Joel Moses PhD ’67

INTERVIEWER: Today is December 14, 2009. I am Karen Arenson. We are talking with Joel Moses, who earned a PhD in mathematics at MIT, and whose interests over the years have focused on computer science, artificial intelligence, and engineering systems. He started his career as a professor in Course 6, which, at the time was called electrical engineering, and is now electrical engineering and computer science. He became head of the department, dean of the School of Engineering, and MIT's provost. He is currently one of only 14 Institute Professors and is acting director of the Center for Technology, Policy, and Industrial Development. Joel, thank you for talking with us today.

You've described your life as inextricably entwined with MIT. What do you mean by that?

MOSES: Well, I came to MIT in 1963. That was the third time I actually applied. I couldn't come here as a freshman because we couldn't afford it, unfortunately. And then, when I was graduating, I applied again and I made the mistake of accepting a fellowship at Columbia, rather than a research assistantship. And then, finally, I saw the light and I came here in '63 and I've been here ever since. When I applied for faculty positions in '67, about to finish my PhD thesis, one of the department heads said, "Can you cut the umbilical cord to MIT?" I guess I haven't. I just couldn't. It's been so great.

INTERVIEWER: What made it a good match? Is there something about the chemistry between you and this particular research university?

MOSES: That, I don't know, but what was amazing about the situation was I didn't think much about what I was doing. Things were fun. Math was fun. Doing computer science was fun. I didn't really start thinking about -- what was I doing? What's been my goal? What have I accomplished? -- until I became department head in 1981. So, it was a tremendous ride.

INTERVIEWER: Did it remain fun all along the way? Or were some things more fun than others?

MOSES: Yeah, I mean, most faculty don't particularly care for administration. I happen to like it more than most. But there are times as an administrator that are tough and not as interesting, if you will, as others. But I found even administration, especially committee work, which most people don't like, I found that interesting.

INTERVIEWER: Why?

MOSES: Well, I don't know. There's some times you get ideas in the process of a discussion that goes on for a while. You really get at the essence of issues, surprisingly, and ideas come forth that you wouldn't have expected. And that's happened to me on several occasions while I was an administrator.

INTERVIEWER: But, they're usually different kinds of issues from the ones you worked on as a scientist or mathematician.

MOSES: Yes, indeed, the administrative issues. But, nevertheless, I've found it interesting.

How did I get into administration? Well, it's almost a trap. You have early success and you get excited about doing more for the Institute, for the department. I've had early success.

INTERVIEWER: Tell us about the path before you got to MIT and what led you here. Where were you born? What was your childhood like? Your family?

MOSES: I was born in Israel -- well, at the time, it was called Palestine -- two weeks before Pearl Harbor. And I lived, for the first five years, in a village of farms. So I was born on a farm, if you will. And, my parents were not great farmers. In the midst of World War II, they came to Palestine from Germany. My father was a businessman. He was very successful in Germany. And so, he, basically, became a businessman for the farm community. He sold the produce better than anybody else could sell the produce in the nearby city called Netanya. After a while, we moved to Netanya in 1946 -- 1947, I think -- and we lived there until 1954. Netanya, when I left in '54, had a population of maybe 30,000. Now, I think its population is about 200,000. It became a major place for people to vacation.

In any case, ah we left in '54. Arrived here on September 1, 1954. Actually, we were supposed to arrive on September 2, but the Israeli captain thought that since July had 31 days, August must only have 30 days. So, we quickly had to telegram everybody waiting for us that we'd arrive a day earlier on September 1. In any case, we wound up in Brooklyn, New York and stayed there for a number of years going through junior high, high school, and then at Columbia College.

INTERVIEWER: What was it like to come to a new country? Did you find it exciting or frightening or?

MOSES: Neither exciting nor frightening. Much of it was, sort of, surprises. My mother decided to prepare me for coming here. She bought me ten pairs of pants that were short. Well, in the '50s, we didn't realize that the kids in Brooklyn weren't wearing shorts. So, that was a surprise. I didn't speak the language much at all, so I had to learn a lot. That was difficult -- took about nine months before I was reasonably acclimated to the language. A lot of things were surprising-- we didn't have supermarkets in Israel. You had to go to a milk store, you had to go to a meat store. Bakery shop -- that was what people expected to do. Here, it was all in one store. Big surprise.

INTERVIEWER: Do you think coming to a new country shaped your willingness or ability to deal with new experiences later on? Or approaches to creativity or research or anything or?

