Oral-History:Adele Goldberg

From ETHW

About Adele Goldberg

Adele Goldberg was born in Cleveland, Ohio in 1945. Her family lived there until she was eleven, when they moved to Chicago, Illinois. Goldberg had an early interest in math and majored in the subject at the University of Michigan. She continued her academic career at the University of Chicago, earning a Ph.D. in information science in 1973. Goldberg began working as a researcher at Xerox's Palo Alto Research Center (PARC) that summer. There she was part of the team that developed the object-oriented programming language Smalltalk. She became the manager of the System Concepts Laboratory in 1979. Committed to bringing Smalltalk to a wider audience, in 1988 Goldberg cofounded the spin-out company ParcPlace Systems. Goldberg served as CEO and Chairman of the company, which created development tools for Smalltalk-based applications. After leaving ParcPlace Systems in 1995, Goldberg established a consultancy. Among her many consulting projects, she has designed systems to author and deliver online courses meant to enhance equity and high achievement in secondary school math and science.

Goldberg was president of the Association for Computing Machinery (ACM) from 1984 to 1986. She won ACM's 1987 Software Systems Award with Alan Kay and Dan Ingalls and PC Magazine's 1990 Lifetime Achievement Award.

In this interview, Goldberg discusses her educational and work history. She recalls her experiences as a student at the University of Michigan and at the University of Chicago. Next, she covers her stretch at Xerox PARC, sharing her views on the work environment. Here she speaks at length about her work on Smalltalk, including her leading role in its commercialization. Goldberg is candid about the challenges she faced in forming and running spin-out company ParcPlace Systems. In addition, she discusses her two-year tenure as President of ACM. Finally, Goldberg offers advice for young women who are considering a career in computing.

About the Interview

ADELE GOLDBERG: An Interview Conducted by Janet Abbate for the IEEE History Center, 3 July 2002.

Interview #591 for the IEEE History Center, The Institute of Electrical and Electronics Engineers, Inc.

Copyright Statement

This manuscript is being made available for research purposes only. All literary rights in the manuscript, including the right to publish, are reserved to the IEEE History Center. No part of the manuscript may be quoted for publication without the written permission of the Director of IEEE History Center.

Request for permission to quote for publication should be addressed to the IEEE History Center Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.

It is recommended that this oral history be cited as follows:

Adele Goldberg, an oral history conducted in 2002 by Janet Abbate, IEEE History Center, Piscataway, NJ, USA.

Interview

INTERVIEW: Adele Goldberg
INTERVIEWER: Janet Abbate
DATE: 3 July 2002
PLACE: Goldberg’s home in Palo Alto, California

[Notes courtesy of interviewer Janet Abbate]

Growing Up: Family and Education

Abbate:

It’s July 3rd, 2002, and I’m speaking with Adele Goldberg.

I always start at the beginning, so can you tell me when you were born and where you grew up?

Goldberg:

I was born in Cleveland, Ohio, in 1945, and I lived in Cleveland till I was eleven, when we moved to Chicago. We lived in the North side of Chicago: in the city proper, not in the suburbs, but in the North side. I was there until I went to college.

Abbate:

What did your parents do for a living?

Goldberg:

My father was what would be considered an industrial engineer; he ran production for various companies. We moved to Chicago because he changed jobs. His father died when he was three, so he basically finished high school and then went to work.

My mother’s college degree was in math, and when she went to get a job there were no jobs in the schools, so she ended up working in a high school—but I believe in a secretarial job and not actually teaching. Then when I was born: I’m a twin, so I’m the third in the family with three girls; I had a sister a year older, and then my twin; so my mother stopped working when there were three little ones in the house.

Abbate:

That sounds like a handful!

Goldberg:

Yes, that’s a handful.

Abbate:

So both your parents had kind of technical backgrounds?

Goldberg:

Not really. I don’t think they would think that. There was no table conversation about math and science. I think my parents were always bewildered by what I was up to.

Abbate:

Really?

Goldberg:

Yes. You know, I can only contrast that to what I know went on here with my girls, where there was a lot of table talk about what we were doing, either in research or in business. There was very little table talk on any of those levels—short of, if I remember correctly, my father commenting on management doing a better job of taking care of its people than the unions [laughs]. In those particular examples he gave, he was right; but it’s not always the case!

Abbate:

Were you interested in math or science from an early age?

Goldberg:

Absolutely! It was math more than any particular science, and I’m not sure why, other than I thought it was great fun doing problem solving, and I was good at it. When you’re good at something you enjoy it. Even as a youngster, but certainly in high school, I quickly discovered that I enjoyed the math, and I focused on math. I was very fortunate to be in a high school in Chicago where they had two really fine teachers, one of whom was a Ph.D. They were good teachers—having a Ph.D. doesn’t necessarily make you a good teacher—and they had a special track for a small group of kids. I’m trying to remember if I was the one girl. Maybe there were a couple of girls in the class, but mostly guys. But because this one teacher—it was Dr. Claire—really encouraged the good students, and he was part of an NSF summer program for high school kids that was at IIT; I remember going to one of those summer programs. So most of the encouragement I got was really from school, from the school teachers.

Abbate:

Sorry, what’s IIT?

Goldberg:

Illinois Institute of Technology.

Abbate:

Okay.

Did your sisters like math, too, or just you?

Goldberg:

Just me! Yes. We’re very different. You know, my sisters were bright, but they just weren’t as interested, and I think they were more social than I was, frankly. [laughs] I don’t remember. I mean, I was a twin; I always had a pal, I always had a friend; so it wasn’t like being in classes with just guys meant you didn’t have girlfriends. I had my twin sister.

Abbate:

You went to just a regular public school?

Goldberg:

I went to regular public schools. In Cleveland—by the time we got to Chicago, all of our curriculum was ahead of Chicago; so I remember the sixth grade that I spent in Chicago was mostly interesting because I sat in the hall most of the time, doing nothing, because the teacher thought I was sassy. It was just a different style of school. I got my work done faster, because I had done it already; I’d sit there quietly, and the teacher would ask me, “Are you done?” and I’d say “Yes, ma’am.” And then I quickly learned you didn’t say “Yes, ma’am.” It was too polite; that was considered sassy! [laughs] “Out the door!” Except that out the door was more interesting than sitting in the class. It’s funny what you do and don’t remember about school. It’s so long ago. That’s what I remember.

Abbate:

Did your parents encourage you to have a career? Were you expected to be supporting yourself?

Goldberg:

I don’t mean this in a disparaging way for my parents, but I don’t recall them encouraging or discouraging. What I recall was that I just did whatever I did, and they never stopped me.

My twin sister got married when she was nineteen—just before our 20th birthday—and she chose not to finish school. She’s been married for thirty-seven years, perfectly happy, beautiful family—but I remember being very angry, thinking that wasn’t smart! But, you know, it works out differently for different people. We’re all very different. My parents didn’t have any particular mission that they were on, nor do I think that I do with my kids—that there’s a mission, other than supporting what their nose is taking them to.

Undergraduate Career at U of Michigan

Abbate:

Did you go to University of Chicago for college, or just grad school?

Goldberg:

I went to undergraduate at the University of Michigan in Ann Arbor. I started school in ‘63. I don’t know what they do now, but in those days, you could apply your junior year in high school; and of the top ten kids in my high school class, nine of us went to the University of Michigan on early admission. For some reason, probably having to do with the reputation of the football and basketball teams in those days, Michigan, for a Midwestern large university, was the most popular. My sister went to Wisconsin, but she just went there for two years before getting married.

I didn’t like Ann Arbor.

Abbate:

Really? Everyone says it’s so nice.

Goldberg:

For me it was an enormous change. It was the first time I was away from my twin sister, so that was socially a shocker. You know, I had a group of friends who were not the same friends that she had, and so it wasn’t like everything swirled around her, but there is a certain security that you have, where you could go pursue things quite comfortably, knowing perfectly well that she was there. I was very active in extracurricular in high school; I was an officer in the Student Council, did a lot of stuff that I haven’t got a clue why I was doing it; but I felt fearless about it. What was the difference? Because she was always there. She’s really much more the stronger personality.

So I remember being quite an unhappy person, my initial days in Ann Arbor, which I’m sure colored everything. But it was just so big; and it rained all the time! [laughs] You know, things like that that I just didn’t like. I guess the largeness also provided options—socially, you could find different groups—but I was still not comfortable. I found the food terrible; I gained a lot of weight the first year. You know, I look back and think, “I could have done it better.”

I stayed three years, and then took off for Europe and missed a half a year, and then finished up. Then I went to graduate school at the University of Chicago.

Abbate:

At Michigan, did you major in math?

Goldberg:

I majored in math, and they had some computer classes. In perspective, think of the timing. My freshman year, Kennedy was killed. Those were not very good years in this country, and the war didn’t help; as the war was heating up, it changed the social scene. I would say that doing math and science was almost a way to hide socially. I wasn’t doing it purposely, but if I think about a lot of the students, you really did distance yourself a little bit from more of the social events.

So I did math and some computing. Let’s see, Bernie Galler was there at the time, and Bruce Arden. Computers were taught from the perspective of numerical analysis and programming various number-crunching algorithms. They were involved a bit, I think in my senior year, in some of the IBM 360 work, but most of the computing was done with a deck of cards handed to someone at a glass window.

There was a gentleman from my high school a year ahead of me: Richard Weiland. Very smart; he was the T.A., I remember, for one class. The whole style of teaching was to give you a problem; to give you sample data, but not the full data set; and then you had a certain time at which your deck of cards were handed in. At that point, Richard would run the deck of cards with his real test data, and if it worked, it worked; if not, not. I don’t remember ever finishing anything; ever really completing anything. I remember thinking it was just not the way to teach people about doing quality programming, because you never finished anything. Some of this was just the large-school situation: they’re big lecture halls with small lecture groups.

Abbate:

This was within a math class that you were using the computer?

Goldberg:

No, they had computer classes. They taught Michigan Algorithm Decoder—MAD—as the first programming language. But, as I said earlier to you, I like problem-solving, and it was clear that my math interest was probably no different than why I do crossword puzzles. Those are word games, and these are number games. So I thought programming was interesting.

Summer Work at IBM

Goldberg:

After my junior year, I had arranged to go to school in Germany. I had arranged it with my German teacher at Michigan, who actually failed to keep his promise to arrange for housing—because I wasn’t on a Junior Year Abroad program; I was just going to go for the summer and do an immersion and learn German. When I got there, his friend picked me up, but there was no housing. I spent a couple of weeks looking and didn’t find anything, so instead, like all good Americans with their Junior Year Abroad, I called home and got my Eurail pass. I knew where there were friends, and I kept meeting up with kids I knew from school—sorority sisters and other people—and traveled around for months. Everywhere I went there was this huge building with an IBM sign on it, and I thought, “Oh, I get it. There’s this company that does computers. I liked my computer class; it was fun; and you could live all over the world if you worked for them.” That was what I saw as I traveled around Europe. Then I went into the Middle East, and got sick in the sun and came home. I missed the fall term, but I only had ten hours of math classes left to graduate, which is all I took.

My dad’s company had just bought an IBM 6600—I think that’s correct, even though when you hear “6600” you may hear a CDC machine—I think it was the last of the unit record equipment, the last really big one. So he knew the Sales guy and Systems Engineer out of the local Chicago IBM office. When I came home and told him that I thought that computing was perhaps a neat thing to do with my math degree, he suggested I contact IBM. I don’t know how he arranged it, but they agreed that I would be a clerk for the summer [of 1966] in what IBM called an “Installation Center,” which I think is an interesting idea. You know what a unit record machine is? A unit record machine is basically an accounting machine where everything is kept on tape or on punch cards, but you program a physical board that’s one-and-a-half or two by three feet, and it has holes in it just like it was a telephone operator switchboard. Sets of holes serve as temporary storage locations, or registers. You run physical wires from register to register to transfer the data or process the data. That was “programming”!

