Harold (Hal) Abelson PhD '73

INTERVIEWER: Today is August 23, 2011. I'm Chris Boebel.

As part of the MIT150 Infinite History project, we're talking with professor Hal Abelson. Professor Abelson is the Class of 1922 professor of electrical engineering and computer science at MIT. He holds an AB degree from Princeton University and a PhD degree in mathematics from MIT. In 1992, Abelson was designated as one of MIT's six inaugural MacVicar Faculty Fellows. Professor Abelson is co-director of the MIT-Microsoft iCampus Research Alliance in educational technology. He's co-chair of the MIT Council on Educational Technology. And he serves on the steering committee of the HP-MIT Alliance. He has played key roles in fostering MIT institutional initiatives such as MIT OpenCourseWare and DSpace.

Hal, thanks very much for taking the time--

ABELSON: Sure.

INTERVIEWER: --to talk today. So let's start at the beginning.

ABELSON: The beginning?

INTERVIEWER: The beginning. Just start by telling me a little bit about where you grew up and the early years?

ABELSON: So the real early years, I grew up on a chicken farm in central New Jersey. My parents worked on the farm. My father later went and worked in the post office. I was the first one in my family to go to college, although my parents were always very, very interested in education. My parents liked music, education. I had an uncle who sort of was the college person, who worked for Exxon and had a high level position there. So we always cared a lot about education. I'm not quite sure where I glommed onto math and science. I remember when I was I guess 12 or 13, I decided I wanted to be a physicist, mostly because I think I liked the word. But I didn't know what that was.

I did sort of okay in school. When I talk to kids these days, I remind them that when I was their age in seventh or eighth grade, I was getting D's in mathematics. In fact, like a lot of people who end up going into computers, we are really bad at doing arithmetic. Someone once said the reason a lot of people go into computers is to make up for their own handicaps in doing that. So I wasn't particularly good in any of those subjects early on. And then about the time I got in high school, I started being good at it.

One of the things the school did, they hired somebody to help me learn calculus because of course our school did nothing like that. And I don't know, sort of grew a lot of interest in math and science from there. So at some point I said gee, I don't want to be a physicist anymore. I want to be a mathematician because I guess that word sounded better.

And then I got into Princeton on a scholarship, which was nice because my family could never have afforded anything like that. And did pretty well at Princeton. Started getting interested in a subject called algebraic topology. And then got into MIT graduate school on that. And ended up coming to graduate school at MIT mostly because my wife and I were looking for places we could both go. We weren't married then, but she was looking for a place where she could go. She got into BU's School of Social Work. I got into MIT.

MIT also had a draft deferment. So back in ancient days, one of the things that you wanted to do was if you were worried that getting drafted to go to the Vietnam War, so MIT did something where I think because I was teaching or something like that, there was a category of draft deferment you could get. So we ended up coming to MIT.

And MIT was just absolutely an fascinating place. One of the nicest things, when I first showed up at MIT, we lived in Westgate. And one of the great memories that we had of being in Westgate was looking down over the field and at the other end of the field, seeing the tear gas on Mass Ave. Because that was the time when even MIT students were protesting stuff. They were protesting, for anybody who remembers, Draper Labs used to be called the Institute Labs. And there was a protest about how is it that on the MIT campus, there was classified research related to the Vietnam War? It was just sort of wonderful. Just marvelous to be here and look through this thing.

And what do I remember of MIT? I remember going up to the door on 77 Mass Ave. And gosh, they were automatic doors. And coming from Princeton, which is an ivy covered-- it's the last thing in the world you would ever imagine like that. And then there were these long, long halls. And I always had these visions of getting on roller skates and going down the halls these automatic doors. It was just a wild place. It was all one building. I didn't know quite what to make of it.

But then early on, it could have been a couple of weeks after I got here, the Students for a Democratic Society, if anyone remembers that from the '60s, that was the campus activist organization. It took over the President's Office. And they did this great thing. They had open house at the President's Office. And they were sitting there replaying the tapes so they could see what was typed on the electric typewriters in President Howard Johnson's office. And they held open house. And I said, hey, I'm a new graduate student. I'm never in my life, ever going to be able to walk into the MIT's President's Office. So let me go sit around and see what's in the President's Office.