MOSES: It's possible. I don't know. What I think, as I reflect back on what shaped me, it was my parents' background, more than anything. My father was born in Germany and stayed there until they left in '39. My mother was actually born in Romania and left Romania as a result of World War I. Wound up in Berlin. That's where she met my father in '38. Their attitudes toward family, toward, actually, problem solving -- I didn't realize that -- that had a major impact on me. Possibly more than my 12 years or so in Israel. Again, yes, everything shapes you, and I was shaped by all of these countries. But, the surprise was how much my thinking was affected by my parents' German background.

INTERVIEWER: And how would you characterize your approach to problem solving?

MOSES: Well, problem solving and organization -- a little more hierarchical than the American classic approach. In any case, a different kind of hierarchy. In the '50s and later decades, most American thinkers about organization thought in terms of what you might call a tree structured hierarchy, where you have a president and vice presidents et cetera, et cetera. There is more of what I call a layered hierarchy, in thinking of Germans. It goes back, in some ways, to the Middle Ages. That approach to organization and also problem solving, it turns out, had an impact on me and makes me different from a lot of people in this country at this time.

INTERVIEWER: Were you always drawn to science and engineering? Or when did you first, as you think back, when did you first recognize that you might be? That you like these subjects or were good at them?

MOSES: I realized that early on. In the fourth grade, when the teacher had to leave town, they had me grade the papers in the class. Math came easy for me. It was enormous fun. In fact, I was on the math team in my high school. I went to a regular high school. It wasn't one of the exam schools in New York. Stuyvesant, Bronx High School of Science, Brooklyn Tech, were the exam schools. I went to Midwood High School. But in the competition, we beat Bronx High School of Science. A real surprise for them, I suspect. It was fun and we worked at it. Every lunchtime, in addition to eating lunch, we would solve problems.

INTERVIEWER: You and your classmates?

MOSES: Yeah, and the team members, yes.

INTERVIEWER: But, did you have a faculty member who sat with you at lunch and threw problems at you?

MOSES: Aaron Shapiro was the faculty member who was in charge. And I just read about him recently. You get all this information, now, on the Internet. He was treasurer of the American Math Association branch for all of New York for many, many years. He was wonderful. Really, a great teacher. I had him, not only for the math team, but also for calculus later on.

INTERVIEWER: Did you think you would become a mathematician at that point?

MOSES: No, because my parents convinced me I was supposed to become a doctor. And why that? Well, my father's cousin, he was a famous doctor. He was one of the leading doctors at the Hadassah Hospital in Jerusalem. He created the rabbit test for pregnancy and became very well known as a result. Certainly, in that generation. Young people don't seem to know about that. And then my mother's brother became a doctor. I once asked him, Uncle Sam, -- that was his name -- how did you learn seven languages? He says, Oh, that was simple. I got thrown out of six countries. In the process of leaving Romania, then Germany during World War II, he wound up in a number of countries. He learned a language in each one.

So, basically, I was supposed to be a doctor. When I went to college, they didn't have a pre-med major. So, you have to put down a major. I did what was easy; I put down math. Then, next year, I put down math again. Finally, I said, this doesn't make any sense. I'm not ever going to be a doctor. I'll be a mathematician. And then, I wound up being a computer scientist.

INTERVIEWER: Did the medicine part really attract you at all or?

MOSES: Well, I hadn't really thought about it much, as it turns out. Nowadays, actually, we do research in my center on health care policy. So, it's a turn around, if you will. But math was just too attractive, easy, and computer science was even more attractive and even easier, in some ways. So, I really went for it.

INTERVIEWER: What was your life like as an undergraduate at Columbia?

MOSES: Well, actually, I lived at home. We couldn't really afford for me to live in the dorms. What makes Columbia different from MIT is that a large number of required subjects in the humanities. And you have to think about Plato and Aristotle, music and art and things of that sort. Probably more so than MIT students really tend to do. That's the difference that I find between Columbia and MIT. On the other hand, when I came to MIT, I was blown away by the level of work in mathematics and in computer science. It was just much, much higher level than I was used to at Columbia. So, I came to the right place for those fields.

INTERVIEWER: Finally.

MOSES: Absolutely.

INTERVIEWER: Did you imagine yourself becoming a professor when you were in college? If medicine wasn't it, did you begin to try to think about what might be your career path?

MOSES: No. It just fell in place to me. For example, my last year of doing a thesis, I didn't apply anywhere. I got these calls. How would you like to give a talk at Carnegie Mellon? How would you like to give a talk at Bell Labs? How would you like to give talk at Stanford? The head of the department at Cornell called, said, "I guess you're not really thinking about us, but if you are, Cornell would be a good place to go." Basically, I gave talks everywhere but I really didn't think about a faculty position much until I was in the corridor one day and Bob Fano, who was running the lab then, Project MAC, computer science lab, said to me, "I hope you'd stay as a faculty member." Oh, okay. That was it.

