Oral-History:Laurie Keller

About Laurie Keller

Laurie Keller was born on December 15, 1945 in Chicago. Her family moved to California where she grew up and attended school. After attending University of Southern California and Berkeley, she began working as a computer programmer handling support and programming for several companies, including Electronic Data Systems, TRW, Iraq Petroleum, and Texas Instruments. Keller was one of the few women in the field at the time. She also worked for Continental Airlines as Manager of Training and Standards. She eventually moved to Britain where she became the first women lecturer in computing at Open University. She achieved Chartered Engineering status and earned her Master's in Computer Science while teaching at Open University.

In this interview, Keller talks about her childhood and influence of her family on her career. She also goes into depth on each experience working as a computer programmer for various companies. She ends the interview with her description of the work and research she has done at Open University. Keller also provides a reflection on the status of women in computing and her advice for young women thinking about entering the field.

About the Interview

LAURIE KELLER: An Interview Conducted by Janet Abbate for the IEEE History Center, 26 September 2001

Interview # 606 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 at Stevens Institute of Technology, Castle Point on Hudson, Hoboken, NJ 07030 USA. 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:

Laurie Keller, an oral history conducted in 2001 by Janet Abbate, IEEE History Center, Hoboken, NJ, USA.

Interview

INTERVIEW: Laurie Keller
INTERVIEWER: Janet Abbate
DATE: 26 September 2001
PLACE: Laurie Keller's home in Bedford, UK

Background and Education

Abbate:

This is an interview with Laurie Keller on September 26th, 2001. [This interview replaces one that was recorded but later lost, so we will be revisiting] some old territory.

So, just to go back to the beginning: If you would just say briefly when you were born and where you grew up.

Keller:

Right: I was born in Chicago, the 15th of December, 1945, because my father was stationed with the Quartermaster Corps there; but my family was living in California at that time and went back to California when I was just a few months old, and I grew up in Southern California, in the Los Angeles area.

Abbate:

Your parents were both accountants?

Keller:

My parents were both accountants by training, yes.

Abbate:

But your father worked for the Army? What’s the Quartermaster Corps?

Keller:

The Quartermaster Corps are the people who buy and disperse food, blankets, munitions, all that sort of stuff. So he was largely responsible for things like purchasing, particularly meats—that’s why it was located in Chicago. Meats, but also things like blankets, uniform material, and so on and so forth.

Abbate:

So you just moved—you didn’t [spend any time in Chicago].

Keller:

No, I have no memories of Chicago other than going there as an adult.

Abbate:

And your mother stopped working when she had you?

Keller:

Well, that’s actually interesting, because I came late in my parents’ life; they’d been married for thirteen years. My mother was still working, and she was going to stop work on the 24th of December, go back to California, and have me. The only trouble is, I came on the 15th! [laughs.] That was a Saturday, and she was actually working on the Friday, and thinking she had indigestion! [laughs.] I sometimes jokingly say I was nearly born in an office. Not quite, but . . .

Abbate:

So then she stopped working.

Keller:

She stopped working, yes.

Abbate:

So you’re the oldest.

Keller:

I’m the oldest of two. My brother was born in 1949.

Abbate:

Now, you had mentioned before that there were other professional women in the family. Your mother’s sister . . . ?

Keller:

My mother’s sister was a doctor, yes. She was an obstetrician-gynecologist.

Abbate:

And did she live in California?

Keller:

No, I never knew her. She died before I was born.

Abbate:

Oh.

Keller:

But she was always held up by my mother as a suitable, shall we say, ambition to aim for.

Abbate:

And your father’s mother had also done something?

Keller:

Well, she’d been widowed quite young, and she’d worked at various things, including being a cook in lumber camp in Minnesota; but she settled down as being a comptometer operator at some point.

Abbate:

Did you know her?

Keller:

Well, I have pictures of me standing next to her, [but] I have no memories of her. She died when I was about four.

But my mother’s mother, who was a farmer’s wife: after they moved off the farm, she supplemented the family income by what you might call “practical nursing.”

Abbate:

So it was always expected that you would have some sort of career to support yourself?

Keller:

I was quite shocked when I realized that some people’s mothers didn’t work! [laughs.] And that must have hit me really in my teens; because almost all the girlfriends I had, their mothers did something outside the home, or they took in children and were paid for it; so I just thought everybody worked. I never had a concept of somebody not doing paid employment.

Abbate:

And you had ambitions of your own?

Keller:

Yes, I did, I suppose. Well, my parents had always emphasized to me that it was really important to have a career of some kind—as a backup, if nothing else. There was always this: “If you have this, you can take care of yourself, no matter what happens. And who knows what happens in life?” That was their approach. I’m not sure that I quite understood it at the time, because as I say, I saw all these women I knew [who worked]; I was in contact with female teachers and nurses and so on. I just didn’t realize what the flip side of what they were saying was, this economic dependence thing; I really didn’t realize that. I just thought everybody did these things.

I had a number of different ambitions—you know, the usual things when you’re little. I wanted to be a fireman [laughs]—various sorts of things like that! But what I finally settled for when I was in my teens was, I thought I’d like to live an academic life. I’d like to be a writer, write lots of things. I realized it was probably difficult to support myself doing that, so I thought, “Academic life is the way you do that.” And that would give me the chance to spend lots of time in England—because I was very interested in English literature and so on, and I thought that would be ideal: I’d get a job in some nice little college teaching English, and then once every seven years I’d take my sabbatical and go spend lots of time in England, digging around archives and old libraries.

So, I got here [to England, and] I write a lot, but it’s not at all what I planned initially! It was purely by chance, I think, that I ended up being where I’d initially thought I’d like to be—but in a completely different capacity.

Abbate:

You were going to write fiction?

Keller:

Oh yes! Yes. I was going to be a great poet! [laughs.]

Abbate:

What were you reading that inspired you?

Keller:

I’ve always been particularly fond of 17th-century English literature, so it was John Donne, Herbert, Ben Jonson, Shakespeare, things like that that moved me. I’m not a great fan of 18th or 19th century literature, particularly. The 20th-century American literature I was fond of, so another possible ambition was to write the Great American Novel! [both laugh.]

Abbate:

You went to regular American public schools?

Keller:

Yes. As it turned out, the high school I went to was very unusual. It was Dorsey High School in Los Angeles, and at the time I went, which was 1960 to ‘63, a lot of demographic change was happening in that area in Los Angeles. The school actually had a very diverse student body: so about a third of it was black; about a third of it was oriental, which was split between Japanese and Chinese, with a sprinkling of Korean; there were a few Latinos; and the white population of the school was almost evenly split between Jewish and non-Jewish. So it just had this enormously diverse student body, and that in itself was a real education. The other thing was that it had some teachers, most of whom were either World War II veterans or who had obviously been much changed by the war or the Depression, who were really, really interested in pushing the kids as far and as fast as they could go. I didn’t really realize this at the time, because it seemed like a perfectly normal education, but we really were pushed quite hard. I actually found that I could coast at university, because of the high school education. I had such a good background that I could not work too hard for the first year or two at university, just because they’d covered much of the materials, the concepts, and the skills that I needed.

Abbate:

What were your favorite subjects?

Keller:

Everything! [laughs.] I loved science. I did Spanish as a foreign language, and I really enjoyed that. I really did love English. I loved to write; I always have. I mean, from the time I was a little kid I used to write little books and illustrate them myself and staple them together! [laughs.] So I loved English. I think math was about the only thing that I really didn’t like. I did well at it, but very early on I decided it was designed to torture children! [laughs.] I always looked at it as a kind of hurdle that was put in your way, that you had to get over in order to go to college, and I never really looked at the value of it in its own right—until, I would say, only about ten years ago, when it really started sinking in to me how valuable a skill and a knowledge area it is. I did it, but I hated it! [laughs.]

Abbate:

Now, one of the questions that interests me is how important mathematics is for success in computer programming and in computer science. Do you think it’s the central key skill, or do you think there are other things that are equally applicable?