So we walk in. And there were a bunch of students kind of sitting on the floor of the President's Office. And I see a guy that I'd been to high school with, who I knew had gone to MIT, but we'd had no contact since then. And I went up to him and said hi. Gee, hi, Ronnie. This is sort of really nice. And he says, what are you doing? And I said, I'm a new graduate student. I'm looking for what to do. What's interesting to do around here? And he said, oh, why don't you go over to the Artificial Intelligence Lab. I hear they're doing interesting stuff. So I wandered over there and talked to some people. And that's essentially how I got into computer science and AI and all the stuff I'm doing now.

And I like to tell that story to students who worry about what they're going to do when they grow up. Because I like to tell them it will be random, and you have no idea. But at some point, some totally random thing will happen that will have a tremendous influence. So don't worry about it too much.

INTERVIEWER: So backing up for just a moment, had you done work in computer science before that at Princeton, or even previously? How did you first--

ABELSON: Yeah. At Princeton I did-- well, there sort of wasn't a computer science major then. In fact, I'm not even sure there was one at MIT when I showed up. But at Princeton, a bunch of us worked in the computer center. And the Princeton Computer Center had an IBM-- I think a 7.090 or 7.10. It got one of the first 3.60s. And they were working on this very weird thing called time sharing. And a bunch of us went around and messed around with that and did what most of Princeton considered to be a totally worthless, stupid thing with computers, which is actually using it for text processing. Because everyone knew that what computers were really for were solving big complicated equations. And the computer center used to take down the system a couple of hours each week and let those of us who wanted to screw around and waste time, play with this text editing thing. Get tremendous complaints from the astrophysics department because people had to stop trying to solve on science equations.

So I became one of the people who were messing around with that. And I became a little bit of a consultant, an undergraduate consultant in the computer system, teaching people to migrate their programs from FORTRAN to a language that IBM and a couple of other places were pushing then called PL/1, a very modest name, Programming Language 1. And this was supposed to be the be all and end all of the new computer stuff. And there was a big move to say you should translate all your programs into PL/1. So I became one of what was called the PL/1 kids and would consult with faculty in the other departments about how to modify their programs.

So I had some stuff in computing and I knew about that. I knew very little about what I actually went into at MIT. But it was a background in computers, enough to walk into the AI Lab and do interesting things.

INTERVIEWER: But at the time you came to MIT, initially there was not a sense that you would be moving in that direction?

ABELSON: No, I was in the math department. And I was a graduate student in the math department, became an instructor in the math department after that, and only moved into computing later on because some of the work I was doing in topology was related to distributed computing. And then sort of the topology part of it atrophied and the computing part of it grew. And eventually when into computer science.

INTERVIEWER: So you mentioned the AI Lab. Talk a bit more about working there and also you worked with Seymour Papert in the Logo Lab as well, right?

ABELSON: Well, that's actually how I started in the AI Lab. For some reason I went-- there was an early Logo presentation called I think Teaching Children Thinking, which was in 26-100. And it was this big thing where Seymour Papert and his assistant for many years, Cynthia Solomon, talked about the Logo computer language and showed the Logo turtle. And I looked at it and it was pretty fascinating. And pretty wild. I remember later on reading a report from a journalist who's been there. And her mind was just blown about this notion that children should do these weird robot things. And I remember her saying, I looked at this thing and I thought these people were witches.

But anyway, I thought they were cool, whether they were witches or not. And I walked over to the AI Lab, looking for something to do. And walked into the elevator. And then elevator stopped at one of the floors and then Seymour got in. And I said, hi, I like your stuff. I'm a new graduate student. Can I work for you? And he said, yeah. And that's how I got into this whole business. So another bunch of random events.

INTERVIEWER: So you've talked a little bit about what it was like coming to MIT in the late 1960s, what your impression was of the campus. But give me a little bit more information about sort of the culture and the environment at that time. I mean obviously it differed as you've said from Princeton and was unlike other things you'd seen.