INTERVIEWER: Interesting. When did you first encounter computers and was it love at first sight?

MOSES: First time I saw a computer was in 1957, I would guess. I got a job in the summer at Wall Street. I was a runner.

INTERVIEWER: So you were still in high school?

MOSES: I was still in high school. I was a runner. But they gave us a tour of the place and I saw this big, humongous machine with a huge amount of hot air being blown away, et cetera. And they told me what it did was it added and multiplied. That's all it did! This humongous machine. And I was fascinated by this idea. Nothing much happened until about 1961, when I was able to take a course in programming. And boy, that was fun! Enormous fun!

INTERVIEWER: You were in college then?

MOSES: Yeah, now I was a third year in college, but it was actually my last year. I graduated early. And I guess I did so well, they gave me a job working in the center, there. Working for IBM on weekends and then during the summer. And that's how I got into it. So, I moved from mathematics to computer science just because one was more fun than the other.

INTERVIEWER: And did it strike you, at the time, that this was a burgeoning field and that maybe you could make your way in it?

MOSES: I don't think so. I wasn't planning in those days. You just went with the flow. You went with what was most interesting. So, what happened was I got the idea after I graduated college and while I was doing my Master's. I could combine the mathematics and the computing by using computers to do mathematics. And I read a PhD thesis done here under Marvin Minsky at MIT, by a student, Jim Slagle, who did a PhD thesis on doing integration problems in the calculus. I read it and I said, gee! I could do better. I think I should go work at MIT, work for Marvin Minsky, and redo that, building on what Slagle had done.

INTERVIEWER: And you were an undergraduate when you were --

MOSES: I was an undergraduate when I first heard about it and then all through my next year, into my Master's year, I was thinking about how I would go about doing it. As it turns out, I didn't do anything like that. Not what I'd planned, initially. But it was the same topic: integration problems in the calculus.

INTERVIEWER: And so, for your Master's, you were still in the math department at Columbia working on what kind of problem?

MOSES: Well, basically, no problems. You have to do a number of courses and then at the end of the year, they gave you an examination to see whether you could go on to the PhD or did sufficiently well to get a Master's. The department head saw me in the corridor and he said, "We stopped grading the exam when we decided that you did sufficiently well and that's because we know you're leaving." So that was the end of that.

INTERVIEWER: They knew you were going to MIT at that point.

MOSES: Yes, they knew in May of that year that I was leaving.

INTERVIEWER: When you moved to MIT, though, you stayed in the math department.

MOSES: Correct.

INTERVIEWER: Why?

MOSES: Well, the options were going into electrical engineering, which was building up its computer science at that point, or in the math department which had some computer science but not much. But since I had degrees in mathematics, I figured I could do the required courses and examinations in mathematics. If I'd gone to the electrical engineering department, they would've asked questions about electrical engineering, which I wouldn't have been able to answer. So, the math department was quite flexible and it was a great experience because pretty much, what I wanted to do, they let me do it, and I enjoyed myself.

INTERVIEWER: Were the professors in the math department aware that you and maybe others were increasingly interested in the computer science field? Did it show? And did they care?

MOSES: Well, Minsky was in the math department for a while, then he transferred over to electrical engineering. So, they obviously knew that. At some point, they appointed Seymour Papert in the math department, so they knew that they needed to build up a little bit. But it didn't become a major issue, as far as I could tell, until the mid '70s, when Ken Hoffman was department head and he felt that he needed to build up theoretical computer science. And he did so, slowly. And now, they have a very strong theoretical computer science program in addition to the theoretical computer science program that we have in the EECS department. So, they all work together in our CSAIL laboratory -- Computer Science and Artificial Intelligence Lab.

INTERVIEWER: But in two different departments.

MOSES: But in two different departments.

INTERVIEWER: Is that an effective way to go about it? Are there pluses and minuses to doing it that way?

MOSES: It works for us. It really does. I remember once asking the guy who's currently head of the math department, who's a theoretical computer scientist. I said, "Why did you go to math instead of EECS?" He said, "Well, when I interviewed in math, the head of the department was so relaxed, he fell over in his chair." He said, "Any department where the department head is so relaxed, that's the department I should join!" Now, he's head of the department.

INTERVIEWER: What were your first impressions of MIT when you came here?

MOSES: Oh, jeez, this place is tough! It's so much harder. The level of the courses that I initially took in mathematics was so much higher than what I was used to. So, I started immediately taking courses in the EE department, where the computer science courses were. Things that I could handle reasonably well and I enjoyed. And the math department allowed me to take a number of computer science courses in the EE department as well as math courses in mathematics. And so, my first impression was, tough! Oh, my God, am I going to be able to do this? But then, later on, it turns out, it was okay. It worked.

