly the plane. So it was used as a design tool rather than as a training tool. We worked on that for about a year, and it became clear that to build a general purpose analog computer was going to be very difficult. The dynamic range, the flexibility and so on were all very demanding. So Jay decided to change and build a digital computer. Digital computers were quite new at that time in 1945, but from then on we started building this new computer. In the course of things, the other applications from the computer became evident, and there was diminished interest in the airplane stability control analyzer, so it was never actually built.
The central machine that we were designing for that purpose became known as Whirlwind. It was built at what became the Digital Computer Laboratory at MIT which was directed by Jay. It was used for all kinds of real time computer-based activities including number crunching, training, etc. It was really a general purpose computer and could be used for cross purposes. That's how we got into the computer business.
After the war there was the threat of the Soviets having an atomic bomb and that came as a great surprise to most people. Not that there wasn't the expectation that the Russians would get a nuclear weapon or the atomic bomb, but that they did it so quickly. They also had aircraft, which were essentially copies of the B29s, that were capable of reaching the United States on one-way missions. The combination of an aircraft reaching the United States and the atomic bomb made it evident that the United States was, really for the first time in history, vulnerable to attack from overseas. So the question of air defense, which had been worked on during the war and was still being worked on in a voluntary inactive way, suddenly became very important and a tremendous priority -- to design an air defense system with radars, computers, direction centers, aircraft, air to air weapons, communications -- all the things we thought were necessary were being worked on. We got into it because the problem of handling the data for many radars requires a lot of computation. Many radars came from the concern about aircraft that fly in low altitudes and can't see very far away, the idea was you needed lots of small radars to get the coverage. Then collecting that information and doing something with it created various computational problems.
At that time I was working with Raytheon and I introduced Jay, who had the computer, to George Valley, a professor of physics at MIT who lives here in Concord. George was working at the time as a member of the Air Force Scientific Advisory Corps on the air defense problem. Here was somebody who had a problem and somebody who had a solution. We started then to work on the application of the computer to air defense. We'd already done some preliminary work for the Navy in task force defense, so we'd already given quite a lot of thought to the problem.
MIT was at the heart of this, not only because we were there and had the computer and George Valley was there and he was working on the solution, but there were many other people at the Institute working on this. After all this was only a few years after the war, and MIT played a very large part in developing technology for the war. The major activity was the Radiation Laboratory, the Servomechanisms Lab and some other labs, so MIT had done a great deal of work during the war. A lot of the people who had done it were still at the Institute running laboratories, teaching, doing things, so it was an obvious place for major activity of this sort to be formed.
The Digital Computer Laboratory had the work of building this computer that took more and more people and they outgrow the Servo Lab. The solution was to move us up the street into the Barta Building. At first we became the Electronic Computer Division of the Servo Lab, but it was eventually defined as an independent laboratory. It was necessary in the early days for the head of a major laboratory to be a professor, and Jay was not a professor, so this problem was solved by keeping it part of the Servo Lab. After this marriage of the computer and the air defense problem, a lot more effort went into it and the computer laboratory became larger and larger.
The initial work on Whirlwind was supported by the Navy, but the Air Force picked it up. There were people at the Air Force Cambridge Research Center who were working on digital transmission of radar data, and so there was a combination of all these to build a computer-based air defense system. When the Air Force came up against the question of how to go about this, they asked MIT. MIT reacted first of all by having a study called the Charles Study in 1951, which Forrester and Valley were a part of. This study said yes, this was an appropriate way to go.
As a result, MIT set up the laboratory known as Lincoln Laboratory in 1951 to work on air defense, and the primary driver was this computer-based air defense system, but there were also parts of the Laboratory which were working on various and sundry other related things. We started out using the bits and pieces of space on campus, but that didn't work out, so the decision was to build a new laboratory out at Hanscom Field with Air Force money on Air Force land and that became Lincoln Laboratory.
We weren't the only activity. There was a manual system already in operation and activities going forward to complete that. There were people working on fighter planes and all kinds of things. There were several computing ideas of how to build the ground environment, which is part of the air defense system that collected information up in the airplanes and decided what to do to direct the fighters. So there were a number of activities.
After Lincoln was formed, the Air Force made the decision to go with a Lincoln design which became known as SAGE (Semi-Automatic Ground Environment) System. This decision was made in 1952. At that same time we had connection with IBM to build a production machine. There were a lot of other industrial organizations involved including Western Electric. A division of air defense command was set up at Hanscom to work with us in the operational design and software. All that work went on over a considerable period of time leading to the first operational center in 1958. During that period of 1952 to 1958, there was a lot of work going on.