ABELSON: Well to me, it was really strange. I mean MIT is such a big complicated place. I've often thought that well, after being here about 20 years I finally got the ability to think of the place as a whole place. And I just don't even know how it's possible to do that because there are just so many things going on and it's so complicated. But my impression from Princeton is that it was very strange and very regimented.

So I remember hearing about the Experimental Study Group. So what people told me about the Experimental Study Group was that it was a place where you could learn on your own and do stuff that's independently and it was very open. And I remember going over that, because I had some friends who were in the ESG. So what do you do here? What do you do that's independent? And they said something like, well we get to take the standard course curriculum, but we get to study it by ourselves.

And I just remember breaking down laughing. I said, that's this place's idea of what independent learning is? Because I'd been in a program at Princeton, a very '60s program, I think even Princeton stop doing it, where they admitted you. And they said hey, you're a bunch of smart people. You have no requirements. You had department requirements. But the rest of the time, Princeton's a big place, take what you want. And all of the standard things you'd call distribution requirements or university requirements are just abolished for you. Some of us totally exploited it. Some of us had a great time. I ended up with a couple of friends essentially living as graduate students in the math department.

But coming from there to MIT, it was this sense of being regimented and precise. Later I've come to identify that as an engineering culture. I see the same thing happening at Google, because I've been spending time at Google. So you see this culture and you love it. I mean I just love engineer's culture. I'm an engineer and I do that. But one of the characteristics of an engineer's culture is you sort of sit and you say, what's the problem, what are you're trying to do? And then about three seconds later you say okay, what's the solution? And you think for about two minutes and you figure out the solution. And then you just drive to that solution as fast you can. And looking back, I recognize that's how I reacted to MIT. So there was no sense that the requirements were wrong or any of that. It was just we figured out this is the way to do it. And now we are just going to drive to that as effectively as we can. And MIT's kind of like that. And there are wonderful parts about that. And there are sort of bad part of that.

INTERVIEWER: Tell me a bit more about working with Seymour Papert. What was that like?

ABELSON: Well, Seymour was, is amazing. So I guess people know that tragically he was very severely injured in Vietnam. And his brain was very affected. And he's under rehabilitation now. But Seymour was just absolutely amazing-- I mean his childlike sense of wonder at learning. And what Seymour really is about is learning. So people talk about computing and people talk about math. But with Seymour you are just surrounded by this joy of being able to learn.

So it's not just math and things. It's things like, he taught himself when we were there how to work on the bongo board. There's some old AI Lab historic tapes of Seymour doing the bongo board. And what he loved was to explain to people how you thought about how to balance on the bongo board. He also did it with juggling. So he sort of had a tremendous-- teaching people to juggle. And his idea was that you could think about that. There was a method for thinking about how you juggle. So this thing that looked like a very complicated thing that a circus person was doing, it almost magic, you could actually teach it.

And everything he did was infused with that. I remember going to visit him in Maine. And he wrote this map. And the map, you could tell that he enjoyed writing the directions on this map in a way that people would understand them. And of course that went over into math and science. So he had exactly the same philosophy that learning about math and science isn't magic. There's a deliberate, in his way, procedural way of thinking about knowledge and about learning. Another critical part of that is the notion of debugging. So there's this idea that you do something, and you make a mistake. And he would always totally emphasize that mistakes are wonderful. Bugs are just great because that allows you to go back and analyze and think about stuff.

And all of that influence went not only in my own work in education, but also in the computer science work that Gerry Sussman and I did. There's this thing called the procedural explanation of knowledge. One critical thing in teaching people ideas, whether in computer science or anything, is to give them names and to use those names and things. And that goes right through all of our work in the Scheme programming language and our work in making MIT's first computer science course. It goes absolutely, directly back to Seymour.

INTERVIEWER: Were there other key mentors or professors that influenced your development, your intellectual or your professional development as a graduate student.

ABELSON: Oh well, sure. I mean I wasn't a graduate student in computer science. I didn't work for Seymour as a research instructor. My research was actually in mathematics under Dennis Sullivan, who is a very, very famous topologist, who was in the math department at MIT in those days. And again, he was a little bit different. Dennis was one of these mathematicians who's just magic. You look at this thing and you say how could this guy possibly have ever thought about this stuff? And he was just an enormous influence about-- just sort of this intellectual excellence where the stuff springs from nowhere. And you're just inspired that way.