INTERVIEWER: How many of your classmates came in with the same reaction you did? If everybody, or almost everybody, came from the outside, did they all tend to feel, "Gee, this is hard!?" Or were most of the students from MIT, and so they were just going along? It's interesting that you would have come from a very good university and have that reaction.

MOSES: Good point. I didn't know many. Some of the ones I did know in the math department who were Columbia graduates, had an easy time of it. They picked up a field, a subfield, in mathematics and did a PhD within two years, even. So, I can't say what the general impression was. I just didn't know that many graduate students in the math department. So, I can't really say.

INTERVIEWER: What did you do your thesis work on?

MOSES: Eventually, it took a while to convince Marvin Minsky to allow me to do it, but I redid Slagle's thesis using more mathematics to start with. He was, I think, trying to see if he could, in a sense, do it the way freshmen did it. Well, I was trying to see if I could do it the way expert mathematicians did it. So, I read up the literature, which was reasonably -- well, not fully developed -- on how to think of this as a mathematical problem. And then, I did a computer implementation of that.

INTERVIEWER: Which was to use computers to --

MOSES: To do a symbolic integration. And, not the way freshmen might do it, but the way experts might do it. Maybe even better, in some cases. Later on, we figured out how to do it in a way that surprised professional mathematicians.

INTERVIEWER: And why did that question interest you? Did you feel like it was important? Or, what about it tugged at you?

MOSES: It combined my various interests. So, on one hand, mathematics, and the other hand, programming, and the third could be viewed even as AI, or at least initially, was AI -- artificial intelligence. So, by combining all three into that one problem, yeah, gee, it seems like a good thing to do.

INTERVIEWER: Artificial intelligence became one of your strong interests. How did you get into it? And, where did you think you would go with it? What about it fascinated you?

MOSES: Well, Minsky supervised the first thesis -- Slagle's thesis -- on integration. And so, I went to work for Marvin. He was one of the fathers of AI. There are, essentially, four: Minsky, McCarthy, Herb Simon, and Alan Newell. So, I took his courses and I learned about AI from him. And, I tried to advance that part of the field as well as understanding of integration and build, if you will, a mathematical laboratory so that engineers, scientists, could use it to solve their problems. So, it had a useful aspect to it. It had an AI aspect to it. It had a mathematical aspect to it. All combined.

INTERVIEWER: And, the feeling that it was sort of a technical step by step approach as opposed to sort of being fascinated in some grand way about what machines would do in the world or sort of not a science- fictiony love.

MOSES: No, it wasn't science- fictiony. This was real, practical stuff. Now, I have to tell you, there was another graduate student of Minsky's in the EE department, Bill Martin, who was working on creating what he called The Symbolic Mathematical Laboratory, which allowed us to understand and work out step- by- step processes that engineers or scientists might use in order to do symbolic calculations. So, we had to differentiate and integrate and simplify and all kinds of things that people do. And, he was working on the step by step process. He was also working on a beautiful display of expressions which didn't exist at the time. And, I was working on differentiation, integration, a little bit of simplification on my side. And we combined forces after we finished our theses.

So, we had a very practical aspect to it. In addition, of course, the rest of the Marvin's students were mostly interested in the larger issue of artificial intelligence. And, I was, too, but I parted company a little bit from Marvin's approach partly, now, I think, because of my German background, because I created what was, later on, called expert systems or knowledge based systems. And, that was a different approach than the one that was initially used in AI. AI, at the time -- say, mid '60s -- the feeling was that you had to solve a problem by not knowing much about it and just looking at the statement, some axioms, et cetera. Do it very logically. Work out the details, et cetera, et cetera. I said to myself, "This is not likely to work on hard problems." Are you going to be able to re-derive Einstein's Theory of Relativity without knowing something about physics and some of the experiments that people have done? I didn't think so. I thought you really have to know something. That was a surprise to people, I think, to people in AI. I pushed that for a while, but then I dropped out and I became more of a pure computer scientist.

INTERVIEWER: To what extent is that premise accepted by people in AI now and to what extent has the field moved away from it?

MOSES: Well, it's complicated. To some degree, I think, the notion that you have to know something is certainly accepted. A lot of AI, nowadays, is what you might call Modern AI, which is, to people who are interested in certain parts of the problem, computer vision, robotics, learning theory, speech understanding. These can be combined, but they usually are not. People create rather deep specialties in each one of those areas and the goal is, largely, to do very well and be helpful. The goal of trying to understand how the mind works is lost a little bit and we're now trying to get back to it, I think, for example, using some ideas from cognitive science. My own particular pet idea, which is not necessarily widely followed, is to, essentially, try to understand what you might call the organization of the brain. How does the brain do it? And then, try to see to what extent we can use that to utilize computers to better solve problems.