Keller:

I think there are other things equally important or maybe even more important, one of which is linguistics. Abstract linguistic skills: the understanding of grammar, for instance. I now begin to realize that it has lots in common with discrete mathematics. For instance, the types of notations and types of manipulations that you would use in various sorts of grammars are quite similar in a way to discrete mathematics, in that the kinds of things that you are trying to do—like show how things belong to a set, or how sets can be combined and so on—all come within that thing. I actually came at it through grammar, which was something I always enjoyed at school. I think I was the only kid I knew who just loved parsing sentences and doing sentence diagrams and things like that. And there has been some research in Australia that shows that in fact the correlation with that kind of slightly abstract linguistic skill—in other words, understanding a grammar—has a closer correlation to success in computer programming than mathematics does.

Abbate:

Wow, I’ll have to get a citation for that.

Keller:

I’m not sure I can get you a citation very quickly, but it’s there somewhere.

Abbate:

Actually, it seems in a way what you’re saying is that the distinctions between mathematical skills and linguistics skills are maybe not as sharp as people would have you believe.

Keller:

I think it depends on whether you’re talking about discrete mathematics or not. Discrete mathematics is quite different, and there I do think that there is something common. I haven’t really yet worked out exactly what it is, but I’m beginning more and more to feel that this notion of classifying things into sets and being able to manipulate sets according to logical rules is the common thing between discrete mathematics and grammars in linguistics. So there is something common there, but with linguistics you can arrive at it—an understanding of it—in a way that doesn’t require you actually learning the formalism so much as internalizing the sets of rules and realizing, “Yes, there are sets of rules, and I can apply these in different ways.” For instance, children learn to speak a language pretty well by the time they’re three, without necessarily knowing what a noun is and what a verb is, and so on. But when you learn what a noun and a verb is, you can play little games by filling in different things, and then you start looking at complex sentences and wondering, “Well, how do I deal with this sort of thing?” And you can deal with it in this rather abstract way. Not everybody gets there, but if they have that understanding of grammar as a set of rules and grammatical objects as things that you can manipulate, I think they’re very close to what discrete mathematics is doing.

Abbate:

It sounds like you enjoyed school very much. Then you went to college at Berkeley. Now, you were living in L.A.?

Keller:

In L.A. I actually started college at the University of Southern California. The usual story for a seventeen-year-old girl: I met a boy, who was at UCLA, and he was just graduating, and he was going to go on to do a Master’s Degree at Berkeley; so I transferred to Berkeley to be with him. So I started in Berkeley my second year, really.

Abbate:

And then went on and got an English degree, as planned.

Keller:

Yes; yes. [laughs.] Then went on and got the degree as planned.

Introduction to Computer Programming, FORTRAN and COBOL

Abbate:

Now, you’ve said that your first experience with computers was actually through this boyfriend of yours at Berkeley.

Keller:

Yes, that’s right.

Abbate:

And he was an engineer?

Keller:

Yes. He’d graduated in engineering from UCLA, and when he went up to Berkeley he was actually doing what was then a very new degree, the MBA. But there was a free course put on—meaning you didn’t have to pay a fee for it—by the engineering school, and he went for a refresher; it was in FORTRAN. He’d mentioned that was going to go to this course in FORTRAN, and I said “What’s FORTRAN?” And he said, “Oh, it’s a programming language, for computers.” And because I was so focused on language, this really interested me: How could you possibly have a language to instruct a machine? I really wanted to know, “What does this language look like? What kind of grammar does it have? What kind of syntax does it have? What kind of semantics does it have?” So I went along and sat in this room with him and went through the little course, and I wrote myself a little FORTRAN program—which I seem to remember did something like add two numbers. Something really exciting! [laughs.] But it did give me a feel for what programming was.

Abbate:

And you found it easy?

Keller:

Yes! In fact, compared to a natural language, it seemed to me really, really simple—which I suppose it is, in a sense. It’s not quite so free-form as natural language, and it has lots of distinctions. In natural language, people listening to each other will skip over errors—they’ll just sort of understand what the meaning is, without actually having the grammar and the syntax perfect—and of course, machines don’t do that to you. One of the humbling things is writing your program and then getting back great piles of error messages because you have misplaced or misunderstood one single thing in a program. It’s a humbling experience!

Abbate:

Did your boyfriend go on to do anything with computers?

Keller:

Oh, yes. Well, he eventually became my husband. He did his M.B.A., took up a job with the Jet Propulsion Lab down in Pasadena, and we married and moved down there. He was a systems analyst for JPL, looking into the failures of some satellites to deploy once they got up there. That was the first thing he did; but it always revolved a bit around computers, and he’s still involved, in the sense that he now owns a small software company that produces educational games.

Abbate:

Now, you got your first job at the L.A. Philharmonic?

Keller:

Well, yes. My first real job was: I was doing a Master’s Degree, and I’d got about halfway through it, and [I was] getting short of funds, and I didn’t want to keep asking my parents and stuff. I had been working part-time, and I thought, “Well, I’ll swap from being full-time student, part-time worker to the other way around.” I applied for a number of jobs and finally got one with the L.A. Philharmonic, initially having nothing to do with computers. My first job with them, which only lasted six weeks, was that I was supposed to use my research skills to go and find wealthy people connected to companies which were making large profits, who needed to get some sort of tax relief on those profits, who might donate money to the Philharmonic. So that was my first job, and I was only doing it for about six weeks.

At the same time, the Philharmonic plus some associated companies were computerizing their ticketing—the season ticketing in particular—so that when people bought tickets for a concert series, this would all be handled through computing. The idea was eventually that there would be real-time ticketing, so people could walk down to the supermarket and say, “I feel like going to a concert tonight,” and be able to get a ticket. That never happened, really, but the keeping track of people’s season tickets was being done at the time.

The arm that I worked for, the fund-raising arm, thought, “A lot of these season ticket holders are also donors, and it would be useful to piggy-back putting the donors’ records on computer files.” The woman who was doing this left, and I was the only other person in the office who had a degree, so they came to me and said, “Would you be interested in taking this woman’s job?” Well, it was more money, and it sounded more interesting than spending my time going through lifestyle magazines [laughs], and so I took it. This involved me not directly with computers, but with the company which was doing the work, which was an arm of Computer Sciences Corporation. And I suppose because I had this familiarity with programming, I was probably able to do a reasonable job with this, because I understood such things as why computers might mix up one record with another if you weren’t careful to distinguish them, and so on and so forth.

Anyway, I was involved with that for about six months, and then that came to an end. Everything was, then, computerized, and so the job kind of disappeared. The Computer Sciences people came to me and they said, “You understand how the system works, because you’ve been using it, and you still speak English; so you can write user manuals!” [laughs.] It was a chance to write, which was something I’d always wanted to do, so I thought, “Yes, all right; I’ll take that on.” So my initial job with them [in 1968] was actually as a technical writer. But they really needed programmers, and the programmers were so busy they couldn’t explain how the programs worked, so they’d just give me the listing and say, “You figure it out.” And after a while of doing this, I thought, “Yes, I can write these as well as be reading them.” And they weren’t particularly bothered that I would take up what was then very expensive, precious computer time to try out my own little programs; so I just started writing little programs.

Abbate:

This was in COBOL?

Keller:

It was in a combination of COBOL and assembler language, because some of the stuff I was working with was in COBOL, and some of it was in assembler language, and I had to teach myself assembler language to learn how to read the listings. Obviously I had to teach myself COBOL as well, but COBOL, looking a lot more like a highly structured English, was a little bit easier to work out what was going on. So I taught myself assembler language through an IBM Programmed Instruction Text, which was Teach Yourself Assembler 360. Something like that.

Abbate:

Was that a typical experience, do you think? That people had to sort of teach themselves, because there wasn’t time to train?