INTERVIEWER: So tell me a little bit more about your professional path from mathematics to computer science and from getting a PhD at MIT to being a professor at MIT?

ABELSON: Oh, sure. Well when I got a PhD, I was in the math department doing topology. Got an appointment as an instructor. But then because I was working in Logo, MIT started something in the mid-'70s called the Division for Study in Research and Education. And it had a couple of key pieces to it. There were three leaders. One was Ben Snyder, who wrote a very famous book at MIT called The Hidden Curriculum, which is all about how surviving at MIT is a process of constant triage. And students at MIT have to learn that. One's Don Schon, who was in the Sloan School, who was very famous for his theories of knowledge or ideas of espoused knowledge and how you do things. And the third one which was Seymour doing Logo. So I became part of the Division for Study in Research and Education. Got a joint appointment in math and DSRE. And that's where we sort of really started to do Logo work as sort of real MIT research kind of stuff.

But then eventually what happens, MIT decided to close down DSRE. I was sort of too junior then to understand or even be aware of the politics behind that. But in any case, MIT decided to close it down. And then I moved into the electrical engineering computer science department, kind of on the basis of the Logo work and the basis of the fact that I was doing distributed computing. And that's sort of where I've been sitting.

INTERVIEWER: So tell me a bit more about the early Logo work and your work in distributed computing and how that has sort of influenced your research path since?

ABELSON: Well, I don't actually think about Logo and distributed computing together very much.

INTERVIEWER: Well, we can treat it separately I suppose.

ABELSON: Other than they kind of both end up being a vision of computing and being formal about it and things. The Logo work was as I said, magic. It is really hard right now to go back. We're talking a dozen years before the personal computer.

What we used to say in Logo is oh my god, if a computer could just cost $10,000, that would revolutionize all of education. Because what we were doing is these were million machines, a million dollars in 1968 dollars. They were sitting in the Artificial Intelligence Lab, doing the world's very best, fanciest research an artificial intelligence. And we were hooking them up over phone lines to be used by sixth graders in Lexington, Massachusetts. And just a totally outrageous thing. One my first projects as a graduate student was working on a remote display that you could hook up that could sit in Lexington and move over a phone line. And so students could write what was effectively the first Logo turtle and draw designs on this graphics display. So that was like my first job as a graduate student, implementing that.

But it was so off scale. Just the very idea that a computer would be something that a kid could use. We take it so it for granted now that it's almost impossible to put yourself in that mindset. So most people thought we were crazy, because I mean what a weird thing. Why could a computer ever be something that would be relevant to a kid, much less programming it or anything? And it was a mixture of what happened in the AI Lab. Every couple of weeks somebody would run in off the street saying, I want to see the robots. You're making robots. I want to come see them. And AI Lab was just a wonderful mixture of that.

So it was run by Seymour Papert and of course Marvin Minsky, who are just two completely brilliant people with I think a real sense of faith, in that if you're surrounded by a lot of smart people and you give them a lot of freedom and your job is really to inspire people, it will do great things. And the AI Lab was just infused with that. There was just a tremendous sense of community among the AI researchers, which was partly true because it was pretty small. One of the things that's happened more recently is that merged with the Lab for Computer Science. And it doesn't have that same real sense of community and flexibility.

But the original AI Lab, there have been a much of books written about it, all sorts of things emerged from that. The other thing that emerged from that, that maybe you want to talk about later is Richard Solomon and the whole free software movement, which these days is better known as the Open Source Movement. But that was a thing-- I mean in a deep sense it had to do with computer science, but not obviously. But it's a whole kind of philosophy that grew out of this sense that technology is fundamentally a data community and fundamentally about a whole bunch of inspired people sharing and building on each other's work. And that of course goes way past software into things like Creative Commons and lots of the stuff that we see now. But it came out of that germ of sharing around a technology community.