INTERVIEWER: Are you still working in this field now?

MOSES: Very little. It's not one of my major areas, at this point, but I might get back into it.

INTERVIEWER: When you first moved to Boston, even before you started class, you worked for a summer at Lincoln Labs.

MOSES: Yes.

INTERVIEWER: How did that happen? And, what was that experience like?

MOSES: Well, I knew at that point, say in May or June, I knew that I was coming to MIT and I was looking for a summer job. And, lo and behold, I found out somebody was interviewing that day from Lincoln Lab. Oh, gee, I think I know something about Lincoln Lab. So, I got on the schedule. I was the absolute last person he saw. And, he asked me, "What would you do? " And, I said, "Well, I know something about the computers you use. Something about the problems you're interested in." He was pleased enough, he recommended me for the job. So, I went that summer to work at Lincoln. The supervisor said, "Well, we're now developing a special machine with word length of 4,096 bits," where as most word lengths were 36 bits or 32 bits or whatever. This was huge! What could we do with it in order to do language? Gee! And then, he left for the whole summer! And, I was I left to myself to try to figure out what to make of it. And, I, basically, came up with an analysis of how to use 4,096 bits in searching records. Information retrieval it would be called today. For words as well as numbers. And, when he came back, he wasn't thrilled. What can I say? He wasn't there all summer.

INTERVIEWER: This was during the Vietnam War. It was a time when a lot of critics were protesting against MIT's relationship with Lincoln and the instrumentation lab because so much of their work was for the defense department. Did this concern you at all? Were you aware of it? Did you think about it?

MOSES: You're jumping a little bit ahead. The major issues related to Vietnam occurred, I think, '68 to '70. And so, it's five years after I entered MIT. Fall of '63 or, if you will, June of '63.

INTERVIEWER: But the summer you worked there, did it cross your mind?

MOSES: At Lincoln, no. At Lincoln, it did not. Of course, you couldn't help it when you went on campus, as time went on, because it became such a major national issue as well as a campus issue. I was somewhat involved in it eventually, largely because of the young lady who became my wife. She was in Northeastern and she was chosen to represent all the Northeastern students in discussions about what to do about the war, et cetera. Some of those discussions took place here at MIT. She led me into that situation to some degree. We went to Washington a number of times.

INTERVIEWER: If you want to come forward now to the 1980s. I think you were beginning to refer to when you chaired a committee to reexamine MIT's relationship to Lincoln Lab and whether it should be changed. How did that go and what did you conclude? And did your experience during that 1970s summer affect how you approached or how you thought about it during the committee work?

MOSES: Well, indeed. One of the department heads -- the one who hired me into the EE department, Louis Smullin -- and Louis, in the late '80s, he may have been retired at that point. I think he was. But, nonetheless, he made a point to question whether MIT ought to continue its relationship with Lincoln Lab. MIT supervised, if you will -- well, not supervised; that's the wrong word -- effectively managed, but it didn't, in any detail, the operation of Lincoln Lab. And, following the situations in the early '70s that we alluded to earlier, Lincoln Lab, basically, could not go beyond a certain point in developing technology. They couldn't make war- fighting equipment, as Draper did. And, that's why Draper needed to be separated from the Institute and Lincoln. What had to be done is that the president had to promise, guarantee, if you will, to the trustees, that Lincoln was not building equipment beyond a certain point. In any case, Louis made a request that we reexamine this. John Deutch was the provost at the time and he asked me to chair a committee to look into the relationship. I not only went and interviewed people at Lincoln, but I also took a tour around the country. I went to see what things were like at Livermore, Los Alamos, at Johns Hopkins' Applied Physics Lab, at Caltech and decided that, you know the committee decided basically, that not to make a major change in the relationship but to make sure that there were more faculty involved in understanding what was going on at Lincoln and that there was more joint research between faculty at MIT and the campus and the staff at Lincoln. And, all of those things came to be.

INTERVIEWER: Let's go back to when you were hired as an assistant professor. When you were approached in the hallway, or whatever, was that, essentially, an invitation to become an assistant professor as well as a researcher?

MOSES: Yes, it was an invitation to become an assistant professor.

INTERVIEWER: Had the department voted on it or anything?

MOSES: I have no idea. I have no idea. Things were "loosey- goosey" in those days. Searches were not national, necessarily. I don't know to what extent advertising was used. Basically, men -- largely men -- made decisions who was good. And, they would call their friends in other institutions, major ones like Berkeley, Stanford, and exchange information about who's available and things of that sort.