Keller:

It’s hard to know. Of the few women I knew then, most of them had come into computing in a very similar way to what I did. One, for instance, had been a secretary for a programming group, and they’d given her the job of documenting programs, and finally she thought, “I’ve read enough of these. I think I can write one.” And she eventually became a programmer. In fact, I knew two women who’d come through that sort of route: one who’d worked for NASA in Alabama, and one who had actually worked for Computer Sciences Corporation and done that. Another woman had started out working for Goldwater’s Department Store in Arizona. When they started computerizing their stock-taking, they sent somebody off to learn how to punch cards, and the woman they sent didn’t like it, and so [this other woman] said, “Oh, show me how, and I’ll take it up.” Eventually she worked her way from being the key-punch operator to being a computer operator. Then she and her husband had moved to California, where the law said that she could not work as a computer operator, because it required lifting more than 35 pounds—because a box of paper, one of these continuous-form things, weighed more than 35 pounds. So she couldn’t work anymore as a computer operator, so she started doing kind of ancillary work, which was things like setting up jobs and organizing things, and she taught herself to program as well.

So we all came into it through these rather strange and, in many cases, either self-taught or self-defined routes. I have a feeling the men tended to come either from engineering . . . One I knew had been a mathematics teacher, who’d actually retrained with the Control Data Institute. I think the men were probably more likely to come with some kind of formal teaching, and the women more likely to pick it up by guess and by gosh. I mean, that’s just my experience of one place and one time, but that would seem to be—looking back at the people I knew then—that was how they got into it.

Abbate:

So you ended up programming for . . .

Keller:

I ended up programming for Computer Sciences Corporation. The little [group I was in]—first it started out as a division, then it became a subsidiary, and then it became a free-standing company, and then it went bankrupt—was called Computicket. I had started when it was a division, and then it kept getting more and more distant from the parent company, and the parent company finally ended up pulling the plug. But it was through them that I learned all the basic skills and went to being a programmer. When they went bankrupt, the L.A. Philharmonic plus these other allied companies, which were called the Music Center Operating Company, in Los Angeles: they could no long live without their computerized records, of course, because everything now was on computer, and there was no paper that they could adapt to going back to hand methods. So they bought the software, and they hired me and one of the other women to run it for them. So we started a little computing department of two (plus a boss) with the Music Center Operating Company.

At that time, I knew that my husband and I were going to relocate to New York at some point, so I knew that my time there was finite. One of the first things I did was train someone to take my place at Music Center Operating Company. He was, in fact, a flautist, who wasn’t making enough money as a flautist and was interested in some kind of job that he could use to keep him and his family; because his wife was also a musician, and it’s a very precarious existence, so he was interested in establishing himself in something permanent. That was my first teaching job, as such: teaching him how to program in COBOL, so he could read the programs and continue to maintain them.

Moving to New York and Working for Electronic Data Systems

Abbate:

And then you moved to New York.

Keller:

And then I moved to New York. We’re in 1970, and my husband took a job with Mobil Oil in New York, and we moved to New York. Of course, since it was his job that we were moving for, then I had to look for a job when I got there. I found a job with Electronic Data Systems—Ross Perot’s company—working on Wall Street. They had the contract for facilities management, which meant they took everything from getting the data on paper—or over the wires, in this case—right through to producing the reports and responses. It was all run by EDS on behalf of the second-biggest brokerage in Wall Street at the time. I worked there for a year.

Abbate:

As a System Engineer?

Keller:

As a Systems Engineer, because that was the title they gave everybody. That was interesting, because they really expected you to know everything that you needed to know: so you had to know something about hardware, you had to know something about programming, you had to know something about software development, you had to know something about systems analysis. You really had to be a jack of all trades, in a way. But they did provide really excellent training, and it was the first time I had any kind of—I wouldn’t call it “formal” training, because it wasn’t with IBM or something like that, but more experienced people would stand up and offer a course. The way they did this was, they would do it four nights a week: the company would give you a couple of hours, and you would give a couple of hours, so it would start at four and you’d stay until eight, and it was every other week. They were really intense, but they were very, very good. I learned a tremendous amount. I was only there for a year, but I learned a tremendous amount.

My job demanded this, because specifically I was responsible for application performance, which was also kind of a quality assurance thing. Every time somebody had modified a program or written a new program, it came to me. In those days, things like how much memory you used mattered, so I had to ensure that the code was as tight as it could be, in terms of memory use, in terms of time that it would take, and so on. So the expertise I developed there was how to write, or optimize, to get really tight code. Also, I had to ensure that it was adequately documented and that it worked as documented—at least as best we could tell—and to keep the library of all the applications programs and ensure that they worked with each other, so that there weren’t conflicts between different ones. This covered a whole wide variety of equipment as well, because the brokerage from which EDS had taken over their equipment had everything: it had three or four 1401s; it had a 7074, which had a drum on which sorts were run; it had a couple of 360s; it had two CDC 6600s. So it had an enormous variety of equipment, and I really had to get to get to grips with all sorts of different equipment, whereas before I’d been only working on IBM mainframes.

Abbate:

These were mostly IBM machines?

Keller:

It was mostly IBM machines, except for the CDC 6600s, which were the ones that did the online order processing, where it came in from brokerage offices all over the U.S. and would be sent electronically. Of course, they used to break down all the time! [laughs.]

Abbate:

Which did?

Keller:

The CDC 6600s. There was a pair of them; one would be the active one, the other would be the backup. But the active one would go down, the backup one would start up, and then most of the time it would it would collapse under the load as well. [laughs.] The backup to that was, they had a huge room with banks of phones, and people would go rushing in there and go back to the old way of taking the orders over the phones and writing them on little slips of paper, which were then put on little conveyor belts and went round for hand processing. Because of course, with the stock market you couldn’t just say, “Wait a minute! We’ve collapsed; you can’t do any trading.” You had to keep going in some way, so the ultimate backup was this room with all these phones and the little rubber conveyor belts and pads of paper. It was an interesting experience. One of the things I learned was, “Never throw away the old system until you’re absolutely certain that the new system works!” [laughs.]

Abbate:

Maybe not even then!

Keller:

Maybe not even then, yes.

Attending IBM User Group Meeting, SHARE and GUIDE

Abbate:

You mentioned at some point that you had gone to IBM User Group meetings, like SHARE and GUIDE. I don’t know if this was when you were at EDS.

Keller:

No, that was later. From New York I went to England; then, I was here [in England] five years; then I went back to the States. It was when I went back to the States that I went to SHARE and GUIDE, because by then I was a Systems Programmer for TRW. I’d go to SHARE and GUIDE, and that was always an interesting experience. One [reason] was that there were very few women there, so if you were a woman you got lots of attention—some of it not particularly welcome. The other thing: I would have loved to have had a video camera or a recorder at the time, because one of the things that really amused me was the way men from different companies would sort of [boast]—it was this “mine is bigger than yours” thing! [laughs.] You’d get two men meeting, and they’d start talking about their equipment, which was why they were there: their IBM setup. So you’d get one saying, “We’re running twin 168s, with disk farm of so many disks and so many tape drives”; and the other one would say, “Ah, yes. Well, we’re running three 168s,” or something. Or, if they couldn’t top the equipment, then they’d say something about the complement of operators that was required to run this; so there was this funny kind of competition about how big and how complex your system was. Of course, most of the people who were going to be there were going to be big companies with big systems—big insurance companies, big aerospace companies and stuff—who were running quite large computer installations anyway. But it was just this rather funny thing, and I always wondered if there were people who came from small companies with little systems who kept very quiet because they felt embarrassed or something! [laughs.]

Abbate:

Was it useful otherwise?

Keller:

SHARE and GUIDE were really user organizations: partly to hear what IBM had planned, partly to share experiences and problems and fixes for them, and partly to act as a voice by IBM users to IBM about what they wanted, what their problems were. So that was the reason why these places existed. They normally went on for three days or so. But there was also a lot of this socializing and macho posturing to go with it! [laughs.]

Moving to Britain

Abbate:

But I guess we’re jumping ahead. So you were at EDS for a year and a half?