Abbate:

I don’t think I did know that.

Goldberg:

These were to me, physically, a wonderful way to touch and feel and say, “Ah! Information bits are transferring from one place to another, and you’re manipulating it.” Of course, you’d do fairly simple accounting procedures, but the machines would get kind of complicated; they would go from maybe four registers to sixteen. A sixteen-register is the biggest 407 that I think I touched.

So what happened was: IBM had these Installation Centers where they would sell these accounting machines to small businesses; it was mostly for record keeping. Instead of doing the programming for the customers, the customers would do their own, going to class; but if they needed help, they’d come to these Centers, and they could sit with a Systems Engineer and get help. I was in the adjacent Business Office, and there really wasn’t very much clerking work to do, because it was summer. They also did something else fairly smart: they had all these programmed instruction manuals, which I’d never seen before, where you can just sit and teach yourself. So when I had spare time, I could go into what was a reasonably empty Installation Center in the summer and pull all these books off the shelf. There were machines around, and I’d go through the programmed instruction and teach myself. I was minding my own business, because I’m relatively quiet and not a big socializer in the offices; just wiring up and teaching myself stuff. And they’d sometimes have classes where the Systems Engineers were being trained or the Sales Engineers were being trained. Those were the days of IBM suits and white shirts: I saw the Sales guys get sent home if they came with a blue shirt, even though blue was “in.” They didn’t even keep shirts on the shelves for them; they had to go home and feel the pain. [both laugh] I thought, “Hmm, this is interesting!” Socially, it was pretty strange.

But I guess it was noticed that I had figured this stuff out, because at one point, probably the highlight of my summer—mind you, it was the summer that I had cut all my pigtails off in hopes of looking a little older—I was asked to go out and fix a machine on the West Side of Chicago. It was an interesting experience on two levels. They had a machine where the board—we didn’t call them “programs,” we called them “the board”—didn’t work. The Sales or Systems Engineer introduced me to this West-Side-of-Chicago blue-collar manager—which I was used to; I mean, my father worked in that area—and said that I would fix it, and you could just see the disbelief in his face! [both laugh] If you think about it, I was just a kid, and in those days people didn’t think about kids being programmers in any form or doing computing. Computing was unusual enough. So I remember sitting there very carefully. What happens with these boards is, once someone decides they work, they replace each of the wires that plug in and out with permanent wires that have little hooks in them, so you can’t accidentally yank wires out and destroy the program. So this was a permanently wired, sixteen-register, loaded 407; it was the biggest it could get; and I remember having a grand old time following all the wires—i.e., following all the data as it moves through the registers—and concluding that it worked just fine, and therefore it was a hardware problem. So the first thing I learned was that not everyone thought young women should be running around doing this, and the second one was that you didn’t just call up the Field Engineer or hardware person and tell them it was their problem! [laughs] They didn’t really believe me, and it took a couple of weeks for them to come check—but it was true: they had installed a twelve-register instead of a sixteen-register machine. Phew—I was correct! [laughs]

It was this adventure in “Okay, you know how to do this, and this is fun; but now what?” Because I really didn’t know anything about computers; I just knew this little game of wiring up. I had two choices: I could finish school and join IBM and go into the training programs, which were excellent; or I could apply to graduate school specifically for vocational training, which is a very different reason to go to graduate school.

Abbate:

I’m not sure what you mean by “vocational training.”

Goldberg:

It means that I didn’t go there to learn how to be an academic researcher. I went there to learn some very specific knowledge and skills in a burgeoning field. I didn’t go for the credentials of the Master’s; I went for the knowledge, feeling I didn’t know enough. I could look at the ads in the paper and see all the code words, and when I look back now, I think, “I can’t believe I didn’t know what that stuff means.” But it’s another vocabulary, and I didn’t know it, and I felt I needed to know it.

I opted to go to Chicago, which was more applied math: one, because they accepted me with scholarship; and two, because I was interested in going back to the city and seeing if I was going to go back and live in Chicago, because that’s where I had family; and three, because the only other place I applied was MIT, and I got turned down. So it was either go to school in Chicago or not.

Abbate:

Had you asked IBM about being in the training program?

Goldberg:

No, I just knew because I had worked there. I could see what was available; I knew generally what they were doing, although I didn’t know the details.

Abbate:

But you could have done it if you’d wanted to?

Goldberg:

I’m sure if I’d applied, I could have done it. What I actually ended up doing was, I went back to Michigan in the spring for my final semester and looked around for a programming job—because besides going to school to learn, I thought, “Well, I can teach myself a certain amount, so let’s see if there’s something on campus where I’d have a learning opportunity.” That was a transitional experience, because the job I found was at the Center for Research on Learning and Teaching, which was headed up by a wonderful man named Karl Zinn. This was the beginnings of computer-aided instruction, and what Karl had were grants on IBM 1500s running Coursewriter II; this is a language for offering courses. The reason I got the job wasn’t because I was a programmer, but because what they were programming were German classes, and I knew enough German to be the mixture of programming and content. I’d had at least one course in MAD; I didn’t know Coursewriter, but I knew how to wire! The IBM manager in Chicago had told me, “When somebody asks you if you can do something, you always say Yes: because you can; you just have to learn.” Of course, that’s an ambiguous question and it’s a dangerous one, and I see people in business do this all the time. “Can you do blah-blah-blah with your product?” And the answer is always Yes, to anybody technical. So the question should be, “Right this minute, do you know how?” But that isn’t what I was asked.

His name was Bill Lloyd. I remember when he told me he was being transferred by IBM to the strangest-sounding city I’d ever heard of—which of course is Poughkeepsie, but you know, I was just a Midwesterner and didn’t know from Poughkeepsie!—and I said, “Well, you don’t have to go!” And he said Oh yes, he had to go; and that’s when I learned that “IBM” stood for “I’ve Been Moved.” [laughs] I think that turned me off a little bit—not working for a big company, but for that particular style where you get moved around. I thought it was neat they had places everywhere in the world; I didn’t know that you didn’t get to choose where you went! [laughs]

So I went back, and I did my ten hours left of math, and the rest of the time I worked for Karl, and I both learned about this field of using technology in education and understood more about programming at that level. It was much nicer than punch cards through the window; I mean, I was sitting at a machine that was mine to interact with, and it was 1967! There were such machines that weren’t shared; there were these early ideas.

Then I spent the summer with some of my college friends, and we decided that, having seen “the world”—meaning Europe and maybe the Middle East—maybe we should go see the United States. We rented a car out of Denver and traveled all over the national parks, which was wonderful, because when you grow up in the flatlands of the Midwest and the only family vacations you go on is on a turnpike between Chicago, Cleveland and New York to see family, you don’t exactly get a good impression of what is here in the United States. So that was a wonderful trip. We ended up in San Francisco in 1967, and one of my girlfriends, Beverly, and I stayed. She’s from Connecticut, so her inclination was to go find an insurance company to get a summer job; and I went marching into the San Francisco Business Office of IBM and told them—how did I say it? I said I was a Systems Engineer looking for a job, and that I had worked in the Chicago office. Now I figured, “It’s IBM. They have computers. Click-click-click, they’d find out that I was a fall clerk!” But in fact, they didn’t have their network of “who you are” in place—fortunately for me—so they said that they weren’t hiring until the fall, but they could use some help in the Installation Center in the summer. I said, “Piece of cake. No problem!” So I got a summer job—knowing full well that I was going to go to graduate school, so it wasn’t entirely honest of me; but they didn’t actually offer me a job anyway; it was just a summer job.

It was a very peculiar summer! We were living in Berkeley, and we were commuting into San Francisco, working in an office, when a large number of my high school and college friends were in the Haight-Ashbury, with a totally different mindset. I’ve always been a bit more conservative about my personal life, and I kind of missed that era—missed everything: the drugs, the music, whatever was going on! I think back now and think, “I really did miss it, and really should have gone and dealt with it!”

But it was an interesting summer, because I learned from some of the women who worked for IBM: one with a political science major, and one an older woman; I don’t really know how old she was, because I was in my early twenties, fresh graduate, and anyone ten years older than me probably looked old. But I remember she came to work every day in a grey suit, and I remember she thought that my manager had a say on my personal life, because she asked me what I was doing the Fourth of July, and I said “Oh, my roommate Beverly and I are hitchhiking to Yosemite.” We didn’t have a car, and we wanted to see Yosemite. And she said, “You don’t hitchhike. I’m telling the boss.” And I said, “Excuse me?” [laughs] That was kind of weird. But the woman in political science told me that in order to get ahead in IBM in those days, you needed them to put you through their various training courses. They had a wonderful training capability, and learning on their new machines allowed you to learn not only the machine, but also the principles and the basics. She wasn’t a math major, and they weren’t letting her take classes as a result; that was her opinion, of course. Now, I don’t know the truth about how good she was or wasn’t at whatever it was, but it was clear that there was some kind of a filtering, which I thought wasn’t very nice. There were just these two women in the office, and an English secretary who came in bright yellow dresses and shoes, very pregnant—and that was also a no-no, coming in these bright clothes. It was a corporate world eye-opener. It didn’t turn me off at all, I just needed to know what was going on and what your choices were; but I do think it stopped me from ever applying to IBM. I never did. I learned, summers later when I had friends from the University of Chicago who worked for IBM in the Research facility, here in the South Bay, that there is a different world: a much more casual world; a no-white-shirt research world. That was when I found that out, but not earlier.

That combination of getting to work twice for IBM and working at Michigan taught me that you don’t just go to school: you have to do work-study; you have to have jobs while you’re in school. My one advice when we’ve done various career counseling, and certainly with any young girl, was: “You need to go see what the people do. I don’t care if you’re washing bottles in the chemistry lab; if you think you’re going to be a chemist, you need to be there and see what they’re doing. You’re not qualified to do the chemistry work, but you are qualified at some level to be standing there and seeing what they’re doing and learning about it before you jump in with both feet.” For me, that paid off. I mean, it was a little late in my career for doing that; it might have been nice to do it earlier, in high school, before I went marching through years of college; but it was something that really stuck with me.

Comparing U of Michigan and U of Chicago: On the Number of Women

Abbate:

I want to get back to a comment you made earlier, before we started recording, about women at the University of Michigan and the Vietnam War. Were there a large number of women in the math classes?

Goldberg:

No! I had a sorority sister who was a math major, and she and I went to class together. We were the only women. You know, it depends on the year and precisely what you’re doing. At Michigan, at some point, the undergraduate math courses were the same as the graduate Master’s-level classes. I always thought that was pretty peculiar. She and I just would go to class, and there weren’t any other women in there.

I was a terrible student, because I had, in those days, an eidetic memory. Eidetic memory is a visual memory; if I see something, I could see it again. So if what they’re going to do is make you regurgitate proofs of theorems in the exam, I could close my eyes and turn the pages of a book and write it down. It didn’t mean I understood anything. So, I look back and think, “That wasn’t an enabler. That was a crippler, in terms of learning.”

A lot of people used to say, “Don’t you just hate being the only woman? You have to be so much better.” And I said, “Well, it kind of makes me work a little harder and be better.” I didn’t see it as a negative so much. Maybe it is, but I didn’t focus on that. I focused on: This is what I was learning, and it was interesting, and I enjoyed it!