INTERVIEWER: Yeah, I want to definitely in a little bit explore some of these broader ideas in their current form. But let's talk a bit about that culture of sharing and of open networks within MIT at that point. I mean I've read that obviously there's a whole sort of different landscape of security and passwords or lack thereof. But just describe the environment a bit more at that time and how obviously it had effects later.

ABELSON: Well remember for thinking about now, the internet-- it wasn't an internet then-- the ARPANET was a small place. And it really was a very small, very, very homogeneous place.

I like to tell people I've had my email address since I kind of knew all of the people that I could have send email to. I mean it's been written a lot that people who originally designed the internet said they just sort of did not design the internet architecture to think that there would be bad guys out there. Because they knew all the people. All the people were like them. And that reflected over in passwords. I remember there was one sort of dark day when the people from ARPA said, you need to put passwords on your accounts because they're on the network. And for security, you need to do that. And a whole bunch of us were just convinced this was a horrible idea. I mean what a terrible idea that you'd have to put a password in an account and people could get into it.

I remember Richard Stallman reacted to this by making his password RMS, which was initials. So just jumping ahead a little bit, if you move about 10, 12 years later, my daughter, when she was about oh 14, introduced me to what became known as MUDS or multiplayer games on the internet. I had no idea this thing existed. And I was introduced to this by my, what she then, she probably was maybe about 13. And I come home and she says daddy, do you know there's this thing on the computer that all kids in the whole world can use? And I say, gosh no. And she says, I'll show you. And she says you have to dial this phone number. So she starts dialing this phone number. And I say, oh my god, that's my lab at MIT. That's the terminus server to my lab at MIT. And she says yeah, you have this thing. And it types out this funny star or something. And now I'm supposed to type R-M-S. And I say, oh my god, you're typing Stallman's password to get into the system.

And she goes from that and starts up playing this game. And it was great. She had no idea that this was MIT. She had no idea what RMS meant. But there were kids all over the world who were doing this. And that's a little bit of the spirit of sharing and openness and delight that comes from that. That's just really hard to do now. Because again, the network is a much less homogeneous place than it used to be. But that's a little bit of what that AI Lab kind of spirit I remember, just to give you another little anecdote about that, that's fun.

I don't know if you've interviewed Joel Moses?

INTERVIEWER: Not personally, but yeah.

ABELSON: So Joel Moses at that time was the head of the symbolic mathematics project. That's really the great thing that he did, before becoming provost, sort of inventing symbolic mathematics. And there was a whole big dispute between him and one of his ex-students who had gone to Berkeley. And there was a dispute about who should get access to this Maxima program and should it be licensed and not licensed. And there was a fairly heated email exchange that went between Joel and Rich Fateman. And of course, everyone could read everybody else's email at that time. And they decided that people in the lab were wasting so much time reading Joel's email that they simply translated all of Joel's email to message of the day. So you all saw it when you logged on. And that's a little bit again, the sense of playfulness and openness and we're all part of this big community and the computer system is a way of empowering that.

And I do believe that had a whole lot to do especially with the development of time sharing. In a sense sort of Bob Fano and Corby Corbato, when they started the time sharing project, it was sort of about technology. But pretty soon they realized really what they'd done is create a new kind of community. And to their credit they recognized that very, very early and pushed the sense of gee, there's a thing like community. There is an information community. MIT was one the first places that really did that in a conscious way. And again, this all comes out of the same kind of culture of openness.

INTERVIEWER: Yeah, we'll return to the whole question of open networks, Open Source a bit later. I want to turn a little bit in a different direction and ask a bit about your interest and professional development of teaching computer science, of the development of 6.001 and also your work with Gerald Sussman and the book.

And maybe, I don't know if this is a productive way of leading into it, but it's a question I'm really curious about. You mentioned very early on that when you were growing up, you were bad at arithmetic, but found out later that you were good at math. And I think that's very fascinating. How does one think about math in a way that makes that true? I mean they're impossible that could be bad at arithmetic, but good at math.

ABELSON: I don't know. I don't view arithmetic as having very much to do with math in fact. And I think there are a lot of people in kind of the same boat.