Keller:

Just a year. The reason it was only a year was: six months into our time in New York, my husband was transferred to Britain. Initially he was told it was a six-month contract, so, since I had a good job and we had a nice apartment, I stayed in New York while he came to Britain. And then at the end of his six months he was told, “Well it’s going to be another six months.” And we had to think about this, and we thought, “We can’t keep going on like this, trying to live two sides of the Atlantic”; so I packed in the job with EDS and came to Britain and started looking for a job here.

But of course, I thought I was only going to be here for six months, and I didn’t want to be dishonest, so I was looking for something short-term. I found a job with, of all places, Iraq Petroleum Company, who were looking for somebody to do some conversion work; and since that was a finite time, that seemed a reasonable thing to do. It was rather interesting, because the week that I accepted the offer, the Iraqis nationalized about a third of the assets of the Iraq Petroleum Company! —which made me wonder, because I was waiting: I had to wait six weeks for a work permit, and I was sitting there thinking, “When I finally get this work permit, is there actually going to be a job to go to?” But there was. In fact, there were some interesting outgrowths as a result of the Iraqi nationalization. We had to redevelop a couple of systems, because the Iraqis had just nationalized the computer, the software and everything, and kept it there, and we had no access to it, so we had to replace it in some way.

Anyway, I got there, and it turned out that the job actually fitted in quite well with what I’d been doing at EDS. Iraq Petroleum had been using an agency that did all their actual computing. Iraq Petroleum had a couple of programmers, but everything was sent to be processed at a . . .

Abbate:

A software service bureau?

Keller:

Yes, it was a service bureau type of arrangement. The service bureau, of course, had CDC machines that they were running [the programs] on, and they were large, so the size of the program actually didn’t matter too much. What they bought to replace it was an IBM 1130, which had a 16K memory; so part of the problem was actually working out how to take 200K programs and fit them into 16K memory—of which you could actually only use 11.4K, because the rest of it was taken up by the operating system. So once again, it was this business of having to . . . I think what it did for me was, I had to learn a great deal about the way systems operated, inside the box. I had to understand how memory was used; how programs worked, in the sense of how instructions were fetched and executed; and how you could group together things, where you’re doing a lot of work in a small bit of code, which you can then put into memory and just work away, and keep out of memory those things that you don’t use very often. It’s a technique called “program folding,” which nobody ever does anymore; it’s long dead! [laughs.] But it was very interesting and intricate work, to rewrite these programs. You couldn’t just take a program and stick it in and do anything with it. You had to really understand what it was doing, where it was spending a lot of time, where it was spending a little time, and reorganize it internally to fit this very small amount of memory that you could load it into.

Additionally, the programs that I was working with had files that did things like direct access to records, which was not supported on the IBM 1130. So I had to find a way of simulating this, so that we could still do the direct access; but because it wasn’t supported at a physical level, we had to support it through software. I had to design, for instance, a way to record where records were on the disk—because it did support what they called “relative record access,” so you could say things like “give me the 532nd record,” provided you knew that that was the record you wanted. So I had to do things like design a way of building an index, so you could find out where records were and access them directly. So that was interesting, because it was something I’d never really done before.

Abbate:

So you were doing systems and, I guess, databases?

Keller:

Well, they weren’t databases in the sense that you would think of them today. The term “database” was used, which basically meant a large collection of data that you might want this direct access to specific records through. Yes, so it was a combination of systems and understanding how the data were stored on the disk, so that I could write the software that actually allowed me to do this direct access that wasn’t quite a direct access. Because I now knew what the relative record number was; I had a key and a relative record number, and if what you wanted matched the key, then you could go to the relative record number. I mean, it’s a common technique, but it didn’t exist on the 1130.

Abbate:

You probably needed to know something about compilers to actually optimize that?

Keller:

Yes, I had to learn a lot about compilers in order to optimize the code and do the program folding.

Abbate:

What language were you writing it in?

Keller:

Again, that was the mixture of COBOL and assembler language. Because the other thing that we had was a little IBM System 7, which at that time I suppose probably would have been called a “microcomputer,” if they had the word. It was about the size of a large washing machine, and it was only used in that particular system as a disk controller for its disk; compared to the 1130, which just had a little one-platter or two-platter disk, the System 7 had quite a large disk, but it also had no keyboard—it just had a series of switches and lights on the front, which is why it looked even more like a washing machine! But it had quite a large disk, and so we were only using it as a disk controller and disk storage. But all the speaking to it that we had to do had to be done in its own assembler language, which was different from the assembler language on the 1130, so there was the whole problem of getting the two to talk to each other in such a way that the System 7 would accept a command from the 1130 saying, “Access this record, and get it, and then send it to the 1130.”

Abbate:

So a little bit of data communications as well.

Keller:

A little bit of data communications coming in. I didn’t look at it that way at the time; I don’t think I even knew the term.

Abbate:

I don’t think there was a term.

Keller:

We spent time trying to get them to talk to each other, yes.

Well, I suppose there was data communications at that time, but it was a rather different sort of thing. For instance, when the firm still was using this service bureau, they had a remote card reader and a remote printer in the office: so they’d put things in the card reader, the contents of which would then be transferred electronically down to the service bureau, where they’d be processed, and then the printout would be sent back up the wire to our printer. There was one very funny thing there. There was a woman who was the operator of this equipment, and at one point—I don’t know how we got into the conversation, but I said something about how the information on the cards was getting down to the service bureau, which was about three miles away, I think. And she actually thought that these cards were being rolled up, put in a kind of pneumatic tube, zipped down these three miles, read at a card reader down in Victoria—we were up on Oxford Circus; this was down in Victoria—read down there; rolled back up; stuck in a tube [and sent back]! [laughs.] And I said, “But look, when you start running this, it’s only really a fraction of a second between the time the card disappears and it reappears on the other side. It couldn’t possibly be going three miles down!” [both laugh.] But that was her mental model of how this was working, and she couldn’t really understand—despite the fact she operated the equipment—how you could get information physically from one place to another place without this physical medium of the card.

Abbate:

That sounds like it was actually a pretty enjoyable job. How big was the group of people working there?

Keller:

I was one of three programmers, and there were two key-punch operators, and this woman who was the computer operator. Well, she became the computer operator; initially she only operated this remote card reader and printer, but then when the 1130 and System 7 were installed, she was also the operator for them.

Abbate:

And the other two programmers were men?

Keller:

Men, yes. The interesting thing was, I couldn’t travel out to Iraq to do anything; Iraqis didn’t welcome women. Once or twice we had somebody come from Iraq to find out how particular programs worked, and this was a really strange experience, because we all sat in the office—me, one of our male programmers, and this Iraqi—and we had a three-way conversation, where the Iraqi would ask the male programmer, and the male programmer would ask me; I would tell the male programmer, and he would tell the Iraqi. And this was all in English. This wasn’t involving translation; this was all done in English.

Abbate:

Because he wouldn’t . . .

Keller:

Because he would not speak to a woman.

Abbate:

Bizarre!

Keller:

Yes, it was very bizarre. It was a very bizarre experience. [laughs.] And I thought it was funny, because my colleague was actually repeating what I said. Now, this man was sitting there in the room, understood English, must have been able to hear it; and yet somehow it was necessary to go this triangular charade to convey the information.

Abbate:

Maybe if you had been at a terminal in another room, and just typed things in . . .

Keller:

And he wouldn’t have known. Maybe! [both laugh.]

My mother, the business that she ran: she imported shellac into the United States—or lac, actually, and made shellac. One of the things that she found was that the lac factors in India and Thailand—she was told by people in the States, “Don’t ever let them know that you’re a woman, or you will find real difficulty in dealing with them. They just don’t want to deal with you, because somehow to them it’s not real—a woman being in business.” So she never put “Mrs.” on anything; she wrote “Clara Smith,” but she never put “Mrs.” on anything, so that they wouldn’t know. I suppose they didn’t know enough about Western names to twig, or maybe they just ignored the fact and thought, “Well never mind. She pays her bills on time!” [laughs.] But yes, there was this problem with dealing with [certain countries].