When I went to Chicago, there were a lot of women. In fact, I was surprised by it. I was surprised by two things. One was the number of women, and the other was how much I loved a smaller school, where the faculty actually were accessible. Now, probably the difference was between graduate school and undergraduate. As an undergraduate at Michigan, I was in the Honors College, so classes were relatively small. We had fairly specialized curriculum options as Honors College students. University of Michigan is not the land grant school—Michigan State is—but the University, in order to get some of its grant money, had certain obligations. I don’t know if it’s still true, but in those days, the valedictorian of any high school class in the state of Michigan had to be admitted. My freshman year roommate was valedictorian from Houghton Lake, Michigan, three hundred miles due north from Ann Arbor; not the Upper Peninsula, but way up there. She managed to get all C’s, but they had these intense flunk-out courses in English and whatnot to weed everybody out very fast, and they did: these kids who were used to being the brightest were, I think, just misplaced and overwhelmed, and they dropped out pretty fast. But it was a purposeful program; and in order not to cause the out-of-staters, who were meeting other qualifications, to get mixed in with that behavior, they had separate classes for them in what was called the “Honors College.” The point of that was just to say there was a smaller group, but you were still on this very huge campus, and there was an aloofness of the teachers. Perhaps that’s because we were undergrads and they thought we weren’t very interesting. But when I went to Chicago, there were weekly teas, and the faculty were there. The University of Chicago at the time owned the Encyclopedia Britannica, and it was Information Sciences Year—the degree program was called Information Sciences, not Computer Science—and so they had visiting hotshots from around world come for a week. I say “world” because Maurice Wilkes came one week; he was actually doing a project with one of the graduate students. So suddenly I was in this much more supportive environment with respect to actually learning about the who’s who and why they were doing what they were doing, and you had a lot more social contact with the faculty and learned about what they were pursuing, what their research goals were.

The reason I went there was literally what I called earlier “vocational training,” because I felt I didn’t know enough to go get a job in computers; I had just dabbled at teaching myself. Plus, I think school’s a nice place to hide out for a while, and I wasn’t quite ready not to be in school. But there were all these women there for their Master’s Degree. There was one woman from the East Coast doing her Ph.D. program, and I think all the rest of the women were doing Master’s. I’m sure I’m wrong, but my memory is that two-thirds of the class were women. It wasn’t a big class, but still, I remember that. I also remember there were people from Naperville, [Illinois], from Bell Labs; there were two: a man and a woman. I remember a conversation we all had when we all discovered that he was on full scholarship, did not have to work to make up hours lost, and did not have to get certain grades to get the tuition paid for; whereas she had to make up the hours and had to get at least a “B” in order to get the tuition paid for. And her belief was that that was a distinction they made between men and women—which was the first time my ears had heard something like that, so it stood out.

Abbate:

Those were people from Bell Labs?

Goldberg:

Yes. I don’t know the facts; I just remember the conversation. I don’t know that I live my life with blinders on—certainly not anymore—but in those days, I was just enjoying what I was doing. Anyone who could spend the Summer of ‘67 in Berkeley and San Francisco and not see what was going on is really out of it! That’s me. [laughs] Pretty strange. So I remember some of these conversations.

The bottom line for me at Chicago was that it was just a lot more personal, a lot more supportive. Consequently, I found that I was much more heavily influenced by the other students and their opinions about how well I was or wasn’t doing, and what I should do. People were close enough to help each other a lot. So I liked that a lot better. When I say I didn’t like Michigan, it’s probably a relative statement more than an absolute statement.

What we were talking about earlier was, “Why were there so many women?” Especially compared to now, when we say forty percent of the student population should be women, and we’re not even half there. What’s going on? I have this conversation with people around the world, and no one has a really clear answer about what’s going on. People have a lot of opinions, and I’ve seen some research reports. I know what was going on at Chicago—or I believe what was going on there—was that this was a university whose business was having students, and you don’t just make offers out to a hundred people and have two show up. I’m not sure those were the odds; but if a hundred people are all men, and they’re being drafted into the Vietnam War, the odds of your getting a high percentage, between the competition with other schools and the Draft Board, are very low; it’s highly likely that you’re not going to get them. So they had a higher percentage of women they accepted, because they could guarantee that at least they wouldn’t be drafted. We didn’t get drafted.

I didn’t think about it, obviously, in those days. It wasn’t obvious to me that’s what was going on. But there were two moments when I suspected something unusual was happening. The first one was when I got the letter of acceptance, with a generous fellowship, and I showed it to Karl Zinn. He looked at me and said, “What! Why would they give you so much money?” Okay, the school cost a lot of money; it wasn’t that they were giving me money—although in those days, three or four hundred dollars a month in stipend was a lot of money. University of Michigan tuition for the year, my mother told me, was a thousand dollars in those days; I don’t know if she meant for a semester or for the whole year, but we’re considerably different from that today! So that was a good salary, plus the tuition was covered, or most of the tuition was covered. So he was just shocked! And I thought, “Well, I don’t know why, but I got it. Just say ‘Congratulations’ and let’s move on. Don’t be so shocked!” [laughs] So that was funny, because I didn’t know it was a big deal. I mean, I knew I needed it, but I didn’t know that anyone was going to react like that. But then, many years later, I had the good fortune to meet Thelma Estrin. In the Bay Area here, there used to be a group that did career conferences for women—high school girls, re-entrant women returning to college—to talk about math and science and what the career potentials were. They brought women in who would give a plenary session talk; the girls could ask questions of them; and then they did something that I think was marvelous on both sides, which is that they broke up into work groups, and there would be two women in very contrasting careers. I had a stock broker that I shared the time with, and we talked to the girls about what it was we did and why having done math allowed us to do our jobs, because math permeates so many different careers. It was really a wonderful day for seeing the way math is foundational. We did that several times, and one time Thelma was there with two of her daughters, Judith and Deborah, and we got into two conversations. One was, “Does it matter that you get an A in your class?” This was a long debate I think she was having with Deborah about the importance of the grade. We didn’t settle that one. But then she pointed out to me that she got into engineering, which eventually led to a position running a research lab at UCLA, because it was World War II and literally she was the only one qualified. They couldn’t discriminate on the basis of her being a woman, which she felt they otherwise would have. So we talked about it, and I said, “Well, there were a lot of women in school with me, and it was the Vietnam War.” Now, nobody wants another war just to bring women greater opportunity, and of course we have a lot more equality in the armed forces now, where women go as well; but that was apparently the effect. I’ve never tried to track the numbers and look and see if there was a spike or not; it was just my personal experience with why I got such a good scholarship.

Abbate:

I’ll have to see if I can quantify that.

Goldberg:

I’d be interested to hear.

From U of Chicago to Stanford: Computers in Education

Abbate:

After that summer at Berkeley, you went straight to Chicago. Were you living with your parents?

Goldberg:

No. The University of Chicago’s on the South Side, and my parents lived on the North Side. That’s quite a distance. I shared an apartment with four other girls, all of whom I had gone to high school and Michigan with, and they were all in different departments. It was a brand new apartment building that Chicago put up—women only—for undergraduate and graduate students. It was a very interesting idea. So it wasn’t too far a walk to where classes were. I didn’t think of the university as dangerous, but around that time there had been a breaking of the code where you leave students alone and the gang wars are handled separately; there would be drive-by shootings, and students would be injured. So there was a little bit more caution, not always pleasant.

I think I mentioned earlier that Chicago owned the Encyclopedia Britannica. You know, I don’t know the extent to which professors know the high impact they can and do have on students, but I’ll lay you odds that a high percentage of students will tell you that some professor came up with some idea that made a big change in their life; and mine was a very strange one. In those days, in ‘68, there were big sit-ins at the university. I was teaching a programming class, and we did a lot of joint stuff with the Business School. Many of the Information Sciences faculty were in the Business School, and we also had visitors come. Patrick Suppes from Stanford was one visitor. He and Richard Atkinson, who’s now the Chancellor at the University of California, ran the Institute for Mathematical Studies in the Social Sciences—IMSSS—which was an NSF-funded activity to do educational technology. They ran one of the three NSF-funded experiments in computer-aided instruction: in their case, a big time-shared system teaching math and reading. These two gentlemen both were the who’s who in mathematical psychology, with Suppes also in philosophy and set theory; he had written a grammar school–level book on set theory—part of the “new math.” They were doing a lot of work to understand mathematical models of learning, and he had come for the week to Chicago. Roman Weil, who is still a most interesting professor at the University of Chicago—he was at the Business School and was joint with the Committee on Information Science, where I was—threw a dinner party for Suppes. He invited me, amongst other people, and at the dinner told Suppes that I, during my defense for my Ph.D., had expressed an interest in computers in education, and I needed a year of real immersion in that, and shouldn’t Suppes invite me to Stanford for the year? Which was remarkable for Weil to do, because Suppes on the spot said “Sure!” and next thing I knew, I was headed to Stanford for the year—and never returned! [laughs] That’s how I ended up in California in ‘69.

Abbate:

So you had already decided you were interested in computers in education?

Goldberg:

Not exactly. As I told you, I went for my Master’s, not my Ph.D. Every school’s got a different rigmarole; at Chicago, you took an exam to get your Master’s. You took your exam before you actually finished all your courses, and you had to pass at a certain level to get your Master’s. If you passed at a higher level, you were allowed to do an oral exam to see if you could stay to do your Ph.D. I passed high enough to do the oral exam, but I wasn’t going to do it—I mean, I only took the exam because I had to for the Master’s. But the other students convinced me to do the orals. It’s kind of a long story, but I ended up taking the exam, and the exam was this painful three hours of standing in front of the entire faculty while they asked any questions they wanted to. It might have been shorter if I’d understood the game, which was to get me to not know an answer—they just keep asking questions till you’re convinced you don’t know everything! That’s what it came down to. I keep wondering, “If I had just answered the first question wrong, would I have been out of there sooner?” [laughs] But once you do that, they’re sort of satisfied, and they ask you the most important question of all, which is, “What is going to be your concentration for your dissertation?” You’re supposed to have an answer, and I hadn’t talked to anybody; I hadn’t prepared; I hadn’t really thought of doing any of this. I only knew one thing, which was the job I’d had at Michigan at the Center for Research on Learning and Teaching, so I was winging it: I basically said that’s what I was interested in. Which I was—but I didn’t know whether it was the right thing to do a Ph.D. in. I had no idea!

But it opened up a can of worms, because Chicago was funded by the Atomic Energy Commission, which wasn’t exactly for educational research; rather, research was in designing new machines. And by the way, Richard Weiland, who had been at Michigan playing T.A., went to Chicago, too; so he was still there, doing that to the students! [laughs] It was very funny: I said I didn’t think he should do it that way! And I was teaching an undergraduate Business School course in comparative programming languages. Again, it was a small department: you had chances to teach; it was your class, you weren’t just the T.A; so there was a lot more step-up-to-it opportunity. So you get a certain confidence building doing that. But I said I was interested in computers in education, and Roman Weil was right: with very little prospect for proper funding, they needed to do something about me. And, as I said, the school is in the business of having students; and one of the businesses of a new fledgling department—which, in Chicago, is called a “Committee”—is to graduate Ph.D.s if they’re ever going to be a department; it takes like twenty-five years to be a department. They didn’t make it. Basically, they were shut down, and then many years later they restarted a Computer Science Department.

So I went to Stanford for the year. I worked in L.A. that summer at the Southwest Regional Labs, which was education work, and then went up to Stanford and worked for Suppes. During that year, the problem he had posed for me to work on I wrote up as a proposal for my dissertation; I submitted it and defended it, orally again, in Chicago, and it was accepted. And it was just an accident; I didn’t know the rules. The rules are, you get five years to do whatever you proposed and they accepted. There were no other constraints. Suppes offered me a full-time Research Associate job, and in contrast to going back to Chicago and teaching undergraduate programming again, it was much better pay. I’d be paid full-time to do what was going to be my dissertation, so it was a no-brainer: I stayed. That, and the weather! [laughs]

Abbate:

How did you pick the topic? What was the topic?