I'm sure air defense contributed in a major way to electronics development and computers of course are electronic machines, so it did contribute a lot to electronics but more to computers. I think SAGE really was a major impetus to the computer business in the Unites States. Not that it wouldn't have happened anyway because it would have, but the effort, the attention, the money, priority and so on that the Air Force applied to the air defense SAGE project, I think had a big impact on the development of machines and software.
Ken Olsen worked in the computer laboratory and he worked on SAGE. He left in 1957 to form Digital Equipment, which was one of the major spinoffs. There were lots of other activities, people who initially worked in the digital laboratory and contributed to SAGE and went on to other things. Most of the computer companies in those days had taken the knowledge of procedures that we had developed. There were lots of students, lots of people who learned to program Whirlwind and the SAGE machines, although in a formal sense there was a heavy educational aspect that took place during that time.
In 1957 the Soviets placed Sputnik in orbit, but that didn't have very much influence on the work we were doing. We already had a very high priority. It certainly had a major impact on the activitites in the scientific areas. I still remember what a shock it was to have that happen. That really stirred up everybody and that led to all kinds of activities. But we had this tremendous priority of a very important job and we had all the resources we could ask for, and we were coming down the home stretch toward getting the system running. I don't think it would have been possible to work any harder than we were.
The thing that became evident as the design of SAGE went along was that it was not something that you just built and went away and left like an airplane or something like that. It was a software-driven digital computer-based system, and it required continuous work and updating as weapons changed, as enemy vehicles changed, as tactics changed. It became evident that somebody had to continue to do the system engineering work which Lincoln didn't do on an indefinite basis. MIT had not agreed to do that. MIT had agreed to do this job until the first module was built -- the first two direction centers were built, and the rest of the 20 odd centers were supposed to be produced by various industrial organizations that we were working with. This left a big problem as to how this was to be done. The Air Force took about a year to look into the various ways to do this. They looked into picking an industrial contractor. The other industrial contractors didn't like this idea of only one of them having the design job. They talked about a consortia of contractors and that didn't work out. They went back to the telephone company and asked them to do it. The telephone company said this is MIT's baby, we're not responsible for it. So they finally went back to MIT and said you got it. Jay Stratton, president of MIT at that time, said he didn't think it was appropriate for MIT to do this, but what he would do is spin off a part of Lincoln into a new nonprofit organization which would be configured to carry on this semi-permanent job. That's how MITRE was born.
MITRE was spun off from Lincoln Laboratory in the summer of 1958. Almost all were SAGE people with some radar people totalling about 500 people taken out of Lincoln. The design work and the computer work was in Division 6 of Lincoln Laboratory, but Division 2 was responsible for radar communications and things like that for running the experiments, and we took a good part of that division as well. We also took pieces of other divisions.
The vice president of MIT, who was given responsibility by Jay Stratton to set up MITRE, was James McCormick. Jim was a very bright, retired Air Force General. He picked the name. He swore he picked it by looking through the dictonary and came across the word mitre, which means a number of things including a bishop's hat. He also saw a definition of a shield and a bishop's hat sort of does look like a shield, so he thought a shield would be a nice reference for a company that was set up to do defense work, so he picked the name MITRE. He swore he did not know that it began with MIT. When Jay Stratton heard it he said well, you've already done it and it's too late to change it, so it became MITRE. The rest of us thought it might have been MIT with Reseach and Engineering or Jim was trying to get the Rand Corporation to co-sponsor so it might have been MIT Rand. One of the jokes was "Must I Trust Robert Everett."
I became head of the technical side of MITRE. The first president was C.W. Halligan who came from Bell Laboratories. The next year I became vice president of technical operations so I had the technical research.
The corporation is a 501C3 company which is nonprofit. It's formal term is public charity. It's like MIT. It was set up with a charter to work in the public interest and has a group of trustees who run the company and help the management decide the policies. Although we were set up to continue work on SAGE, we were well aware that was not a permanent job. For one thing, about this time, intercontinental ballistic missiles came on the scene, and nobody quite knew how to devise a defense against ICBMs. We also knew that the kinds of systems we were working on were useful for a lot of other purposes in the military and in the civil sector. The company was not limited to air defense work; it was only limited to work in the public interest. MITRE has worked on a lot of air defense systems but most of them were overseas. People are still worried about air defense, but it is now a fairly small fraction of the work.