Abbate:

Your mother was running that business from home, when you were growing up?

Keller:

She actually started with it shortly after my parents came back to California with me as an infant. She started out doing book-keeping for a father-and-son firm, and they couldn’t afford to pay her cash, so they paid her stock; and she eventually ended up owning a third of the company! [laughs.] By “eventually,” I’m talking about a period of maybe three or four years. She ended up owning a third of the company, at which point the father and son had a falling out; and rather than have a stranger come in, she bought out the father’s share, and he left; and then eventually she bought out the son—so by the time we’re talking about, it was her sole business, and had been for some years. I think they left probably in the mid-fifties, and she’d had it from the mid-fifties until she retired—when she was eighty-seven! [laughs.] We think she was the only female shellac maker, probably, in the world.

Abbate:

It could well be!

So I guess she wasn’t afraid to be doing something atypical for a woman.

Keller:

Yes. Again, she never made a big thing out of it, although she did occasionally have problems. For instance, the bank wanted to know that my father was willing to stand as a guarantor, despite the fact that the business was hers, the money was hers, the income was hers; everything was hers, but they wanted this man in the background. And she did have problems when he died, because the bank suddenly said, “Well, no more guarantor”—despite the fact that they knew how she’d run her business for years. My dad died in the seventies, and this was a problem. The laws have subsequently changed, and so have attitudes, but at the time, it was a problem.

But other than that, no. She never made a big thing out of it, but she certainly wasn’t afraid of being in business, and being in rather a man’s world.

Abbate:

So how did you end up leaving the Iraqi . . .

Keller:

Iraq Petroleum Company. It was this short-term conversion work, and that went on for about six months. Then, because the Iraqis had nationalized, there was an IBM 1401 out in Iraq that they had nationalized and taken away, and it was the one that did the stock control for [one of the oil fields]. Iraq Petroleum had three major sites, and this was the one up in Kirkuk oil field, which is the area that’s right in the north of Iraq; and so for months, the people at this huge warehousing facility had had to keep track by hand of what was going in and out. We had to write a stock control system, so they kept me on to do that. That ended up with some interesting and funny stories, because the people in Iraq were keeping track on punched paper tape, so we had to write something to read punched paper tape, because the IBM 1130 didn’t normally have a punched paper tape facility. So we had to write something to read paper tape, and we wrote the stock control program; we had to do it fairly quickly. Then when they sent us [the data on] all these movements in and out of this warehouse on punched paper tape, it came in three or four great big tea chests full of paper tape, and somebody had to run it through the machine, and they loaded it backwards. A lot of it was loaded backwards, and so when this was finally done it turned out that it was all backwards! [laughs.] And so we then had to write another program to flip all the records. The records were right—you know, the boundaries between records were right—but the actual characters were backwards, and so we had to write a program quick to find all these and flip them around! [laughs.] So yes, it was an interesting experience, to catch up with all this.

Anyway, at the end of that I left. But in the meantime, my husband had been told he was going to stay another six months, and at that point I thought, “Who knows how long this might go on?” So I decided I would look for a permanent job, and I found one with Philips Electronics in Croydon, which is one of the large southern suburbs of London. I went to work there as a Senior Programmer Analyst. When I was hired, they said, “Oh, we want senior people”; but in fact they didn’t really have work suitable for senior people. It was the most boring job I’ve ever had in my life. There was nothing to do, or very rarely anything to do. It wasn’t just me; it was an open-plan office, and there was a cluster of about six of us who worked for [a subsidiary of Philips]. We were organized along the subsidiary line, so the group that I was in worked for a subsidiary of Philips called C.E.S., which did selling and maintenance of consumer electronics, primarily things like washing machines, refrigerators, record players, radios—all kinds of things— and they also sold insurance to people who had bought Philips products. And we were all bored, and we’d sit there every day: everybody would bring in a different newspaper; we’d trade newspapers around till we’d all read all the newspapers; and we’d solve all the problems in the world. [laughs.] The Senior Analyst used to lie down on top of his desk at three o’clock every afternoon and have a half-hour nap. It was really boring!

Abbate:

Were people in the other divisions busy?

Keller:

It varied. In fact, I did some work for one of the other divisions on a problem which still isn’t really easily solvable—they ended up giving it up after they’d spent about three million pounds on it, without ever producing very much—which was to go to a natural language interface, so that people could ask for reports in natural English. Very difficult problem! Very ahead of its time. It’s still a difficult problem. It’s still not something you can solve very easily. I worked on that for a bit, just mapping possible natural language statements onto controls for the program to produce the right sort of reports; and for a single set of reports, I think I ended up with something like 990 different ways that you could ask for it in English! [laughs.] But anyway, that never got anywhere.

I suppose because I was so bored, I applied my skills and I started improving the programs that we had. I was going through them, and I was taking out redundant code, and I was optimizing them—because I really had nothing else to do—and I got in trouble for it. Because it turned out it was a cost center model, and the subsidiary was charged by us for the amount of computing time they took, and for the amount of memory they used. Well, when I cut this down to the point where they actually moved down to another charging band, my boss was very unhappy, because A) they hadn’t paid me to do this work—the subsidiary hadn’t paid for my time to do this work, and B) we were now getting less money from them, because their programs were running so much more efficiently! [laughs.] So I thought, “I’d better leave.” At that point I went, again, looking for a permanent job, and I got a job with Texas Instruments, also in Croydon.

Pinups in the Office

Abbate:

Before we leave Philips, you have to tell me the pinup story again.

Keller:

Oh yes! [laughs.] Because it was an open-plan office, a number of men had nudie calendars. This is 1974, and they’d have Playboy calendars and things like this up. But it was an open-plan office, and the women who worked there used to get really annoyed at this in-your face machismo—that word wasn’t used then, but that sort of thing. After making a couple of remarks to people, asking them if they wouldn’t mind putting it somewhere else, and getting the “Oh ho ho, look at you! You’re obviously so terribly upset by the idea of beautiful women with their breasts hanging out,” we were wondering what to do. Playgirl magazine started being published just then, and I think their very first centerfold was Burt Reynolds. [laughs.] We went down and we bought a copy of Playgirl, we took Burt Reynolds out, and we pinned him up—and all hell broke loose, because the men were really outraged. I mean, here was this naked man, sitting there looking at them! They called in the senior management, who were equally outraged at this naked man, and we just said, “Well, what’s the difference between him and the Pirelli Tire Girl over there?” And in the end, what they told the men was that they had to take down the obviously visible girlie pictures; that they could only have something that was on their desk, facing them. So a little discreet calendar was okay, but they couldn’t put it on public display; and we agreed to take Burt Reynolds down. [laughs.] It was a minor victory, but it was a victory of sorts.

Abbate:

[laughs.] A little consciousness raising.

Keller:

A little consciousness raising amongst the men that we worked with, yes!

Abbate:

Good for Burt!

Keller:

Yes. Burt doesn’t know what he did! [laughs.]

Moving to Texas Instruments

Abbate:

So: you moved on to Texas Instruments.

Keller:

Texas Instruments; yes.

Abbate:

A London subsidiary?

Keller:

It was called their “European Information Centre,” and it was the hub of [their European operations]. They had quite a distributed operation: so they had an office in Germany; they had two in Holland; two in Italy; two in France; one in Spain; one in Portugal; and they were mostly manufacturing facilities. They were linked to this European Information Centre, which was in Croydon, which also provided services for a Texas Instruments subsidiary called Geophysical Services, which did oil prospecting. For those people, it was [data from] the North Sea, the west coast of Africa, Saudi Arabia, and that sort of area that was coming into the London thing; and that was linked to Dallas, which was linked to Attleboro, and then there was something out in the Far East.