Goldberg:

It was actually handed to me. I mean, basically Suppes proposed—and I liked—the topic. It’s a problem that still persists today; it has not really been properly solved. If you’re going to use computers in education—if you’re going to use a computer as either some kind of mentoring or some kind of substitute teacher in any form—the student that’s doing work eventually is going to say, “I’m lost”; “What do I do now?”; “How do I do this?”; “Why is my answer wrong?”; or “What do I do next?” Specifically, Suppes was teaching a course in the philosophy department on set theory. He was doing online proof construction for axiomatic systems. A student is given the axioms and does “symbolic manipulation” with a little command language on a teletypewriter; when the student gets stuck, they ask for help. The only thing they knew how to do—which is mostly what they have now—was to have stored answers; stored hints. So you give them a hint. Now, had the student asked for help before they did any work at all, the hint would probably have been a perfectly reasonable suggestion. But in a situation where somebody has actually wandered their way down and done five or six lines of proof, it is likely that what they did is finishable. Maybe you’ll say to them, “Well, there was a simpler pathway,” or “There’s an alternative way to do it,” or “You missed the point that this is really just an instance of an existing theorem you’ve already proven.” So you’d want to be able to guide them. A human being working with a student would never give them a canned hint; they’d look and see what they did. Suppes called that the “Dialogue Problem”: how do you have a dialogue with a student? It manifests itself in different subject areas differently, but in this particular case, it was about theorem proving. So it was an interesting crossover: I could both do theoretical applied math and work on a design for a new way to teach symbolic logic—one in which students could actually define a new symbolic language, state the axioms, and derive theorems. So you actually could up the level to graduate level, where they’re really thinking about math as a language, a formal language in which you can posit theorems and prove them. I ended up building a fairly sophisticated system that Suppes then used at Stanford for many years, in the Philosophy Department, to teach with.

Abbate:

What was that called?

Goldberg:

It was the symbolic logic course; that’s what it was called.

Abbate:

I mean the system. Did it have a name?

Goldberg:

No. [laughs] We were not in the computer group in the AI Lab; we didn’t give names to things! That’s probably why we weren’t really very well known.

But the Dialogue Problem piece of it then became a challenge—to say, “All right, we’re teaching the students how to do theorem proving. Can you write a computer program that would do what the students did?” In those days, artificial intelligence and intelligent computer-aided instruction assumed that the computer, as tutor, could do what you asked of your students. So I wrote a mechanical theorem prover that captured what I felt we were asking the students to do in their behavior: a mechanical theorem prover that could actually take the axioms and the rules of legal transformations, pose a theorem, and prove it. Then, when the students asked for help, you would take the lines they did as also givens—already proven—and starting from that point, see if you can finish the problem. In addition, the system could recognize that there were other solutions to show the student. The long-term goal we never really did, which would have been to pedagogically have a basis for knowing when to say “Let’s ignore what you did” or “Let’s continue with what you did. After all, it’s just one more line; you’re almost there; but notice there’s another way to do it.” There’s a lot of interesting questions you could ask educationally.

Abbate:

What if what they did was wrong?

Goldberg:

Then you’d know that. You’d just discover that nothing there was going to be useful. Actually, you pose the interesting math question, which is, “How do you know, when you start to construct a proof, that it’s provable?” Mind you, this is curriculum; you wouldn’t put in the curriculum a proof request that wasn’t provable. We didn’t do the problems where you come up with counterexamples. For Chicago, remember, it was applied math; so what I did for the dissertation were completeness proofs, which is to characterize the space of expressions that could in fact be handled by the theorem prover, and then to prove that that space was covered by the theorem prover. So the dissertation actually ended up being mathematics, which is what was required; but behind it all was this other more general AI work, and also more practical work, because you ended up with something practical you could use with real students. We did a lot with the college students, so that was fun. That was very nice, though I assumed it was not going to be my lifelong job.

I ran into some problems getting the dissertation approved. You know, when universities have rules or allow things, I’m not sure they really think clearly on impact; but the rule was that every faculty member had to approve every dissertation. It was a small department. There was a faculty member who was actually at IBM Watson Research, and I had gone and gotten his approval before I went to Chicago and did my oral defense—which, mind you, was a very weird day, because I was doing a degree for Chicago under a Stanford professor. I knew more than anybody in Chicago about the problem, although of course they were all bright and they could certainly understand. But these are public meetings: faculty, students, public, anyone could come to a dissertation defense at Chicago in any of the departments and ask questions and give you a hard time. And some of the students, that was their job in life—to give other students a hard time! [laughs] I remember it was not a pleasant moment! Fortunately, I knew the subject area. But after the fact, because all my proofs were done in set theory (because that was Suppes’s approach), this professor located at IBM decided it should all be completely rewritten in category theory, because the theorem statements would be stronger. I tried to explain that this was not my life ambition, but I was obligated to rewrite it all, so I did it. I went to Brazil to teach a school term, and while I was there I rewrote it all. I was not happy about all that!

Abbate:

How did you end up in Brazil?

Goldberg:

Well, that’s what’s so wonderful about universities, especially in the computer field: they’re just full of students from all over the world. And for me, that turned out to be a big attraction, as Stanford especially is just full of people from all over the world. That diversity is just compelling as an environment to be in. And because the Institute for Mathematical Studies was in psychology, education, computing—it was an interdisciplinary program anyway—it got mixed in with this international group. So there were a number of people there from Brazil who were doing their Ph.D.s, who went back to Brazil. They were doing research in AI, which I was knowledgeable about. In those days, we had email, but we didn’t have the Internet exchanges and the support we have today; we actually had to write long letters back and forth. They finally said, “Can you just come and teach?” It was teaching a class for the faculty.

I just went wherever my nose was. You know, if someone offered up something interesting, I didn’t have any solid roots in the ground that said, “Don’t go, don’t go!” That happens when you have children, not before. So I picked up and went to Brazil. This was 1972, and when I found out I was going to have to rewrite my dissertation, I figured, “Well, it’s okay. I’ll do it down there,” which is what happened. I was down there from after Mardi Gras up till the summer. I then decided that I didn’t have to fly home; I was there with the guy who was my boyfriend, who became my husband, and we decided, “Well, we could go by land. We could get back to the United States by taking land vehicles. Let’s see if we can stay on the ground. Let’s see how long we can stay on the ground and get back to the United States.” We had a two-and-a-half-month trip that way. It was great. You can only do that when you’re young!

Abbate:

That’s so adventurous!

Goldberg:

Yes, that was fun! I didn’t speak Spanish well enough to be doing something like that, either. I was never really very good at foreign language—but so what! Actually, people were very nice and very helpful and very kind, even though we must have looked like we were a pair of dummies! You know, not speaking the language properly.

So that was a good adventure, and then I came back—that would have been the end of ‘72—finished up some papers with Suppes, and went to Xerox the year after.

Abbate:

Where did you meet your husband?

Goldberg:

At Stanford. I’m not married to him now. He was a student in the department as well.

Abbate:

And that was computer science?

Goldberg:

No, it was interdisciplinary, with students from education and math and psychology and computing at the Institute for Mathematical Studies. It was a very large program.

Getting a Job at Xerox PARC

Abbate:

How did you get the job at Xerox?

Goldberg:

How did I get the job at Xerox? Good luck! I was really lucky.

It turns out, despite the fact that I’m relatively shy and don’t join things, that I became active in the ACM. Dexter Fletcher, who worked for Richard Atkinson, said, “You know, we have one of the large National Science Foundation grants. We really should participate in the professional societies, so that people are more aware of what we’re doing and we can learn about what they’re doing.” The ACM National Computer Conference was going to be in L.A., and he said, “We should go down and go to the ACM Special Interest Group on Computer Users in Education; go to their meeting. We should join and participate.” It was in 1969 that I joined the ACM. As part of that, we also started out of the Institute, with other people in the Bay Area interested in computer uses in education, a local Special Interest Group that met monthly. These are wonderful things to do, and they’re wonderful things to participate in, because you can’t just use who you meet on the job as part of your learning. Whether you’re at a commercial company or whether you’re academic, you really need to find out the world around and learn from that world around—and not always things that you think you need to know that moment. A lot of people make that mistake; they really put up their blinders and do their work and they get kind of pigeon-holed. It’s important to see the context of your work.

So we had these monthly meetings. Alan Kay was at Stanford, and he joined up at Xerox PARC in the ‘71–’72 time frame and was hiring people. John Shoch was one of them; John later became a venture capitalist here in the Valley, but at the time he was a Stanford student and was in one of Alan’s classes and joined him at Xerox. So John used to come to these monthly meetings, because the project they were starting at Xerox was presented as, “We’re doing a dynamic book, a hand-held computer”—remember, this was 1972, and everyone thought that they were blooming crazy—”and it’s really going to be for education.” They were very focused on the educational uses. So they came to these meetings as well. So that’s how I met John, and at one point he said to me, “Alan would like to meet you.” So I met Alan, which just totally changed my whole way of thinking about everything.

We had been, at that time, arguing with Suppes that we were continuing to institutionalize education. One of things that I think goes wrong is when people think you learn in a classroom. You send your kids to school, and the teacher teaches them in this building; but when they leave the building, learning stops, and you just act on what you know. Of course, the well-educated people in this country don’t operate that way; they’re in homes where they’re learning all the time; but in a large part, if not a majority of the world, that’s not what happens. So that’s what I meant by “institutionalizing”: you go to the building, the institution. I don’t know if you remember when calculators first came out, and they were going to be such a marvelous tool in the classroom; but did you know that this portable, carry-it-with-you, have-the-tool-with-you-all-the-time invention had a little attachment to it that locked it to the desk in the school, so that kids wouldn’t steal them and take them home? Sort of the antithesis of the idea. What we were worried about, and tried to get Suppes to understand, is that with teletypewriters, or even if you moved onto display screens (we had Imlacs) for your technology base—and there was the big PLATO project as well—the students still had to go to a place where they talked to the machine and learned; and that really you needed to use computers as tools for helping you in your learning process, and that was all the time, and it should be carried with you like your pencil and paper. He didn’t want the Institute to be working on that, for whatever his reasons were. So meeting Alan, and seeing the Dynabook Project in this group, I thought, “Right! That’s exactly the point! The point is, you’re going to have your own computer. You’re going to carry around a PDP-10.” That’s what we thought it was, a PDP-10, which is lightweight compared to what we have now. It was going to be there, like a notepad, but it was going to be a computational-based device that allowed you to build models of the world and test your understanding of the world. What students who succeed really do is, they interact with teachers and parents and other people—and it could be other students—who challenge the models they build of their world; and this is true whether it’s science, math, or social studies: that you’re constantly constructing and deconstructing. And when someone really pushes you to think about your models, make the models explicit, talk about those models, and be able to build those models and then challenge those models—I mean, every well educated person I know remembers in their childhood that that was what was going on all the time. You know, it could be physical—the guys who take apart engines and put them back together again, to understand how things work—you can do that physically, or you can do the same things mathematically. And you need that ability to take apart and to build with you all the time; it’s a tool. If you get an idea, you need to jot it down; you need to capture it. So that’s what Alan was all about. And I thought, “Wow! A whole group doing ‘the right thing’.”

So I told Pat I was going to Xerox, and he said, “But we didn’t finish our papers.” I said “We’ll finish our papers, and then I’ll go to Xerox.” I preferred that to academic jobs. In those days, the law required—I mean, federal grants required—women faculty, and they were looking for women, and there just weren’t many graduating. You can’t fault a university for not hiring women professors when there weren’t any to hire! But that wasn’t what excited me, what I really wanted to do.

It was the summer of ‘73 that I went out to Xerox. I stalled it out a few months to finish up at Stanford. And I went there pregnant with my first child! [laughs] I remember saying, “Are you sure it’s okay? I’m pregnant!” “Sure, it’s all right. No problem!”

The Work Environment at Xerox PARC

Abbate:

What was the environment like there?

Goldberg:

Well, you know, there are always two parts to any situation: there’s what’s given and what you take. The environment, when I first joined, was brand new, which means that they were not fully capitalized; so you could decide what you wanted, whether it was furniture—they chose beanbags, which for a pregnant lady wasn’t a really good idea [both laugh]—or what the machines were going to be. It was the first time around. But I remember saying, “Well, this’ll last a few years, and sooner or later they’re going to have to recapitalize.” But I remember Alan saying, “We have five years to produce the dream.” We were told, Do anything that was our dream, and the company would then take it from there. Sort of this idea that you were creating a whole new industry for the company.