Again, there's theories that people who go into these developmental fields are doing it in some sense to make up for their own deficits, to try them. So I sort of view it kind of that way. But again, I've always been interested in puzzling out things and explaining things. I wasn't very much interested in teaching until I, gosh, until I got to MIT in fact. But it was mostly explaining things to myself. Being in a couple of very good summer high school programs, one at University of Notre Dame, which has since moved to Ohio State, which had a long, long, long track record of inspiring people in mathematics. One at Cornell, in physics. So that's another kind of educational link that I was exposed to.

And later, my colleague Andy diSessa and I ran that kind of summer high school program at MIT around computer science and Logo and teaching things. So I've been interested in education kind of from that perspective.

I've also played around with computers in education. One of the things I did in the math department was I was teaching linear algebra. And I think just for the heck of it, I did what must have been one of the very first MIT computer exercises where students used the Multics computer system in those days to do some exercise at the end of the day about matrices. Now I didn't know what I was doing. And it wasn't particularly good and wasn't particularly successful. But it was kind of foray into saying, gosh these MIT computers might actually get used in courses. Not much of that happened for a long time after that though.

INTERVIEWER: Thinking about teaching computer science specifically, can you talk a bit about the Structure and Interpretation of Computer Programs and what that grew out of and what the sort of fundamental concepts are behind your approach?

ABELSON: So Gerry Sussman and I, I think were really privileged in that we wrote a book that really was able to draw on that whole of a 20-year history of how MIT and that whole AI culture thought about computing. And it goes directly back to again Seymour Papert and Marvin Minsky. And it really has to do with-- what's the phrase we used-- a computer program is a way of expressing ideas and communicating ideas, and only incidentally about getting a machine to do stuff. That is straight out of-- I don't think Seymour used those words, but that's straight out of the Logo philosophy of we're teaching you-- like I talked about juggling before. This really is, what's the procedure for how you juggle? And critical in being able to do that is having the right set of words and having the right descriptive framework for saying things like that. And so the key idea in Structure and Interpretation is what we ended up calling the linguistic approach to programming design, which says if I would like to do some very complicated thing, one of the critical tasks is picking the right descriptive framework for saying that.

So there is a complete blend. I think somewhere in that book we wrote that when you make a complicated program, you're actually designing a special purpose computer language to solve a problem. And it comes through in lots and lots of ways. So it goes back I think to the AI culture, which has the property that you don't quite know what problem you're solving because these things are really complicated. It's not as if someone says compute the total payment on this mortgage where you know exactly what you're going to do. Rather it says, you want to be mucking around in this area to solve this problem and lots and lots and lots and lots of similar problems.

And that permeates both the design of the book and the actual programming technology that people use. So the style of programming languages that go back to LISP is really sort of muddier. It's not a very precise kind of thing. You make ways of programming things where you're solving not just one problem but a whole family of problems in that.

And again, it goes back to very, very technical things, like LISP is a very good language for creating other computer languages. Because when you're programming in LISP, you actually have the mentality of gosh, I'm actually creating some special purpose language to solve not just one problem, but lots of them. And that's the critical idea in Structure and Interpretation.

I mean we can't take credit for that. We were just blessed that there was a whole tradition of 20 years of really deep thinking at MIT and a couple of other similar places. And we sort of got to be the people who got to put that in the book. And it's had enormous influence, but it's not because of us, it's because the tremendous power of those ideas and the communities that they came from.

INTERVIEWER: So moving beyond those communities that the book and that thought process came out of, did you encounter resistance, maybe that's not the right word, maybe misunderstanding-- people who were focused on the syntax?

ABELSON: So we published this with MIT Press. And MIT Press sent it out for reviews. And the first review came back. And I'll never forget. It said, quote, the question is, is in MIT Press publishing this book going to advance or set back computer science? So there was a lot of resistance to this because it's just very different. And there's a lot of people who really don't believe in that approach, who really do think it's the wrong thing. Because again, it's this notion of you don't actually focus on solving this problem very, very, very precisely.