Just at the point I arrived, they had finally set up this network of linked mainframes; linked to the mainframes were minicomputers in all these various manufacturing facilities and so on; and then, in turn, everybody had dumb terminals that were linked into either the minis or directly into the mainframes. It was the first example I saw—this was, again, end of 1974: they had just developed a system called “Message”—MSG—which was kind of a forerunner of email. Up to then, the company had been using Telex as a means of communication, and now you could use the computer links to do this; also file transfer, so we could transfer things without having to send disks or cards through shipping companies and so on. So that was really exciting, because here was this system where you could talk to people all over the world! And although it was intended strictly for business, the company didn’t mind—because it didn’t cost them anything—if you would put little personal messages on these. Now, the end points of this were not actually individuals, they were offices; but you knew who was in the office, so you would generally communicate with a known person, or two or three people who might be sharing a terminal, and you knew who they were. Almost immediately, when we started with this system: one of my first jobs was to teach people how to use it, and almost immediately, the interesting thing was that people started putting in these personal messages. “What’s the weather like today?” “Do you have any kids?” Things like, “I’m leaving now. I’ve got to go shopping for dinner.” “Oh, what are you going to have?” And it was amazing that really deep friendships developed over that network very, very quickly. I mean, it was almost instantaneous that you got to know people, and in a very personal way. You’ve never seen them; you’ve never heard their voices, or anything like this; but you got to know lots about them, because—I suppose subsequently people have noticed that [with] email, people will reveal all sorts of things that they wouldn’t do face-to-face, or even over the phone. So that was one of the really nice aspects of that particular system.

My job was as a systems troubleshooter, so I was getting messages from all kinds of people, that they were having problems with this, that , or the other, or they needed to know how to work some piece of equipment which had been delivered without any instructions—which was a very interesting exercise! [both laugh.] How do you tell somebody how to operate a piece of equipment that you can’t see? I think I learned an awful lot about writing instructions from that experience, including things like—it sounds silly, but things like: “Is the light on in the front? If not, look behind and see if it’s plugged in. Is the plug switched on?” [laughs.] Right down to quite detailed difficulties with systems compatibilities and so on.

Abbate:

So you were . . .

Keller:

I was a Systems Programmer, basically. That wasn’t the title, but that was the job.

Abbate:

You were doing support, basically?

Keller:

Yes: support for the users, who were all internal people. It was a lot of troubleshooting, and no two days were alike. But it was all through the medium of this MSG system. It was quite primitive: you had to know where somebody was, because although there was the name of the terminal that you used, if it was actually attached to a different mainframe, you had to know that and invoke a little program called “CPU Switch,” which would switch you from the Dallas one to the London one, or to the Attleboro, Massachusetts one, or the one out in—I think it was in Japan; we didn’t talk to them too much, because everything tended to come in to Dallas, and then back out again. You had to know what the path was.

Abbate:

Source routing.

Keller:

Yes. It was a bit primitive Almost all the terminals were not screens; they were thermal paper with dot-matrix-type printing heads—which was nice, because you could tear off the piece of paper and go show somebody! [laughs.] So there are advantages to paper. And I’ve kept a lot of the messages, actually, that I had from people. When I left, people sent me messages, which I ended up keeping.

Yes, that was nice. I think one of the most surprising things that happened with that was: one person I communicated with a lot in Dallas—who I still am in communication with, twenty-five later—he was a great enthusiast for this system. So much so—I mean, I thought it was wonderful, but he was just so ecstatic about it—that I thought, “Well, yes, it’s nice, but really, do you have to be that ecstatic about it?” And when I met him for the first time, [I realized that] he has a terrible stammer, which has obvious had an enormous impact on his life, and for him, this was the first time that he had a medium [through which] he could communicate with people fluently. I could suddenly see, when I spoke to him for the first time: “Gosh, you know, this must really be good for him, because he doesn’t stammer in print.” Subsequently, it actually gave him, I think, the courage to address his [speech problem]. He still has a terrible stammer, but he’s been an award-winning—what’s the organization called in the States that provides dinner speakers?

Abbate:

Toastmasters?

Keller:

Toastmasters, yes. He’s been an award-winning Toastmaster, and stuff like this. So I think that was kind of a first step in him taking his disability by the horns and saying, “There’s something more to me than my stammer, and I can overcome this.” So in that sense, I think it did some real good. But it also made a lot of friendships.

Abbate:

How long were you at TI?

Keller:

Unfortunately, again, only a year; because at the end of the year, [laughs] that was when Mobil finally said to my husband, “We don’t need you in England anymore,” and we were supposed to move back to New York. He was supposed to join the company back in New York; in fact, he only did that for a few months and then resigned, and we moved back to California. So there was a period there of about September to January where I was unemployed, but then we went back to California, and I was in the position again of looking for a job. We had bought a piece of property, and we were going to build a house on it, and he stopped working to act as the contractor for building the house—so I became the sole bread-winner, and that’s when I ended up with TRW.

Abbate:

So you had made this decision that he would stop working, and you would support him, before you had a job?

Keller:

Yes! [laughs.] Possibly a very shaky thing to do, but those were pretty good times in terms of jobs in the computing industry, and I didn’t have grave worries that I would be unemployed for any extended period of time. It was more whether I could find something quickly that I liked.

Abbate:

And he, evidently, had no problem with that.

Keller:

He had no problem with that. The bargain was supposed to be, I would support him for three years, and then he would support me for three years, and I would go on and do my doctorate, which I had always wanted to do and never did.

Abbate:

In English Literature?

Working for TRW

Keller:

Yes. [laughs.]

But anyway, I got the job with TRW then, with their Management Systems Division, but as a Systems Programmer. It was a funny kind of job, because we didn’t have any direct access to the computing equipment at all, so we worked through another Systems Programming Group in a different division of TRW. Again, we were doing a lot of what I did at Texas Instruments, I suppose, in that I was dealing with the user end of things and translating their requirements and needs to the Systems Programming Group in this other division. I was doing a fair bit of training. I did a lot of debugging—there was an enormous amount of debugging to do—and that was actually how I started training people on a sort of regular basis, because I kept thinking, “If they only knew a little bit more about what they were doing, they’d be able to do it for themselves instead of bothering me!” [both laugh.] So I did things like [teaching] them how to make use of some of the facilities that compilers provided for them—to learn how to read data maps and procedure maps—so they could do their own debugging. How to understand what the various error codes meant, and how different error codes had to be dealt with in different ways, and so on. But that, again, also meant that I had to really keep up to date with what the operating system does, and what the system does.

I enjoyed that, but it seemed a bit dead-end in terms of promotion, and the opportunity came—I was asked by the people in the other division whether I wanted to join them, so I went to what was then called the Systems Engineering and Integration Division of TRW.

I never quite understood exactly the whole of what the division did, partly because I think it was classified information.

Abbate:

It was military?

Keller:

Military satellites; yes. There were various strange banks of computers that I could look at through the glass window but never get in to see, which were doing things like acquiring data from these military satellites and processing it. The area where I worked was basically looking after things to do with satellite construction and maintenance. So it had, for instance, large databases for inventory control for parts, and so on. It had some time-sharing aspects, in that it provided time-sharing support to programmers in the other divisions, or within our division in other buildings and so on, so there was some telecommunications involved with this.

I think one of the first things I did was, they sent me to find out about distributed systems, which was just beginning to kind of poke its head over the horizon back then. So one of my first things was a little research project to go down and look through various journals and write a report on what distributed processing was; whether we had any [laughs] or should have any; and if we should have any, what sort of form it might take. That was one of the first things that I did with them, which was quite interesting, but it was a terribly intellectual exercise, in that I spent a lot of time looking through journals and things. I went to visit some companies that had distributed applications of various types, to look and see what they did and how they did it, and then gave a report on all this to senior TRW management and some generals from the Air Force. They decided, in fact, that we already had distributed processing, because we had—would you believe—a machine in another building where somebody took a magnetic tape off, and walked it over and put it up on ours! [both laugh.] So the idea of distributed processing was a bit shaky there. They hadn’t quite made the leap to the fact that they didn’t need to put it on a tape and walk it across; they could send it electronically.