It was very comfortable. There were wonderfully smart people, who were very supportive of one another, at least in our group; I don’t know all the groups. Our group was a mixture of academic and non-academic backgrounds, mostly not; and a mixture of men and women. Alan was very good at spotting people who were bright, so there were actually two women, I remember, who had been secretaries, and he said, “No! You’re programmers,” and he brought them on as programmers. He was very supportive; gave them whole new careers; and he was right, in all the cases! So we had a certain amount of flexibility; but eventually bureaucracy set in. We had machines we wanted to take to schools for experiments, and you couldn’t just take them out of the building. You had these arguments; I mean, that kind of stuff sets in eventually in any organization. We thought that the hardware would happen in five years, and that it was a software problem, and it took longer, because the packaging took longer. There was frustration getting people to believe in the same thing you believed in, because it was a big idea that required a lot of cooperation. So there were a lot of things that did happen, but they happened by subterfuge, not because top management knew what was happening. This is written up; there are a lot of classic stories about that.

From my personal perspective, doing research that way was an ideal thing, because I was a mother with young children. I could come and go as I wanted; if they were sick I could stay home. We had a really big project that we started at the end of ‘77, which was to teach Xerox management about software and the flexibility of software and how software changes what the hardware provides. You see, in those days, if you made hardware, you made hardware; hardware was the product. This whole idea that software changed hardware was just revolutionary. So Bob Taylor, over in the Computer Science Lab—we were in the Systems Science Lab, in a different lab—he wanted to run a big workshop, a multi-day workshop, for the President of Xerox and nine other people. So there I was, asked by Alan to put together all of the software courseware and run the Lab, and I had just had a baby! But the computers of those times were Altos, and I think I got the first one at home. We considered it the heater for the winter! [both laugh] It heated up the house. But I had it at home so I could work at home, and then I’d go in at night sometimes; but I was nursing, so I’d sometimes take Rachel with me. One of my favorite nights was when Rachel was crying, crying—maybe she had a stomach ache or whatever—and I was trying to work with some people, and we only had nine weeks to put everything together, so it was a fairly tense time. Terry Winograd, who’s just a wonderful person, a Stanford professor, was there, and he took Rachel in his arms and went down the hallway with her—screaming—just so we could get the work done. And I thought, “Wow! I can’t believe he’s doing that!” At the time he didn’t have his own children; of course, he does now. But I said, “Boy he’s going to make a good father!”

[START DISC 2]

Goldberg:

I found it very supportive. I remember, when I first joined there, going to IFIP [International Federation for Information Processing] in Sweden to give a talk, and there were Bob Metcalfe and Butler Lampson, coming to my talk on a topic that I didn’t think they had any interest in at all; and they said, well they were just there supporting me. Just little things like that were very nice. So when things got a little more sour and became more of a zero-sum game, it was kind of too bad, because it had started out really nice. But you know: money’s money, in terms of people’s egos and what they were trying to do, and eventually the company was going to have to have some payback for it. There were a lot of things that went on there that gave Xerox good payback. Xerox’s entire high-end printer business, for example, was certainly driven out of work of people like Severo Ornstein. There was good work that went on, but people know the big things that didn’t happen. And things got more complicated when there was a division next door, the Office Systems Division, that was doing a particular product line in the eighties and expecting more support from Xerox PARC, from the research—and getting it; but you can’t always solve the problems when they’re marketing problems, they’re not computer science problems. It’s a different issue.

Developing Smalltalk

Abbate:

Tell me about Smalltalk.

Goldberg:

Okay. First, I should tell you there are some papers; one in particular that I wrote, called “The Community of Smalltalk.” It was in a book on programming languages.[1] There are the History of Programming Languages books, and in the second, HOPL 2, there’s a wonderful paper that Alan wrote that was the history up to the late seventies, his history, until he picked it up again doing Squeak.[2] Then there was the transcription of the talk he gave—because that’s what he really wanted to put in the book, but it wasn’t what Jean Sammet, who was editing the book, would let him do. After each speaker they had somebody who commented on what they said—this was the way the history conference did it, so all of those were there for history purposes—and I wrote that. “The Community of Smalltalk” was a separate paper that quickly summarized what Alan had done and took it further into the business issues.

So: what do you want to know? When you say, “Tell me about Smalltalk,” what are you asking?

Abbate:

What were you doing? What was your experience?

Goldberg:

Smalltalk was the name given to the software for this personal computer—this hand-held, carry-it-with-you, have-a-network, talk-with-other-people, exchange-ideas, build-models. The basic idea was that anyone who wanted to could learn to construct and deconstruct models to understand with, so the language had to be a simulation language. The way the research worked was, you’d come up with an idea for the language. In this case, a lot of the ideas were a synthesis of prior ideas from work done in the sixties in a language called Simula, which is where the initial computer object construct comes from; but it also comes from work done in graphical interfaces out of Lincoln Labs, Ivan Sutherland’s Sketchpad. What we would do is, we’d design a language and implement it. The primary implementer in those days was—well, still is—Dan Ingalls, who is just a wonderful person.

I guess there was a Smalltalk 71 that they had done, but the first one I think I really remember is Smalltalk 72, which was done in BCPL on a Data General Nova. They really were simulating what they thought things would be like on the vision, the Dynabook, and doing some of the initial low-level work in animation—freehand drawing in animation—and music, and looking at Doug Engelbart’s work. Doug was the inventor of the mouse, so they were looking to see what kind of devices could be used to give better access, so you didn’t have command lines with all the funny syntax. Then the Alto was built in cooperation with the Computer Science Lab, and Smalltalk 72 was put up on there.

Smalltalk 72 was an extremely odd language. It was the first Smalltalk I saw, and when I first arrived, my job was—the way I put it—”to find a way to teach this to anybody.” And I recall thinking this wasn’t going to be teachable to anybody; and it wasn’t because there were a lot of very strange font characters; it was because the nature of the language was that it wasn’t statically parsable. You couldn’t look at it and say, “Okay, I know what it means”—because as you started looking at a token, that token would actually execute as a method to an object and eat up things along the line. So it would look ahead and see if it wanted the next token, and if it wanted, it might eat it up; and so depending on the implementation of a method, the object was gobbling up the message stream! So you had this notion of objects and messages, but you couldn’t really look and read and know what the message was; you had to actually interpret it in your head.

Abbate:

So you had to follow the entire thing from the beginning to see what the meaning of each line was?

Goldberg:

Oh, yes. Yes, it was very strange. So for very simple things—very, very simple things that were being mirrored, which was the LOGO turtle geometry capability—there was a little picture of a face, which was Myrtle the Turtle; and you could tell Myrtle to go a hundred, and it would go a hundred. So the first thing that Myrtle did was see “go,” and then gobble up the hundred and do that. Well, that’s okay, but when you get more complex things: maybe it wasn’t as bad or as condensed as APL, but it’s the same idea. In APL, at least, you actually can parse it without knowing the implementation. Here, you had to know the implementation. So I wasn’t entirely very happy with the teaching implications. But I did do some experiments with kids to see if they could learn and what was going on, and I brought from Stanford with me, from Suppes’s lab, a student who worked in the lab for a while, Steve Weyer, who was also interested in education. We were looking at a variety of uses of the Dynabook as a dynamic medium: what does it really mean to capture information and be able to make it accessible for manipulation? So the idea isn’t that you store and retrieve information; it’s that what you retrieve becomes something you can manipulate, which is a very powerful idea. And the notion that you, as an average person, could express how to manipulate it is even more powerful—and, given that it was 1973, pretty unusual.

So the research program was always: do a language, do its implementation, create uses, try it out. We had a target set of media-based applications we were interested in—document editors, and various simulations, and music, and art, and how they would interact—and so we would build these. And the question was, “Is how you built it—how you expressed it—readable, understandable, accessible?” and “Who could build what?” Because we were interested in getting young students—ages ten, eleven, twelve—who could understand a certain level of abstraction, to be able to build tools. They weren’t just going to paint; they were going to write painting systems! That was the idea.

What was kind of interesting was that Alan was very much influenced by Piaget as a psychologist, as a pedagogical mentor, if you will; and Piaget always said that you start with the concrete and go to the abstract, and that’s how kids learn, and that there are certain ages when they’re able to move from the concrete into the abstract. So we took that as a given, as to how we would devise a course of study. The team was investigating implementations. You’re always worried about things being fast, especially as these were interpreted systems at the beginning; but Smalltalk 76 was a redesign (the date is the year of the design, not necessarily the year in which it was finished and deployed), and what we were trying to do was find a way in which you could have a language system. I want to say it that way: it wasn’t a programming language, it was a programming language system, because it included the kinds of tools that help you express your ideas in the language and understand what’s going right and what’s going wrong—both are valid; and be able to look at things that do work and take them apart—see how they work—so that’s kind of the “deconstruction” notion; as well as have parts that you can compose with. So, regardless of what names you give them, that’s conceptually the right idea. For the kids, what you really wanted was to think about was: “What did you want them to learn?” What were the basic powerful ideas that you wanted them to understand about constructing and deconstructing in a computational environment, and therefore, what were the fundamental components of a language system library that would be useful to them? What would you teach them, to have them have a way to think about searching and finding? Actually, all the Smalltalk interfaces were done as a challenge on “search.” How do you search and find how things work? It could cross-reference questions: “Who sends this?” “Who implements that?” “How are these parts related to one another?” “Oh, I like what that looks like; how do I take it apart and see what it’s made up of and then change it?” So this incremental change became a part of our teaching approach. So it was very driven, not so much by the academics of computer science as by the notion of who the target users were going to be. They were not professional computer scientists.

Smalltalk 76, then, had a class structure with a hierarchy. The previous one was mostly objects, classes with instances; but in 76 we had more of a hierarchy, where you can define an object as being just like another object, except—and we’re talking about exceptions either being characteristics or behavior. So I can say, “You’re an employee, except you’re a special employee.” And what would that be? “Well, when you get paid, you always get a little extra under the table.” I shouldn’t say things like that! But that’s an example—versus an instance, where you’re all cars, and the instances are distinguished by what color you are, or how big your engine is, or whatever.

What was very interesting about Smalltalk 76 was its implementation with a memory manager that was garbage collecting, and a way of writing a system that Ted Kaehler did for snapshotting frequently onto the disk—because we were always worried about these little personal computers that were being built, that things would crash, and it’s too frustrating to lose information. And you wanted to be able to replay, to recapture work potentially lost due to a crash. Bravo, the early text editor that was done in the Computer Science Lab, because it was under development, would keep a complete incremental track of exactly what the user did; so if you lost something, you could always replay it. This turned out to be the major aspect of the text editor that made it, in its early days, a tool, not just a research project. So we kind of did the same thing, mirroring it in the memory manager for Smalltalk; and that was very interesting.

Remember I said earlier that we’d had this big project to teach the President of Xerox and the Chairman of the Board and actually the top ten people? The way it was designed was as a discrete-event–driven simulation world. The implementation was general purpose, based on a Simula system called Demos. We specifically had five or six simulated contexts that the participants would work in. One of them was the Xerox Copy Center; another one was the production line for VersaTech—VersaTech was making the chips for Xerox printers. So you had things that were about the business, so that the workshop participants could be interested; and it was very visual, so you could see the queue for workers, and see the work get done, and collect data on it and stuff; and they were going to build these things. It turned out the best way to do that, of course, was to have a framework for simulation that was specializable: there were components provided, but those components were only starting points. There was a general notion of a “worker,” but then you’d have a sub-class of a particular kind of worker, which would be a different Xerox copy machine; and there would be a center full of different machines with different characteristics, in terms of how fast they were and how fast the toner ran out and various other things. So that was a big push on sub-classing. I remember we were a week before the big day, and we had just been working full-out, when two of the guys, Ted Kaehler and Dave Robson, came to me and said, “The memory manager was designed for lots of instances and a few classes, and what you’ve done—which, mind you, is conceptually the right idea—is created a system that is class-hierarchy–oriented, with only a few instances. And we think we’re going to run out of memory. We think we’re going to have problems.” And I asked them what they suggested. They said, “Well, we think we should write a new memory manager, and we think we have an idea how to do that.” And I said, “One week before the workshop you’re going to put a new memory manager in!” And see, the whole idea here of components— where you can pick any level of a system, and take something out and put a replacement in, because the interfaces are stable—was now about to be fully tested. So I suggested we do the following: that Ted come up with a scheme he wanted to do; Dave check it out; and if he approved, implement it. If it works the first time, we’ll go for it. If it doesn’t work, we pull back, and we just be so careful; we just test the curriculum, and we just skate on this thin ice; because it wasn’t worth it to put a memory manager in that would crash all over. That’s not a good impression. It had to work.