There's a whole movement in computer science, a tremendously respectable one, and it says a program is like the proof of a mathematical theorem. And there are people who do prove things about programs. That in some sense is the mainstream almost. Whereas we're saying well, a program is sort of like a language. Like you can say lots of things. And I don't imagine I'm going to prove everything I'm going to say. It's just that the great thing is I get to see all these different things. So we've had a lot of influence, but I wouldn't even say that we're mainstream, even now.

INTERVIEWER: From my limited experience when I think back to my computer programming classes, it was more like the mathematical theorem approach. Yeah.

ABELSON: Exactly right. It is.

INTERVIEWER: So talk about 6.001 and the development of that class, which is obviously closely related to the book, right?

ABELSON: Gosh. So the other person who really deserves enormous credit for this is Bob Fano. There was a tradition of, what was the first MIT computer science course? There was one taught by Mike Dertouzos and Steve Ward for awhile. But when I got involved with this, Bob Fano was teaching this. It was called 6.035. And what was it called? I think it might have been called Structure and Interpretation of Computer Languages. And it was very focused on, here's this kind of language. The language then was ALGOL people used. And then it did a little bit of LISP. And it was kind of an okay course. Mostly famous because it was the first course.

But Fano did say the key thing. So I talked about what you're doing is making a language and how you make languages to make languages. The technical term for that is an interpreter. So an interpreter is something that effectively takes the description of a language and lets the computer effectively speak that language. And Fano was very-- I mean he would say this in his lectures. I don't think any of the students got it. I didn't get it until like the third time through, where he sort of said the really, really key thing is that you build up complexity by constructing an interpreter. And again, that's the core idea of 6.001, said in a slightly different way. So he had an enormous, enormous influence on us, but also more directly.

So about this time, there was a tradition in the electrical engineering computer science department of when the department did the transition from kind of power engineering to signals engineering in the late '50s, led by people like Ernie Guillemin and those sort legends at MIT, there was a very, very influential series of textbooks that came out, I think in the late '50s, early '60s, that almost sort of set the academic marker for the field of electrical engineering.

So when we were there in the in the late '70s and very beginning of the '80s, Joel Moses and Gerry Wilson wanted to do that again. So there was this notion that there would be a series of new courses and new books that the department would put out. And then what was going to be the computing one? So Fano basically backed that Gerry and I would do the computing one and do the computing course. And again, there was a lot of a lot of controversy there because people didn't believe it. Fano was sort of the power in the department who really made it possible for us to do that. Gerry and I sort of happened into this. So we were sort of lucky. And just very, very fortunate that there were some smart people backing us.

INTERVIEWER: How has introductory curriculum in ECS changed since that time? I mean clearly it's--

ABELSON: Oh, it's changed a lot.

INTERVIEWER: Yeah.

ABELSON: I don't know. People keep using the phrase bittersweet. It's changed in style. So what happened in the department is the department kind of decided that the way students get introduced into the department should become more horizontal and less vertical, by which I mean when you go into the department before the new curriculum, there were a bunch of very intellectually deep courses. So 6.001 really is a deep course in software engineering. And then the idea was well, you shouldn't actually do that because if students want to think about what they want to major in, they ought to get an experience that's a lot more broad. So they should learn not only about software, but something about circuits and something about signal processing and all that. So 6.001 sort of didn't fit into that world.

I sort was on the committee. I had this funny feeling. I sort of loved the course. But I also had this feeling that the course needed to change. And both Gerry and I felt that as long as we were part of it, it wasn't going to change. And in fact we consciously got out of it. We had an official, what you call passing of the baton ceremony in one of the lectures where Bob Fano was there as the originator of this, who let us do 6.001. And then Eric Grimson and Duane Boning were there. They were going to take it over. And we kind of officially passed the baton. And we said we now did this in front of 350 student witnesses. We are never going to teach this course again. The department can't ask us. But then what happened is even after we got out of it, the course didn't change as much. And we can flatter ourselves and say well, the ideas were just really great. But sometimes you just need a kick in the pants to go someplace else.