INTERVIEWER: But after getting your degree in Course 18, in math, you ended up as a professor in six.

MOSES: But that was natural.

INTERVIEWER: Was it?

MOSES: Yes, it was absolutely natural.

INTERVIEWER: They didn't ask you if you knew any electrical engineering.

MOSES: No, indeed not. But computer science was a growing field at the time. Where did the faculty come from? Well, Minsky came from math. Corbato came from physics. Many of them came from electrical engineering. And, the locus was in the EE department. And so, the transition, which Minsky already made from math to EE, was natural for someone interested in computer science and AI and computer science theory. Yes, all of those. And so, it wasn't a significant transition.

INTERVIEWER: Was that fairly typical in other universities, too?

MOSES: Every university seemed to be dealing with things in their own way. Probably, most complicated may have been Berkeley, where they had some computer science in the math department, computer science in the EE department, and then they combined faculty at some later point than we did. Let's see, in '67, there was just barely a computer science department at Stanford. Just recently, maybe, one at Carnegie Mellon. In the next 10 to 20 years, of course, there was a significant growth in the number of departments of computer science. Now, EECS was an unusual model.

INTERVIEWER: Although, it didn't become called that for a while, right?

MOSES: Yes, it didn't become called that for a while, but having so much computer science within an EE department was unusual and it led to interesting events later on.

INTERVIEWER: Which were?

MOSES: Well, in '73, Louis Smullin, the person who hired me, announced that he was going to retire. At that point, we had already hired quite a few computer scientists. Barbara LIskov, Albert Meyer, and Jerry Saltzer. These are the new generation. I was in that generation. There was a move afoot to split the department into an electrical engineering one and a computer science one. Now, mind you, in the '70s, the president, provost, the chancellor were all electrical engineers. Wiesner, Walter Rosenblith, Paul Gray. The power was, if you will, in the EE- side and I thought that it would be wise to stay as part of the EE department because of such a tremendous history, such great strength, and so powerful. I didn't buy the notion that by splitting, you'd gain power, which was the key argument. And so, I worked against it.

INTERVIEWER: And, the three of them were predisposed to keeping it together? Or did they have any -- Paul, and Jerry. I didn't realize Walter was in Course 6, also.

MOSES: Oh, yeah. I don't know. I didn't talk to them. Because, at the time this happened, I was untenured. And it so happened that I got elected chair of a committee that Louis created called the advisory committee in the EE department and I was elected chair. So, now the question is, how do we advise a department head about what we should do about this issue? And, basically, a number of us thought it would be wise not to split. And, I came up with a questionnaire which I sent to every faculty member. And, I said, "Well, some of us have thought about changing the name of the department and here are three options: electrical engineering and computer science, electrical and computer science and engineering, electrical engineering and science, -- well, electrical science and engineering and computer science and engineering. What do you prefer?" I don't remember the exact number but it was around 80 percent said electrical engineering and computer science. Only one person said, "Gee, can't we keep the old name?" That was Louis Smullin. Anyway, so I used that with the visiting committee -- you know, these all powerful visiting committees of trustees and others that MIT has and is one of the reasons it's so strong is because of this particular concept. And so, we presented the results to the visiting committee but at that point, we had a new administration in the department. Bill Davenport was chair. Corbato, Penfield were the associate chairs and they took that suggestion forward and more than a year or so, the name was changed. Electrical engineering and computer science. And, as a result, I think, the discussion about breaking it up had pretty much died.

INTERVIEWER: Because you were giving fuller recognition to the computer science part of it?

MOSES: Correct.

INTERVIEWER: And, did you know that when you put that survey out?

MOSES: Absolutely.

INTERVIEWER: Had you thought that through?

MOSES: Yeah, oh, absolutely.

INTERVIEWER: So this was a political move, in a sense?

MOSES: It was a tricky political move.

INTERVIEWER: Your first one?

MOSES: No.

INTERVIEWER: No?

MOSES: That was the one that some people remember.

INTERVIEWER: Is that, do you think, the beginning of your path toward administration?

MOSES: No, what happened there was the advisory committee was created in '71, maybe. Maybe '72 -- let's say '71. And elected its own chairman. It was Michael Dertouzos, who later became, for many, many years, head of our Laboratory for Computer Science. And, Michael called me one day. I was in Michigan giving a talk. And, he said, "I'd like you to chair a committee." And, the committee's job was to figure out what the impact would be on the department if it could no longer grow in size. See, what happened, apparently, -- well, I looked at data, so it's not so apparent -- yes, what happened was, beginning in the mid- '50s, the size of the faculty at MIT grew. And, it grew in the EE department, especially. So that, by '68, we had 125 faculty. Unbelievable. Wow.