The company was set up to enable it to have a long term relationship first with the Air Force and subsequently with the Army and the Navy and some of the other parts of the government. At the same time, this means that the company is privy to the needs of the government, what their thinking is, which way they are going. It's also privy to proprietary information of the industrial defense industry because the government needs to know where the authority is, where the emphasis is going, so it would be totally improper for the same company that is privy to all this information to compete for jobs such as in a manufacturing situation. So the rules were that MITRE is strictly an organization which has been designed to avoid conflicts of interest, and it is nonprofit so people don't make money out of the company. It doesn't manufacture anything because that would put them in conflict with other people manufacturing, and it doesn't compete with the defense industry for jobs. This enables the Air Force to give MITRE sole source contracts. This is a class of organization with federally funded research and development of which there are quite a few, eight or ten now in the Department of Defense, aerospace serves the Air Force in rockets and satellites, Lincoln Laboratories, Rand Corporation. In the event that the company is ever dissolved, the assets go back to the Federal government. Compensation for work varies from time to time. Most of the time it is on a cost plus basis with a negotiating fee. There are certain expenses that you can't collect. There have been years where we've had fixed price contracts.
There has always been a partnership between the defense industry and the Department of Defense. I think the growth electronics business has led to much more complex devices and a much more complex process of tying these things together. The classic militaries for years were given tools, like guns or bows and arrows, but the fundamental system here is a computer system. We now have sensory systems not just eyesight but radars and all kinds of things, and the electronic communications and satellites and the whole business of tying all those together has become much more complex. There is a continuing debate about the way these organizations compete and perhaps in some better world the government will do the government's job itself with its small laboratories, but that turns out to be hard to do under the rules which are in place now. They are very inflexible, so there's a great deal of flexibility in being able to go to an outside organization which can be bigger if you want bigger or be smaller if you want smaller which can change their technology, change the way they compete. I think the fundamental thing that MITRE and MITRE-like organizations provide for their sponsors is the flexibility.
That was a wonderful time that grew out of the war, the freedom to do things. This went on for a long time afterward where the amount of procedural decision making approval, auditing, wasn't very great. The Air Force wanted something done and then MIT would do it. I remember on the SAGE project we spent money on what we thought was important and did what we thought had to be done, and people accepted that and paid for it. You can't do that anymore. The bureaucracy procedures have grown very complex. It's hard to get anything done. It was a delightful time at that time, but having a war is a high price for getting there, but there is no reason why you shouldn't enjoy it when you have it.
I said before that we at MITRE understood that we would be doing other things. We started out working for the Air Force, but we really had worked for the Federal Aviation Agency because the technology that was developed for air defense was applicable to air lines. The work originally done in Bedford then was better done in Washington. Also the Defense Communications Agency has set up to work on common communications. People were worried about common land systems for the Department of Defense as a whole. We became involved in that and set up people in Washington so after a while we had a significant number of people there that we established a site in Washington. MITRE was set up in Bedford because the Air Force had its command control communications business in Hanscom.
When we started out it was just command control -- command being decision making as to what to do and control being sure that we have the systems to do it. To this was added communications after a while but to begin with MITRE, and some of the other systems divisions did not have too much communications, so it became evident that this was also a piece. Some people add a "C" for computers or an "I" for intelligence. I've just been asked to serve on a committee that is going to oversee a study on C4IS&R, Command Control Communications Computer Intelligence Surveillance and Reconnaissance, but fundamentally it is the same thing.
The company is expanding into work for the Environmental Protection Agency, as well as rapid transport, law enforcement, health care, the IRS. We recognized that we were an information systems organization, and everybody has to have information systems, so we started working for other parts of the government because they all have information. First, it spread out within the Department of Defense with the Army and Navy as well as the Air Force. And it worked for various defense agencies in the office of the Secretary of Defense. Then we started working for other parts of the government, the IRS, etc. Back in the 1970s when the energy crisis came, there was a lot of concern about that, so we started doing some work in the energy and environmental business because that appeared to be the real need. It was our feeling that if we worked for these other people that it would not only strengthen that technology, but it would also enable us to measure our capabilities with other activities and it would give us opportunities to grow in some areas. It is very important that we keep the company in some sort of stable condition.
The company is at a cross roads in terms of the nondefense work becoming a separate entity. MITRE is dividing itself into two companies, one of which will do the military and FAA work. The Department of Defense decided that their political relationship with Congress was such that the new company should restrict itself to nondefense work. The solution to this problem is to divide into two pieces. The new company will be called MITRE Tech Systems. We don't say we work in computers any more, we work in information systems. It's been a long time since we designed a computer.
When we started working on the computers back in the 1940s, the memories we worked with were either very slow or unreliable. We were forced to build today's computer memories We had much smaller memories than the ones we get in the desktop computers these days. The biggest step forward was Jay's invention of a core memory. The digital laboratory did all the research work and built the first computer and had an absolute first order of attack on the speed and reliability, which really dominated computers for the next twenty years.