We were really having a good time. We had color output; we were using the color printer; we had music playing. We were just getting a little carried away, having a good old time—which was making Bob Taylor very nervous! And Bob failed to tell me that I had to also give a talk to introduce this workshop; so when they did the dress rehearsal, we weren’t prepared, because we didn’t know about the talk. He was livid! He went to Alan, livid; Alan beat me up; and I said, “Wait a minute! We didn’t know! We had no idea; no one told us.” Because it turns out that the workshop was mostly the guys in the Computer Science Lab talking, and we were concentrating on doing. So, that wasn’t very pleasant. Well, I don’t know if it worked the first, second, or third time; all I know is, they pulled out the full memory manager and stuck in a new one, and it worked just fine.

It was really amazing! I mean, for a researcher—this is about computer people: When you do research, you create something that’ll be useful, because nobody in computer science wants to build anything that no one’s going to use; it’s no fun! What’s your prototype is somebody else’s product. You don’t know! Many years later I did a special project called “LearningWorks,” which was a simplification of the commercial Smalltalk system for learning purposes. The Open University in the UK was doing a new first course in computer science, and for a variety of reasons we agreed they would use this kind of experimental LearningWorks system—version 0.7 was about the time when they had to freeze. [laughs] We might have gotten to 0.9, but then we stopped. And they put out this course, maybe four or five years ago; they’ve taught five thousand students every year; and it’s just been totally robust! I remember when we went to England to get the bug list, figuring we had a long list to deal with; and the bugs were all in their curriculum, not in our system. There were obviously things we could have done to improve it, but it was basically good enough. And that’s just really a testimony to the system on which it was built. There are a lot of systems like that out today, where people really understand the layering and the incremental development, but Smalltalk was pretty early in that.

So we had that adventure with the Xerox executives, and we had a very successful workshop, but I don’t think it changed their mind that much.

Smalltalk 78 was an implementation on a machine that we worked on with another group to design, called the NoteTaker—which was a luggable, not a portable; about suitcase size. We had it done about the same timeframe that Osborne had his luggable, but Osborne mispositioned in the market and messed up his business. We never got to build the refinement; we just had several prototypes. Doug Fairbairn was the key designer, in Lynn Conway’s group. We couldn’t get Xerox interested in portables. We had spent a lot of time with East Coast Xerox groups; we were really starting to work with the divisions and find out who was interested in our ideas, who could benefit; and we went after the guys who managed the technical documents, thinking the technicians really had a problem with getting hands-on information. They could carry this around, take it right in the office. They’d all look at it and they’d nod their head, but it was just too new; it was just too novel. We’d say, “But their trunk is full of all these out-of-date technical documents, and they can’t take them in to their work sites. This way they can carry it in; it would all be there, and they can download up-to-date stuff every night.” You know, it seemed very obvious to us, but it wasn’t very obvious to anybody else! It was a project I worked on with Larry Tesler. It was a disappointment that they didn’t buy into that.

Managing the System Concepts Laboratory: Publishing and Distribution

Goldberg:

Alan, in ‘78, went off on a sabbatical down to USC-ISI, the Institute down there, and basically didn’t come back. I sat in as Manager, and then became the Manager in ‘79. We had done a lot of years of work that hadn’t been published, or hadn’t been made very public, except for these talks that Alan gave; and his was always very futuristic and confusing: Did that exist? Did it not exist? I had been much more involved than other team members because of my experience with SIGCUE in ACM, in the professional society, and knew a bit more about publication opportunities. So I said, “We should publish all this!” The easiest way to publish so much stuff is to do a book, so we tried to all do a book together; and it was a terrible experience, because people hadn’t ever tried to write about it before, and it was very confusing. I took a liking to the idea that before you build a system, you document it; so we decided that really what we needed to do was to rewrite the Smalltalk system, knowing we were going to share results broadly, and write a book about something cleaner that could be distributed. I went through a process with my manager, who was Bert Sutherland at the time, that we wanted to do this; and he was so wonderful. We got Xerox to agree that nobody in the company was interested in Smalltalk [laughs]—that way, we knew if we did the work we could publish it. Bert helped us in the eighties to round up several companies that would participate in the specification of the implementation and do implementations; because public distribution meant standard ASCII character set, which for us was crazy enough—because we’d always been doing unusual font-based stuff, and we could do it—and it meant standard microprocessor technology, which was now becoming available. So over the next several years—what was planned as a year project took three years—we ended up coming up with three books, and a system that was distributed under license to universities and some commercial organizations.

But this system now changed its goal a bit, because up to this point, it was a very loose system in which we were experimenting with what you could teach kids. Alan even taught adults, but that didn’t go very well. I think it’s because kids will sit there and they’ll listen to you and try to learn—kind of respect the adult, because you’re the teacher, right? When you have trouble with kids, it’s when they don’t respect the adult; but basically their attitude is, “We’ll wait and see what you have to offer, and we’ll try out what you want us to try out”; and then if it works, they could move forward. We shoot down a lot of creativity in school very young, and our project was a way to try to say, “No, no! It’s okay! Do what you want!” I remember a lot of very bright students—I worked in Palo Alto with the “Mentally Gifted Minor” Program—where you’d say, “Okay, do what you want. What would you like to do?” It was startling to them: “Wait a minute; I don’t know!” And that was a surprise; that was too bad. But here, they had a way of doing it and did a lot of really interesting stuff. In fact, a couple of the kids we hired actually wrote papers for one of the personal computing magazines. One was a girl, one was a boy. As young adults, he went off to Apple and did the first Mac Finder. She didn’t stay in computing, but she was great. She taught some of our classes. You know, we really let the kids take responsibility, if they wanted. There are a lot of papers on all this.

For me personally, it was like, “Follow your nose to do what felt right.” I wasn’t as academically pushed in terms of having to publish or perish, and I didn’t have to be prepared for a classroom every day, although I taught a lot. I used to teach computer literacy as a volunteer in the junior high school; there’s a certain amount of preparation required, but it wasn’t the same pressure as teaching a full load in K-12 or college. But this project was full-out, full-time, because we were driving it from writing a book and creating something that was going to be distributed—and distributed to software engineers.

Our goal was to teach hardware designers and their software counterparts what we thought might be the need for something special in the hardware. It turned out we were wrong. There was a wonderful project at Berkeley that was part of this, which was the SOAR Project, which later became the Sun Microsystems SPARC. SOAR is “Smalltalk-On-A-RISC,” a project by Dave Patterson and David Ungar. It turned out that you really didn’t necessarily need special hardware, and the Motorola 68000 was handling Smalltalk quite well. I had some run-ins with folks, because I wanted to bring someone else’s hardware (in this case, Sun) into a world that had always been proprietary: Xerox machines. I wanted to know what everyone else was doing, whether our stuff would run on anything else.

I think that when we finished the books and got everything out, a couple of things were true. We were doing this really to feel complete, and good about what we did, and to get people’s help. I don’t think we ever guessed ahead, ever planned ahead. You know, you start a business, you plan a certain success. You have a plan: “I’m going to get to a certain place.” We didn’t do that. We were communicating. So the next years after that were astounding for me: to discover impact that you didn’t plan to be so broad, and to see the most amazing things that people did with what we put out, and the learning cycle that then came out of that. But it was a diversion from the educational activities; it was a learning cycle in software engineering: what worked and what didn’t work; what we were not taking care of with respect to software engineering needs. What I call the “halo of methodology”—methodology both in finding objects and knowing how to run projects for reuse—and a certain class of tools that were needed for team support were where the real payoffs were. So that was very interesting.

Starting a Spin-out Company: ParcPlace Systems

Goldberg:

But our big commercial involvement was with the CIA.

Abbate:

I didn’t know that.

Goldberg:

Neither did I, initially, because we had a group at Xerox called Xerox Special Information Systems (XSIS), and they did intelligence community contract work. Very sharp people, down in Pasadena. Their customer was taken on a tour of Xerox PARC and shown a lot of stuff. I was Lab Manager by then and should have been told, and found out after the fact that XSIS had committed to doing a big, new approach to workstations for analysts, using our research system. And I thought: “This is crazy! This is a research result.” This was undocumented. It was 1981 or ‘82. I mean, we were on a publication pathway, but there wasn’t the book; there wasn’t documentation. I found out about it, and I said to the guys at XSIS,“Look, you’d better show me what you’re up to, because making these kinds of commitments without an engineering plan behind it doesn’t make sense, and I don’t know how stable we are, in terms of whether we’re going to want to change everything.”

Alan always said, “Never have customers,” because then you have to support them! It was very interesting advice, but of course without customers, you don’t learn; you really don’t learn enough, so they’re really the most important thing you can have—never mind revenue! [laughs] I mean “customer” in the more generic sense: someone who cares about and is using your results.

I went off to Washington to see what these people had done, and what I saw was a completely different interface, with some experiments on what they were going to build—and what, in the end, XSIS built for them—which was a new approach to an analyst’s workstation. I said, “How did you do that?” I mean, I was just astounded! “How did you do that?” The guy I was talking to had been the one who’d programmed it, but he wasn’t a programmer; he was a photographer by trade. This was NPIC: the National Picture Interpretation Center. So he showed me; and I said, “How did you know how to do that?” And apparently he’d been at another talk I’d given, where the idea of deconstructing and constructing had been presented; you take it apart. In Smalltalk, you can be in the middle of doing something, you hit “control-C,” and it will stop and open up a debugger and a traceback; so you can see, for the thing you were interacting with at the time, what it’s made up of and how it works. So you pretty much could go find out how things were working and piece things together, and that’s what he did! And I said, “Well, how did you piece it together?” Because we had overlapping windows—Smalltalk 76 was the first place that was done—and he had a tile system, because he thought it would be better for the early learners. So he showed me the code, and it was a mess! I said, “Oh my God, that is a terrible program!” And I showed him how to fix it, and he said, “No, I’m not a programmer; it doesn’t matter! Programmers will come along later on, after I show them what I want, and they’ll write it correctly; but I now am empowered to show them—to design, and show them what I want!” We gave him a language for communication, which was phenomenal!

They built wonderful systems that they used for years, and there was really no good commercial competition for it. But it was because of them—and their community of subcontractors, all of whom they demanded learn object technology and use this stuff—that there really was a commercial need for support. I had suggested to Xerox that they might want a business unit or something like that. So I think I ended up in business without thinking through the impact on me personally; what it really meant. This needed to happen, and it wasn’t happening, and the team was frustrated with the lack of interest on Xerox’s part, and I thought, “Well, we should just get into business.” We had a customer, who represented maybe ten percent of the money once we did get started; but we were making money! We were selling licenses for this published thing. This was happening at the same time that I was Editor-in-Chief of [ACM] Computing Surveys, so I was trying to stay in touch with the computer science community more broadly, and had agreed to run for National Secretary of the ACM. I was National Secretary, and then I was President from ‘84 to ‘86. It was just after that—it was like someone said, “Oh, she doesn’t have enough to do now”—that we were trying to get Xerox to start thinking about this more seriously as a business. Their problem was that they’re a hardware company, and to start a new business line you had to put a hundred million dollars of revenue on the bottom line and show how you can get there, and the only way you could get there was to claim you were selling hardware, and we weren’t. Nobody in those days did hundred-million-dollar software companies. That was a big surprise, too! And that’s not that long ago.