INTERVIEWER: In Course 6?

MOSES: In Course 6. And, that was not sustainable, given the budgetary situation.

INTERVIEWER: The kind of growth or that size?

MOSES: The size and beyond that, the growth. Certainly not the growth, but even that size could not be sustained. And so, the question is, given budgetary restrictions, what was likely to be the outcome? And, I came up with a fairly simple model that -- one variable model -- that said if the probability of making tenure of the people we currently have is similar to what it has been in the past, we have now, currently, 45 percent tenured. Within a decade, we'll be 80 percent tenured. If the numbers don't grow -- total numbers -- our ability to hire will be much reduced. May be able to hire two, three faculty a year as opposed to six or seven. It was so simple and straightforward, people looked at it and said, "That's got to be right." Of course, it was right. It came to be and the dean of engineering, I think, sent it around, the Institute and people knew who I was all of the sudden. It was fun. Another fun thing to do. So, I became a creature, in some ways, of that committee -- the advisory committee -- to Louis Smullin. We not only made suggestions about the name of the department -- of whether it should split or not. We made suggestions about what the required core courses in the EECS should be.

INTERVIEWER: So this is the very same committee that Mike Dertouzos had --

MOSES: Yes, he chaired it for a year, or so, and then, he suggested -- of course, he couldn't impose it as well -- he suggested that I replace him. And, I became the chairman the next year.

INTERVIEWER: And you were already on the committee.

MOSES: I was on the committee.

INTERVIEWER: As an untenured --

MOSES: Oh, absolutely. As an associate without tenure.

INTERVIEWER: And, did it ever cross your mind that perhaps being on the committee, or the outcomes, or what you said would affect whether or not you got tenure?

MOSES: No. Basically, again, I didn't care. Gee, I'm having a good time and these are important issues and let's do something about it. And if I get tenure, fine. And, if I don't get tenure, oh well. Remember those calls I used to get? Well, they'll call again. I just wasn't planning for myself. And, I just did what seemed like a natural thing to do. Things were going well on the research side, in my opinion. I was never an outstanding teacher -- I was okay, I suppose. So, things were going okay in that regard. I was spending more and more time on this thing and, hey, it's doing good for the Institute. I loved it.

INTERVIEWER: And, the research you were doing at that time was?

MOSES: Was an outgrowth of my PhD thesis and Bill Martin's work. At that point, Bill Martin had already separated himself from the projects. I was running the project. We built a system called Macsyma. It's a name that I invented because it means different things in different languages. Latin and it even has a Hebrew meaning. It's related to the word, kismet. Maxeema -- magical, wondrous. And so, when I realized the multiple meanings, I thought, oh, that's a good name. And, we were building this system and at that point, we were probably getting a lot of users on the ARPANET, which was a precursor to the Internet. The ARPANET published the leading sites and for a while we were the number two site in the country because people were logging in on our computer.

INTERVIEWER: The number one site being?

MOSES: I don't remember.

INTERVIEWER: Another university one?

MOSES: I don't know. It may have been something like BBN or, I don't remember what the number one site was. And, of course, we weren't number two forever. Just for a few -- some months -- or, maybe, a year or so. Again, people were having a good time using our system. Finding errors, yes, every once in a while. They were publishing papers. They were publishing papers.

INTERVIEWER: What did it mean when they used the site?

MOSES: Well, they had problems, let's say, in engineering or in physics -- often in physics -- which, you'd have to use formulas. They had to differentiate or, in some cases, integrate, simplify, do power series expansions of one kind or another. All kinds of people using it. And they were talking to us by e-mail. It was fun.

INTERVIEWER: Did they tend to use it because it would have been difficult to get the answers, otherwise, or it was just faster?

MOSES: It was faster. And, in most cases, more accurate. Now, every once in a while, a bug would come in, so I can't guarantee the accuracy. But, yes, the accuracy was an issue as well as the speed. Sometimes they would calculate -- this would have taken me six months of handwork, et cetera, and now I can do it in an hour or whatever. And now, it's become a standard situation. In the field, you have systems that millions of people use. The outgrowth of what we did in Macsyma. Systems like mathematica --

INTERVIEWER: Now, a calculator -- a sophisticated calculator -- can do some of that or no?

MOSES: Yes, yes. For a long time, TI's calculators were able to do some of that. But, again, the large systems -- again, are outgrowth of work we did in the late '60s and '70s.

INTERVIEWER: And, besides the committee work and then, this research what was it like to be a young professor at MIT at that time?

MOSES: Oh. Good question. Teaching, of course, was another issue.