So this new department curriculum structure turned out to be the kick in the pants. And so I was part of the group, along with Leslie Kaelbling and Jacob White and Tomas Lozano-Perez, who sort of designed sort of the new course, which consciously was not about 6.001. But a lot of the biggest themes of 6.001 about extraction and modularity and the way you express things, they carry over into the new course. But the new course is also doing robotics. And it's thinking that signal processing. And there's even some stuff about circuits. So it's not a deep course in that sense.

And there's a lot of people who have come through MIT-- I don't want to take credit for 6.001 and that, but I want to say it was a very deep experience in how you think of software-- who criticized what we did. And they say they really miss it. They talk about their own intellectual development as programmers. It was just profoundly influenced by that kind of very-- we would sit in that class and talk about the philosophy sometimes. The new course does not do that in the same way.

On the other hand, you have to sort of ask what's the population you're serving? And the fact that you talk to the minority of students who really, really glommed on to our way of thinking and they say, gee, we really wish MIT were still doing that. I don't know what the right way of serving the most students are. But in any case we're just doing something rather different in motivation now.

INTERVIEWER: So it's broader, but less deep.

ABELSON: And very consciously less deep.

INTERVIEWER: Right. What are MIT students like? I mean how do they respond to the deep dive on the one hand or the sort of broad fire hose on the other of having to deal with all of these different things?

ABELSON: I think you can't generalize. Right. There are certainly some, I'm going to say kind of the mathematical ones. So even though it's not mathematics, that kind of deep thing is a math kind of experience. And they really would rather have something that's more like the original 6.001. But then there are a whole bunch of others who just want to-- I just want to understand the breath of this field I'm getting into. I don't want to just because I'm going to major in electrical engineering, have to take a whole semester of this other stuff, most of which is kind of irrelevant to me.

So again, I don't have a good sense of how the students are feeling about the population it's serving. Because we did this other thing at the same time, which as far as students I think has a much bigger impact. So there's a whole other stream which we haven't talked about at MIT and active learning, which comes from the stuff I've been doing with the Educational Technology Council.

But we made a radical, radical change in the style of the course, which was removing most of the lectures. So in the new course, the main thing you do with your time is you work on a project. And there's a big giant open lab. And the entire teaching staff of the course is present in this open lab. And there's kind of a pyramidal structure where there are undergraduate laboratory assistants, and each group of them is kind of supervised by a graduate student TA, and each group of graduate student TA's are supervised by the actual course faculty. And we're all sort of mixing it up in the course and sitting there talking to individual students. So rather than sitting in lecture, it's this very kind of open, interactive project-based thing.

And that's such an enormous change in style, that when you kind of ask the students well, do you like it more, how you react to it more, that's got to dominate what we were doing in terms of content.

INTERVIEWER: Yeah. I want to actually ask a whole set of questions on educational technology and the developments you've been involved in. Before turning to that, at the risk of asking to generalize again about MIT students, is there any way of talking about how or whether MIT students have changed over the course of your career?

ABELSON: Oh, god. You get the feeling that about every three years there's a new generation. So the main way MIT students have changed I think is that they're more docile, as much as I hate to say. Marilee Jones, who was admissions director, actually thought about this a lot.

Remember in the '60s, MIT took half of its applicants. And now it takes what, 10 percent, under 10 percent? And the students who end up getting into MIT are coming from a way more competitive environment. And in general, and in tremendous exceptions, they have been successful by kind of listening to their parents and listening to adults. And that's led to a tendency that students I think have too much respect for what adults or like faculty or something are saying. So there's really a sense of just kind of well they sort of know better and they have good things. And they go along it. And there's not as much a sense of really challenging the system.

Now I don't know whether that's a generational thing or whether it's this particular kind of influence at the top schools like MIT. When's the last time you saw students take over a building at MIT? Like I said, when I showed up at MIT when I saw that happen, I said what incredibly an healthy thing to happen. And now we get upset if students crack into an elevator. So there's really this sense of the students becoming-- I mean docile may be not quite-- the wrong word-- but sort of too much respect for what the Institution is doing. And I sort of wish that would change. So that's sort of the big change that I see.

INTERVIEWER: What price do we pay if the students are more sort of quick to fall in line or to do what's expected, rather than to push boundaries, I mean if any?