We ended up writing a proposal: Xerox had an Innovation Board that we went to and presented the option, and they said, Well, there was no place in the company, and if we wanted to spin it out separately, we could. Those were the days when venture capitalists wandered around universities and research centers saying, “Do you want to start a company?” So we’d been approached. It took a long time to negotiate an exit, but we were the first group to spin out a separate, venture-backed company out of Xerox.

Abbate:

And that was ParcPlace?

Goldberg:

That was ParcPlace Systems. I don’t know, if I look back—I mean, you don’t look back on your life and say, “Should I have done that?” “I should have done it differently”—but I knew nothing about running a company; absolutely, totally nothing!

Abbate:

What would . . . ?

Goldberg:

What would I have done differently? I learned over the years to bring in a lot more—for each position, to bring in very experienced people. Learning on the job costs you a lot of time. We lost eighteen months of negotiating with Xerox, trying to do things right, and I don’t think we won because of it. If the stuff was public, we could have gone out and started something on our own. I liked the idea of partnering with Xerox; I like the idea they made money; but I think it cost us too much in lost time. And it didn’t really teach Xerox anything. They ended up with a totally different approach to starting companies that Bob Adams manages. I thought I was doing the right thing globally, but it was the wrong thing locally. If we’d known it was going to take eighteen months—but you just don’t have that kind of foresight. Hindsight’s wonderful; you need a little more foresight!

That was one problem. The second: I had no business experience. I’d never run a company. I didn’t really understand what I understand today. I think if I’d gotten a venture capitalist and said, “Look, we can do all this, except for one piece: we need somebody with some business experience in here, and the instincts for how you hire marketing people, which really could be your business.” We didn’t have good luck; a series of nice, great people, but not quite what was needed.

And then, when I was fairly exhausted, and we really had built this wonderful company with a fabulous, quality reputation worldwide—and it took a toll on the family; it took a toll on everything—I could see our going public and thought, “I can’t do that!” I look back now and think, “Well, that was stupid. Of course you could have.” It was just that I wanted to do technical work; I didn’t want to be doing that. So I hired somebody, and I hired very badly. He took the company public. We had a good IPO, it went up, and then pow! It went bad, pretty much because of his bad business relations and decisions.

So I learned a lot. I don’t care if you do an MBA or you’re in somebody else’s company when you learn those things; I just know that when I work with people with a lot more experience in different areas, they bring in the experience and the collaborative learning. For me, I learn a lot more; there’s a little bit less personal stress, because at least you’ve got people with some of the instincts. I mean, I have instincts in some things, but you can’t have instincts in everything.

I probably just learned a lot about team formation. I’ve become very interested in that. I did a project afterwards to understand more about distributed teams and how to help them work together. I acquired more experience with what doesn’t work and what does—because there’s a lot of things that don’t work when you’re remote. Team formation and distributed cognition, where you really have to have all the pieces together, but in a forum where they’re teaching one another, and you’re aware of who knows what, and so there’s a great deal of respect for who knows what and how they can share. So I guess I’d like to go back the twenty years and do it over again with what I know now—although the market’s different.

Abbate:

Wouldn’t we all! [laughs]

Goldberg:

Yes, wouldn’t we all! [laughs] I don’t know; there are some people who never want to go back. I mean, would you ever really want to be a teenager again? [laughs] Some people say Yes, some people say No. It’s just different.

Abbate:

Was ParcPlace doing licensing and support for Smalltalk, or were you developing applications with it?

Goldberg:

We were doing Smalltalk, primarily. That was our skill base; that’s what our engineers knew. The venture capitalists obligated us to do C++ environments as well, and that was our first mistake. Had I had more experience, and more credibility in the business world, I could have talked them out of that big-time mistake; but it was an obligation of the funding. The reason it was a bad idea was simply that the kind of systems that Smalltalk represented were not amenable—were not interesting—to the C++ community. They were a different breed of programmer, number one; number two, we weren’t part of the community. We invented the Smalltalk community. We were kind of the core of it. There were all these wonderful companies around us, and there were competitors, and all of that made for a great market: a lot of competition, and a lot of choices for customers, which they liked; they liked setting companies against each other. It was wonderful. We knew all the people; we knew everything. We didn’t know the C++ world, and you can’t compete in a situation like that. You know, there’s probably somewhere out there some wise business teacher who would tell you, “You only try to do that which you know well.” Focus, focus, focus!

Abbate:

Smalltalk and C++ are both object-oriented, but aren’t they very different otherwise?

Goldberg:

They’re very different styles. C++ is basically a language, where the support around it was very traditional C-like support; whereas Smalltalk was really a language system, and very graphical. The world headed that way eventually, but we were just really too early for those people. We were fine with people who were COBOL people; people who were into more visual languages; who were interested more in simulation. Although we had customers who did amazing stuff with real-time embedded Smalltalk systems, that wasn’t what we were targeting; whereas the C++ had different performance characteristics for different kinds of applications. I mean, you know that about programming languages; one language isn’t going to be for everything.

So that was just a big-time mistake, and it created a lot of mixed messages to our customers, and mixed messages to our engineers. It was very problematic. Our engineers knew what was going on, but that would have been one thing I wouldn’t have done. So there are at least a couple things I would have done differently. But I would never have given up going to Xerox PARC. [laughs] That’s for sure!

Abbate:

I just think if I had to pick a user-friendly language, it wouldn’t be C++.

Goldberg:

No.

A Changing Research Culture: On Industry's Involvement in Computer Science

Abbate:

You were at ParcPlace from 1988 to ‘95?

Goldberg:

Maybe ‘94 as an employee, and till ‘95 on the Board. Very, very unpleasant, the last few years. Not fun at all. I learned that I’m not evil enough. I think that there’s a brand of people out there whose business style—driven by personal greed, basically—they have this evil streak in them, and they come up with the most evil things you’d never imagine, and if you’re not evil, you can’t second-guess them; you can’t avoid it. I just watched the most remarkable, terrible behavior; greedy behavior. I would just as soon never have had that experience. Maybe it makes me wiser, but I don’t think so. I just learned—I mean, there’s a lot of times where I see people in situations where I see them confronting those kinds of people, and I’ll say, “You really don’t want to do that; you’re not evil enough.” I’ve walked into that other times, too, where I can’t plan—you know, it’s like a chess game: you figure, “You do this, we’ll do that”; but you can’t ever second-guess evil in the business world.

To me, that’s a big difference from the academic world. There’s jealousies, and there’s going after grants, and there’s competition, but it’s academic, technical competition; it’s not about money in your pocket. It’s really a group of people who are goading each other on to learn more, and ultimately to share what they learn. The more business has impinged on academics—for computer science, it’s terrible. I mean, are you going into computer science for research? Is this science? Is this academic? Or is this about setting yourselves up so you can start companies and go after money? Industry’s involvement in computer science has really changed the tone a lot. I’ve heard a lot of people complaining. There isn’t a professor—especially in Europe; in Germany I meet them all the time—that’s not involved in some company. You suddenly find out that, “Oh no, this is about our company.” And you sort of don’t know who you’re talking with; what are their motivations? It’s gotten pretty messy. You can say “Well, the physicists are like that too, and the chemists are like that too, because there’s things they do,” but I think they’re cleaner about it. There’s a more foundational notion of what it is you learn, in your staging, so that at least you’re not in quicksand; and there’s a certain obligation on the part of students to know their history. I mean it’s interesting; you do history: I was just in Switzerland, and I gave the Distinguished Lecture, but I also taught the Advanced Object class, which I haven’t done in ages; and you ask the students, “Okay, where does this come from? Who did it?” And they just haven’t got a clue!

Abbate:

Hmm.

Goldberg:

They haven’t got a clue! Some of them do; some of them read it; but most of them don’t. History isn’t very important.

Abbate:

Do they think it started with Java or something?

Goldberg:

Some people do. And they get a little confused about the ordering of results. It doesn’t really matter unless they’re filing for a patent! [laughs] The number of software patents, at the end of the eighties, started springing up, and I hear a lot about this because people come running to see if it’s in Smalltalk. We did an enormous amount of these things that people have done patents in more recently! But at the same time, there’s always a twist, and it’s not always the same thing. Sometimes it is, sometimes it isn’t.

But that push, that business: It changes the tone, enormously. Not always bad! I mean, sometimes it’s really good to say, “I want to do something that someone would care about.” Why write a dissertation and do all that work, if what you learn, no one cares about? But it’s the immediacy of the caring: is it a contribution to the field, or is it just a way to come out with a product? If it’s both, it’s probably ideal. So the research culture has changed a lot.

I don’t know what I would do if I were starting out now, where I would have ended up. This whole question that we were talking about earlier, about women in computer science: I don’t think I’d go into computing. I think I’d probably be more interested in biotech; something where I think there’s a higher impact, and there’s an interesting mixture of what you can accomplish. You’d use computers—I mean, now it’s more that computers are a basis for computation—but compute about what? The biotech’s interesting; one of my daughters is doing physics and nanotechnology, and that’s just fascinating. I don’t know if that’s what I would do, but there’s a lot of other things that just feel more attractive, and I wonder if that isn’t the reason why you don’t get as many women in computer science today.

Abbate:

Now, is that because things have already been done in computer science?

Goldberg:

Well, I look back at the sixties and seventies in artificial intelligence, and I think, “Wow! Some of that’s pretty Mickey Mouse.” It was such a big deal back then! Now, that always happens; you make a discovery and then it becomes everyday. People forget what it was like not to have television, not to have airplanes; and if you’re born in a world where you’ve already gone to the moon, you take it for granted. You weren’t there the day we landed on the moon; that was pretty remarkable! So I think for a lot of people right now, computers are really just tools, and there hasn’t built up much of a foundational science; and the confusion at the universities between teaching and industrial relationships makes it more like you’re going to a form of business school than going into a science. So if your inclination is to do math and science, you’re not as likely to go into computing. I mean, a theoretician doing theoretical computer science: do they do theoretical computer science, or do they do applied math? I don’t know. I haven’t researched the departments recently to know what is the right thing, but you’re seeing a lot of re-treading of ideas.

A lot of kids realize quickly that you don’t need an advanced degree to make money. I’ve been through this with people: “Why do you want your Ph.D.?” And I’ve had people say, “Well, I want to teach in a university.” That’s a good reason, because you need that; it’s a good ticket. Or someone may say, “I’m interested in pursuing a certain idea, and the best structure under which to do it is a Ph.D., with that kind of advice.” Well, that’s fine. But if you’re going to do it to get respect or to make money? Forget it! That doesn’t make any sense at all. It’s not a ticket! So that doesn’t make any sense. I even wonder why the men are doing it, to tell you the truth. I haven’t looked at the population, but there is an increase, especially undergraduate computer science; and that’s because it’s a required tool. Just as, in K-through-12, you have to do English and you have to do your language skills, you just can’t go without the computing—not and live in this world. But that’s different. That’s vocational, see? That’s back to the beginning in our discussion here!

Abbate:

That’s what you were doing.

Goldberg:

That was what I thought I was doing, was going and getting vocational training! [laughs]

Abbate:

You were just wrong. [laughs]

Goldberg:

Yes.

Consulting: Neometron

Abbate:

I guess we should talk about your current business, Neometron.

Goldberg:

That’s just an umbrella for a consultancy. That’s not a particular business. What I’ve been doing, the last few years, is I’ve been very interested in knowledge management and how teams work, and those interests show up in a lot of different companies’ needs, so I’ve done consulting. In this economy that’s not as much fun; there aren’t as many start-up companies. So I’ve done Advisory Board roles and Acting CTO roles, which I enjoy a great deal. What I liked about ParcPlace was that there were so many different customers in different industries, and I could learn all about the different businesses, and that was really a gift. This is another way to do that.