INTERVIEWER: How much did you teach?

MOSES: I taught a subject a term.

INTERVIEWER: Only one.

MOSES: Like everybody else. Yeah.

INTERVIEWER: Undergrad?

MOSES: In the School of Engineering, the tendency was one subject a term. Later on, I found out that it wasn't true everywhere. So, for example, in mathematics, they teach three subjects a year. One in one term, two in another. In humanities, they teach four subjects a year. I didn't know that at the time. I thought everybody taught just one subject. An undergraduate subject one term and a graduate subject another term. That's what I was teaching. And, the undergraduate subjects were a lot of fun. Especially, if you could get freshmen in the first term because they hadn't yet been inculcated into the MIT system and they were still quite excited about what they're learning.

INTERVIEWER: And they weren't later on?

MOSES: By second semester, they were already into the issue of, "Let's see. What's the minimum I need to do to get an A or to do well in this course?" Less so in the first term. Especially before Thanksgiving. Something happened around Thanksgiving. But, the key is that the older students had a lot of influence on the attitudes of the freshmen. Yeah, that's how it was. It still is.

INTERVIEWER: Did the existence of pass-fail affect attitudes at all, that you were aware of?

MOSES: I was there in the transition. Pass-fail came to be, I think, around 1968. I started in '67. My first semester of freshman was well before pass-fail.

INTERVIEWER: It may have been 1970, or so. I'm not sure, exactly.

MOSES: '68 is when, I think, Walter asked Amar Bose to chair a committee looking into either pass-fail or IAP, I forget which. But, those two issues came out in the late '60s so, I don't know enough about the change that occurred because I only taught for one year -- or maybe two -- before pass-fail came to be.

INTERVIEWER: What courses have you liked teaching the most? Are there any that are just more fun?

MOSES: Certainly, in the early years, the undergraduate programming course was fun. In fact, it's interesting; in '78 I was supposed to teach a new approach to the introductory course. And, Hal Abelson and I were going to develop this course. And then, it turns out, I became associate department head and I wasn't teaching. And, Abelson then changed to Gerry Sussman and they developed this remarkable course, 6.001, which they taught it for 25 years, or whatever. It was one of the most popular courses at MIT.

INTERVIEWER: And 6.001 is?

MOSES: Is, essentially, an introduction to programming using a LISP- like language. LISP, being the fundamental language used in AI and Sussman, in the mid '70s, figured out how to fix a technical problem with LISP and created a language called Scheme. That's another story, in itself. How did it get to be called Scheme? In any case, it was extremely popular and much better than anything I could have been involved in. I'm glad that I got out in time.

INTERVIEWER: So, it was actually about 14 years before you were pulled into administration. It sounds like you were doing a lot of committee work.

MOSES: Actually, no, it was 11 years.

INTERVIEWER: I guess before you became department head.

MOSES: Yes, I was associate department head. Associate department head, in that context, was, basically, a full time position. At least, I made it pretty much a full- time position. I was in charge of all the computer science faculty. 30 to 40, I forget the exact number at that time. And, that's a sizeable department in itself. So, Dick Adler was the associate on the EE- side and Gerry Wilson was the department head, and he was pretty much -- pretty amazing -- what he got done in those three years before he became dean of engineering.

INTERVIEWER: And, what did he get done in those three years?

MOSES: Well, he started a lot of things, some of which I was able to finish. What a key issue was the department's move into VLSI -- Very Large Scale Integration. Because, it's EE and CS department, the EE- side could deal with the, if you will, hardware part of chip making, and the CS- side could deal with the design part of chip making. And, they could combine forces readily. Tremendous argument for maintaining an EE and CS department. We were looking for such arguments. The other thing is we got Jerry Wiesner to get some money for us to buy some computers. We really didn't have educational computers worth beans! And he did it. It's great.

INTERVIEWER: He didn't tell you to go build them.

MOSES: He didn't tell us to go build them. He made some calls and he got, I don't know, a million or so. Which, obviously, was worth a lot more in those days.

INTERVIEWER: These were from where? Digital Equipment?

MOSES: Yeah. Largely DEC. Yeah, largely DEC. And, we got a big machine from DEC. And, we got a computer center to put it into with air conditioning equipment and all that. All of that happened. That was a minor thing. And then, he started a number of things. The department was going to have its 100th anniversary of electrical engineering. Not the department. The Institute was going to have the 100th anniversary of electrical engineering education. It started in the physics department in 1882 and in 1982 would be 100 years. And, in preparing for that, we did a history of the department and we hired a writer to work with one of our older faculty members. That book came to be. Gee, I'm forgetting one or two major things that he did. Oh, yeah, the EG&G building.