Abbate:

So that was just a buzz-word, as far as they were concerned?

Keller:

Yes, it was a kind of buzz-word.

My work over there was: I was in charge of software and hardware integration. So I had about five or six people working for me, and we took care of all the system aspects, which ranged right from making sure that the applications were compatible with the operating system, through to the fact that various bits of software needed to be compatible with the hardware. Anybody still using these [mainframes?] knows that sometimes you get a piece of software and it suddenly falls over, because in fact you now need some piece of hardware that you haven’t got, or the hardware that you have doesn’t support some aspect of the software; so it was to avoid that sort of thing. We worked very closely with a group that did performance monitoring and tuning, and slightly less closely with the telecommunications group, because we were providing the underlying support for the telecommunications group, so that they could actually conduct their work. They were a systems programming group as well, but they only did the telecoms side of it. Then there was a modeling group—a group of two! [laughs]—who did systems modeling, to work out how to keep the same hardware that we had and just push it harder and harder and harder. Because at that time we were starting to guarantee 24-hour access, quick response times, and so on, and more and more work was being loaded onto the machine; so you were constantly having to readjust things, so that you could do this without having to buy a whole new mainframe, because they liked to run between 98 percent and 99.5 percent availability all the time. So we were always running very close to the edge of what the system could actually cope with! [laughs.]

Abbate:

Not much margin there.

Keller:

Not a great deal, no. Amongst the other things that I did was some performance tuning of databases. They had very large databases, which were inventory databases; they fall into the category of very large databases, and I did a lot of performance tuning for those, to keep access time reasonably quick. One of them was so large that reorganizing it was a terrible problem. They could only do it over Christmas or Thanksgiving weekends, because it would take more than two days just to read all the records—well, back them up first, obviously, in case something went terribly wrong; read all of [the records; and then reorganize them and write them back. (OK? End of sentence cut off.)]

[DISC 2]

Abbate:

Were you supervising a group of people? Or was this just you doing all this?

Keller:

I was supervising a group of people, as well as doing some of the work myself. I don’t know why, but my boss decided I was good at dealing with difficult people—so I always got the really difficult ones to deal with! [laughs.]

Abbate:

You mean, to work under you?

Keller:

Well, we definitely had one like that, yes; but, no: difficult users. For instance, we had a couple of database administrators: they knew what they wanted, and it wasn’t always realistic, and they used to get very upset about this, and I was the one who had to go and break the bad news, or do what I could to try to go some way towards what they wanted. One of those [occasions?], I always felt, was my real triumph in optimizing code. He had charge of this huge database—this was the one that took 48 hours just to sort—and he asked me if I could do anything, because sometimes the sort would fail partway through, and because it took forty-eight hours and you rarely ever had more than a weekend, the database would have to go on in this unreorganized state for maybe as much as six months or a year before he could get another weekend to do it. So he asked me to look at this, and I got his sort from forty-eight hours down to two hours. [laughs.]

Abbate:

Wow!

Keller:

When I realized, of course, that he was doing this on weekends and nobody else was around, [I saw that] you could afford to take all the memory that was available; he’d only been using a small portion of the memory to do this, and so it took a long time. It wasn’t any great technical trick, it was just sort of realizing, “If you’re there all by yourself, you can have everything that’s going.” But I also did the optimization for the database so that, for instance, the most-accessed records were the ones that were the closest to where the read-write heads parked, so you got close access to those, and the ones that you rarely went after (maybe once a month or something) were the ones that were way out in “no man’s land,” and you didn’t mind if it took you a second or two seconds to get at them.

But I think that was the most startling bit of optimization I ever did, in terms of results! [laughs.]

Abbate:

Did you get recognition for these types of things from your management?

Keller:

Yes, I did—in one sense. I didn’t in the other sense: the other sense was promotion, and I kind of got caught, as one often does with company policies. TRW was a wonderful place to work for, in that they were very aware of their minority and female employees and issues of equity; so while I worked there, for instance, my salary went up by leaps and bounds, because they used to compare the job I was doing with the average male doing that job, and then I would get quite a bit larger than normal increase, in order to bring me up to that sort of level. So they were very good in that way, [and] they were very good with training and things, but I got caught when I moved between the two divisions, because they had different means of setting grades. So I came across as a Grade Five, and I was actually wanting to apply for Grade Nine jobs, and I was doing what I felt was a Grade Seven job, but they kept saying, “Well, actually we don’t use those grades.” They had some other, rather complicated formula for determining things, and there were different grade numbers, and they didn’t equate to each other, so there wasn’t an easy way that I could come in doing a particular job and just slot right into this. There were strange things, like it took into account educational background, for instance, and I wouldn’t have done so well on that, because I didn’t have a technical degree. And they weren’t sure whether they were going to change from that system to the Three-Five-Seven-Nine system, or what; and so I just got caught in this.

[telephone rings; recording pauses]

Abbate:

[We were talking about] promotion systems . . .

Keller:

Oh, the promotion system, yes.

Abbate:

So that wasn’t personally to do with you.

Working for Continental Airlines

Keller:

No, it wasn’t personally to do with me; it was just the system. I’d bring it up two or three times, and nobody seemed to make any move on it; and my boss left, and I didn’t like working for the man who replaced him, particularly; so I thought, “Well, maybe it’s time to go.”

In fact, I followed my old boss to Continental Airlines. He’d gone to Continental Airlines to effect a conversion; they were really living in the second generation [of programming languages?]. And with airlines, you need 24-hours-a-day availability, seven days a week; and so the conversion isn’t a particularly simple matter, because you have this live system, and you have to do all your programming and your testing and everything in the background, and then at some point you flip the switch, and you want an instantaneous swap between the old and the new. So it wasn’t a case of minor modifications; all kinds of things had to go on in the background. That was what he went there to do, and he asked me if I would like to come and be involved, but as a trainer, really.

I was dealing with all kinds of systems. They had your typical COBOL-type systems, which were personnel records and inventory. You had the online systems, which were things like ticket sales, check-in, and at that point they were automating baggage handling, so they were beginning to get the bar codes attached to your suitcase and it being automatically routed to the right flight. And there were scientific systems, which did things like prepare flight plans, which were largely written in FORTRAN, and they were looking at getting them out of FORTRAN and maybe into something like PL-1, in order to be able to perhaps wed them in with some of the other systems. All of the online stuff was written in assembler language, because it’s the only thing you can get to do fast enough.

Abbate:

Or so they say.

Keller:

Yes. So all this was going on, and I came along actually to train [people] in all of those areas.

What do I say about this job? It was an interesting job. I had to learn how the whole airline control program thing worked. It’s now called TPF-2, but at that time it was called ACP, I think: Airline Control Program. It had derived directly from work done in the very early ‘60’s with IBM and American Airlines, called SABRE, so it was sort of a child of SABRE—not even a grandchild! [laughs.] As I say, there was this business of it all having to come together at some point, where somebody was going to flip the switch and it was going to go from one to the other.

So, I went there and was one of two trainers, but the other one was shown the door. It turned out he was running a little business of his own on the side, which they weren’t too happy about, because he was using Continental facilities to do so. He ran a little magazine for airline employees, and he was using their systems to do things like keep track of his subscribers, and so forth. [laughs.] Anyway, they showed him the door, which really left me in sole occupation. Then they decided: they had never really had standards as such; there were ways of working that they had, but they’d never had standards that they’d all sat down and agreed on; and they were trying very hard to bring together these three rather disparate groups—the online systems people, the personnel and inventory-type people, and these scientific people—to work at common standards, so that if need be, people could move from one area to the other without too great a disruption. So there were all kinds of issues about ways of working, organizing projects—and of course, what we were talking was this huge project, because all of these systems were going straight from second-generation to fourth-generation. So there was this huge project, under which there were all kinds of sub-projects, and there were all kinds of questions about project management, applying standards, and so on.