But I’m also back to educational technology. I had been on the Board of a company that did CD-ROM–based college supplementary materials, and we were moving it on to the Internet when they ran out of money. But there was a project that was started there, early; it’s really professional development for teachers of advanced placement, but it’s also courseware for the students. The target is to make sure that no child, because of their race or location where they go to school—no matter how remote they are—is denied access to college preparatory courses, starting with advanced placement, for which there’s a national standard; but also trying to really focus from the angle of the capacity of teachers to be able to teach these classes, and to give the teacher a support network. I’m interested in community-based knowledge-based systems, and in this particular case it’s helping teachers. We have foundation money to deploy this fall, and we’re primarily in Texas; this is a joint project with the Charles A. Dana Center at the University of Texas, Austin, which is a phenomenal group of people in educational policy and professional development. They are really responsible for the authoring of calculus and statistics courses—those are our first two courses; and then we’re also partnered with the AVID Center in San Diego, where they will do the history and the English literature courses. And we have, out of Neometron’s umbrella, designed a whole authoring process, an XML-based way to specify all the curriculum; and a learning-management system for delivering it in the context of a community, so that teachers can mentor one another and be mentored by experts who are Master Teachers, and do this remotely. Part of the foundation grant obligation is to teach fifty classes of calculus and fifty classes of statistics. We started very late, and we’ve signed them all up now; but teachers start the first of August, and school starts mid-August, and we’re not done! That’s why my time’s a little limited right now.

Abbate:

Well, it sounds like you’ve been there before!

Goldberg:

Yes—but I don’t like it! I don’t feel good about putting out untested software. I learned that a long time ago! So we’ve got some work to do.

On the Rewards of a Career in Computing

Abbate:

What have you found to be most satisfying about working with computers?

Goldberg:

You know, that’s actually an ambiguous question—because there’s the satisfaction of what you do as an individual sitting in front of a computer, doing things; but there’s also the satisfaction of the consequences of what you do. Certainly the consequences of what I’ve done would take precedence, because it really has been an interesting adventure.

The satisfaction comes from being part of a really creative group of people who not only invented an idea, but made it real enough that other people could take advantage of that idea in a way that was enormously empowering. The number of people who said to me, “I’ve always wanted to be able to do XYZ; now I can do it,” was just amazing! And what they did was not predictable. We didn’t sit down and say, “If we create A, then other people will be able to use A, and they will create XYZ.” We never could have guessed, never could have predicted what XYZ turned out to be, because we didn’t know the industries; we didn’t know enough; we weren’t domain experts. So the domain experts—our customers—taught us; and that feedback loop, which allowed us to improve our systems and then that helped them, was just a wonderful way to evolve your business, involve your thinking, your life. That’s very satisfying. You don’t get pigeonholed in any way; you’re always learning.

Serving as President of ACM

Abbate:

Was there anything specific when you were President of the ACM that you accomplished, or were trying to accomplish? I’m not really sure what the President does.

Goldberg:

[laughs] What does the President of the ACM do? The President of the ACM is actually supposedly the Chief Executive Officer during that period of time—which is funny, because it’s an elected volunteer job, while there is a professional staff. Depending on the President or depending on the decade, the staff is either given loose reins or tight reins. I came in specifically at a time in the mid-eighties when, like now, the economy was down. You couldn’t just spend against budget unless the revenue was coming in. We needed to tighten up and tighten up a lot and, for some reason, I knew how to do that. So that wasn’t very pleasant, but that’s what we did. If I remember correctly, I was the second woman President of ACM. The first was Jean Sammet, and Jean ended up in the same situation where there was a belt-tightening time, which is kind of funny.

That’s an interesting person. Have you talked to her?

Abbate:

I did interview Jean Sammet.

Goldberg:

Yes.

What did we do during that time? One of the issues at hand was that ACM was too much of a secret. In Europe it was much more respected, but in the United States still a secret, and we wanted to have a lot more visibility. One of the ways we thought of to have that visibility, consistent with our mission, was to have an ACM Press: have our own book series where ACM’s name would be on the shelf. So I started the ACM Press. I started the History of Computing Series, where we did about five conferences, and the very first one was the History of Personal Workstations, and so I edited the book, sort of setting the tone for what would happen in the subsequent books.[3] It was a good project, although an expensive project. History books do not always sell: people don’t read history, and history books don’t sell broadly, but they’re really important things to do, as you know. We formed a partnership with Addison-Wesley to be the distributor/publisher, but it’s an ACM Press book. As editor, I received royalties, but I returned all the money to ACM to be used to fund ACM memberships for female computer science students.

There was a time of enormous turmoil, in the eighties, about the good and the bad about computing. I used some of the President’s resources to start the Risk Forum that Peter Neumann did; I gave him the money for that. When you want to do something new, there’s a pot of money that the President has that you can use outside the budget cycles.

I did decide that I had to tighten up staff and get it reorganized a bit, and I think we did a good job of that, and holding forth on not losing our shirt. I pulled us out of the AFIPS Office Automation Conference, and I pulled us out of the National Computer Conference. The President of ACM is on the AFIPS [American Federation of Information Processing Societies] Board. What basically happened is that after the last Office Automation Conference in Dallas, they were talking like there was going to be a next one, and I came up with a scheme of three different criteria that, if any one of them occurred, would signal that this was a dead conference and we shouldn’t go on, because we were going to lose money. Everyone agreed that those would be the criteria, but no one at the table believed it would happen, except ACM. All three happened! So that way, you didn’t have too big a debate. With the National Computer Conference, the writing was on the wall as well. It was a situation where a lot of small societies depended on the revenue from the NCC for their income, and they didn’t want to pull out and shut down. And so we didn’t talk about pulling out and shutdown; we acknowledged that this was an important source of revenue for them, but that the ACM was concerned about its finances and would like to withdraw its ownership position and distribute it to the rest of them, who could then share in the liabilities and rewards. I remember the eyebrows from the DPMA [Data Processing Management Association] and IEEE when we did that, but you see, that caused everyone to look a little more closely. So that’s not a wonderful thing, but we got out before it was too late, which was a good thing.

Specialized conferences were more likely to be successful, rather than these big umbrella ones for which you had very few choices of locations, and you ended up in the heat [laughs]—the heat of Houston or wherever! The guys in my lab started one. Remember, this was the birth of various object-oriented projects—the Smalltalk books were out; there were other people doing object-this-and-that; there was Trellis; there was C++; there were a lot of interesting research projects getting started—and I remember Glenn Krasner, who worked for me, coming to me and saying, “We want to start a conference.” “We” was a lot of these people we had worked with on those books, from Tektronix and Apple and Hewlett-Packard and DEC; and I said, “Well, that’s good. Now, who’s going to guarantee the hotel?” And they said, Well, they thought I would. And I said, “Did you mean the Research Lab at Xerox would, or did you mean ACM?” His answer: “Oh, whichever!” [laughs]

I thought they were great! And so I thought about it and realized this was not going to be a Xerox thing; that wasn’t right; and we were out of the budget cycle for the ACM special interest groups, which should have run it. So I made them swear that they wouldn’t lose any money, because we were really tight, and I gave them the money they needed to start OOPSLA, which has been going on since ‘86 and has always, I believe, been profitable. After the first year, SIGPLAN took over managing the conference.

For some of it, it was just a lot of work to deal with the financial situation and to know what the right thing was to do. We ended up changing Executive Directors, but that had been planned for some time. We had been working a lot with the IEEE Computer Society—I think the past President, my predecessor, David Branden, had really worked with the IEEE Computer Society President to see if we could just merge the groups, rather than be competitive—but they have a one-year turnaround for their President, and this woman came in, and that just wasn’t her thing. So I said to Dave that I wasn’t going to concentrate on that, because it was too painful to try. You know, we tried to do tutorial stuff together, which was a failure.

With John White, who was at Xerox and head of the SIG Board (and now the Executive Director of ACM), we revamped how funding went to generic activities. ACM core was putting out a lot of money for very-long-standing projects that never were having proper reviews, and we came up with the idea of just stopping funding them and letting them apply—in an NSF grant application sort of style—to the SIG Board, because the SIGs would provide funding. So we changed the financing structure and the review structure for some of these projects. That included a wonderful education project, which was a set of slides and instructions for people to go to school and explain about computer science; career counseling stuff. They thought, “Well, you can’t just cancel that. It’s important!” And I said, “Then the SIGs will pay for it,” and they did. I said, “If the project’s worth anything, they’ll pay for it,” and they did. So we changed the tone of financial decision making a bit. Two years isn’t very much to do much.

Abbate:

I was going to say, that sounds like a lot accomplished for two years.

Goldberg:

Yes, it was; but again, my mission was simply cleaning up the finances, not doing all those other things. And my mission was to stop being Editor-in-Chief of Computing Surveys, which was a lot of work! [laughs] A wonderful thing to do, but a lot of work.

Advice for Young Women Who Are Considering a Career in Computing

Abbate:

Do you have any advice for young women who might be thinking of going into computing?

Goldberg:

I think I said this earlier, and I’ll just say it again: Don’t go into anything that you don’t get some work experience about, because you really don’t know. You don’t know if it’s really right for you; you don’t know if you’re really interested in it. It’s true that the kind of things you’ll do and the jobs you’ll have as a part-time worker might not be very exciting, but by the time you’re a senior in high school you should be able to get some of these work-study jobs.

My second piece of advice is: knowing about computers, and knowing about math, is just a basic skill these days, so you need to learn something about both. You can’t not do it because socially it’s not the thing to do; you’ve got to find a way around that, one way or the other. It’s hard to say that to teenage girls, but they do need to stay focused on these things. So their parents need to help encourage them and make math and computing be fun. I’m not sure how much it goes on now, but you know, “nerdy” girls never were very popular, so that’s always been an issue.

And then there’s this one piece of advice I remember giving when school started at Reed College one year. I’d gotten an award from them and gave a speech about a lesson I had learned from a guy I worked with at Xerox PARC, Bob Flegal. Bob was always drawing, and he told me if you really want to learn to draw, every picture you start, you have to finish. For me, every project you start and every course you take, you finish it. You don’t just quit in the middle because it looks like you’re grade’s going to be bad; you do something about that; because you have to get in the style of realizing that when you start something, you’re making a commitment, and you finish it. That doesn’t mean, “Do it for the rest of your life.” Your dissertation doesn’t have to be your life work, as I’ve told many a student. It has to be something that, if you do it, you ask a question that people would be interested in and want to know the answer to, a question whose answer makes a difference; but it doesn’t have to be what you’re going to do the rest of your life. But you do have to finish it! There’s just that notion, and I think that’s important.

Abbate:

Good advice!

Now, you have two daughters?

Goldberg:

I have two daughters. Dennis has two, so I have two stepdaughters. My oldest is here in the local area, and she’s Director of an emergency shelter for homeless families that also does training and transitional housing. She’s actually helping me on this educational project as well. She’s finding it really fun and interesting. And then my younger daughter just finished her second year in graduate school. She’s doing a physics Ph.D. at Wisconsin. Amazing!

Abbate:

Great. Well, thank you so much for talking with me.

Goldberg:

You’re welcome!

Notes

1. Adele Goldberg, “The Community of Smalltalk,” in Handbook of Programming Languages, Vol. 1: Object-Oriented Programming Languages, ed. Peter H. Salus (Indianapolis, IN: Macmillan Technical Publishing, 1998).

2. Alan Kay. "The Early History of Smalltalk," in History of Programming Languages, Vol. 2, ed. Thomas J. Bergin and Richard G. Gibson (New York: ACM Press, 1996).

3. Adele Goldberg, ed. A History of Personal Workstations. ACM Press History Series (New York: ACM Press, 1988).