So I ended up becoming Manager of Training and Standards, so that the standards was another one of my little bags there. While I was there, I brought in standards—well, I actually bought in a system of standards for project estimation and project control. The training: I didn’t do all the training myself, obviously, because it just wouldn’t have been possible, but I bought in packages, or brought in speakers, or sent people off on courses, which they would then turn around and teach with us. It was interesting, because I really had to work with all different areas. I worked with, for instance, the computer operators; the programmers in all these three different areas; managers; so really I was working with a spread of people right from the bottom to very near the top of the company. And I got cheap flights! [laughs.]

Abbate:

So you were kind of re-engineering the whole process of managing . . .

Keller:

Yes: the behind-the-scenes process of managing projects. Because they had spent probably close to twenty years doing things the same way, there was very little experience of doing things in a different way. I mean, they’d keep making little changes, but they hadn’t done anything major for nearly twenty years. So this was a whole new culture coming in: a lot of development; a lot of rewriting; a lot of re-development; changes in the ways of thinking; changes in the ways of doing things. I think one of the very first things that happened that really got to me and made me realize what a leap there was to make, was: they hired a new programmer, and they sent him in to me and said, “This man has never seen a key-punch machine. Would you please show him how to use one?” [both laugh.]

[recording pauses]

Abbate:

You were at Continental . . .

Keller:

Yes; this whole culture change. I don’t want to make it sound like I was the only person involved with this, because obviously my boss had come there to do this. They were hiring new people who came in with more experience of the fourth-generation way of doing things, but the example of the programmer shows that they also didn’t have much connection with this rather old-fashioned way of doing things, and all this had to change. It was a period of about two-and-a-half years that this project went on, very intense; and I’m pleased to say that, although it overran on the amount of time that it took to do it—which was made up by people working nights, weekends, and everything else—on the day, they flipped the switch and it worked! So, for all the software projects I’ve known in my life that failed in some way, or had terrible problems, that was one where it actually happened on time, and it worked—and they didn’t have to go back to the old system! [laughs.]

Abbate:

And you can take some credit for that.

Keller:

Yes, I hope I can.

Abbate:

When that project was over, was your job there over?

Keller:

Well, no. Actually, what happened was that in the meantime, my husband and I had broken up; and I decided, since I had these nice cheap flying privileges, that one of the things that I’d like to do was come back to England, just for a vacation. I thought, “Okay, I’ll take a week”—it was Easter week—”take a week and come over; and since I’m coming to England, it would be nice to see people I had worked with.” So I contacted those people I was still in contact with, and said, “I’m coming over Easter weekend,” and got some invitations from various people, and came over. One of the men whom I had done support work for at Texas Instruments was—he’d been with Geophysical Services, so he was working the geophysical side, and I got together with him. We’d been good friends anyway; we got to know each other fairly well when I was working at TI, and we’d kept in touch. Anyway, at the end of the week, he proposed to me! [laughs.] And I said I’d think about it, and I went away and thought for about two days, and phoned up and said “Yes!”

It took us about another year to actually get married. We decided we had to get married, because our phone bills were running into four figures a month! [laughs.] We kept saying we wouldn’t do this, [but] we did, and in those days private email didn’t exist, so there was no cheap way to talk to each other except on the international telephone line. So, in the end, that was what caused me to leave Continental Airlines. I got married in January, and I left in March—which was enough time to pack up things in the States and use one last free flying privilege to go on honeymoon to Hawaii, and then to go to England with one dog and twenty bags. [laughs.]

Abbate:

End of job, at that point?

Moving Back to Britain, Working for Open University

Keller:

End of job at that point. So I ended up back in Britain.

I’d been quite ill when I worked for Continental. I don’t really know what it was. It started out as a kind of flu, and it went on for months, and I was still really suffering from it—getting really, really tired, for instance—so I decided I just wouldn’t take work immediately. I came end of March, ‘81; started looking for a job in October; and got the job here at Open University—I was called for an interview in December. Universities are notorious for not being very quick to [expect you to start working] once they offer you the job. They sort of assume you’re already on some kind of academic contract and you have a whole term at least to work off; and when they offered me and they said, “When can you start?” and I said, “Monday,” they couldn’t cope! [laughs.] But I was unemployed at the time, and I was now really looking forward to getting back to work.

When I applied for this job, they were looking for people with industrial experience. Otherwise I don’t think I would have looked for an academic job, but the ad said, “We’re looking specifically for people with industrial experience to come and work on a postgraduate program for industry.” So that was what I was actually hired specifically to do.

I was interviewed in December. The other interesting thing was, I wasn’t actually short-listed for the job. I didn’t know this, because I didn’t know how universities worked. I know in the States they have search committees and things. Here, you apply; they short-list you; and they might even send you some kind of a task to do—in a conventional university, this would probably be to come and give a lecture; here, it was to prepare some written material, because much of what we do is writing—and then they have an interview panel. So when I came, I was expecting to talk to one, maybe two people, and I walked into a room and here was a panel of five. When I finished the interview, they said, “Well, everyone else has done this written assignment for us, but you haven’t. Would you go away and do the written assignment?” So I had to go and buy a typewriter! [laughs.] Didn’t have one. Bought a typewriter, went down to the library, and discovered there were no books on computing at all, so I had to pick something that I could do right out of my head. The assignment was to write about a computing topic for somebody who knew nothing. I think I chose something like parallelism in operating systems, which isn’t terribly easy to do, but it was something I knew very well; so I thought, “Well, I’ll do it”—and evidently they liked it, because they offered me the job! [laughs.] But I actually had not been on the original short list. Somebody had dropped out. They’d obviously thought carefully about me, and because this person had dropped out, they phoned me up—literally that day—and said, “Can you come this afternoon?” But having worked in industry, I was used to that, so I never twigged that there was something a bit funny about this.

Anyway, when they offered me the job and I said, “I can start on Monday,” they said, “Well, we really can’t cope with that!” [laughs.] So the first two weeks they got for free, because I came in. I wasn’t being paid, but I came in. I thought I might as well get to know the place, and get to know the people and the job.

First Lecturer in Computing at Open University

Abbate:

And this would have been in the Mathematics Department at that time?

Keller:

At that time it was the Faculty of Mathematics. And I was actually the very first person hired here whose title was “Lecturer in Computing” and not “Lecturer in Mathematics”—which was something else I didn’t realize when I came here. And that led to some funny things, because we have a requirement in our contract that we must have student contact, one way or another, and the normal method at that time was through summer schools; you would teach two weeks of summer school. [But] there were no computing summer schools, and in fact there were only two rather small computing courses available. So they said, “Of course you’ll teach a summer school, won’t you?” And I said, “Oh yes, of course. I’m happy to do that!” And then it turned out it had to be a mathematics summer school, and I thought, “Hmm! I haven’t done any of this in a long time, and I didn’t go that far.” So I kept saying to them, “You understand I’m not a mathematician?” And they kept saying, “Yes, we understand that.” If you know Venn diagrams: there was no overlap between my “not a mathematician” and their “not a mathematician.” They were just miles apart! I asked what course the students would be doing, and they showed me the course, and I said, “Well, what would they be working on in the weeks that I’m doing the summer school?” And it turned out it was a block on probability and statistics, which is something I don’t feel too unhappy about. I wouldn’t say it’s the most fascinating subject in the world, but it’s something that I know reasonably well, and I felt, “Oh yes, I can cope with this.” Well, that’s what the students would have been studying that week, but what they were actually studying at summer school was something quite different. And when I came to the summer school they said, “Would you rather do differentiation or integration?” And I said “What?!” [laughs.] And it was terrible, actually. I almost ran away! [laughs.] I almost thought, “Just chuck the job! Get in the car, drive over the horizon, and never see them again!” [laughs.] I did stick with it, and the second week, the guy who was directing summer school was actually very good. He’d sit down with me each night and help me prepare for the next day. So I was literally only just the one day ahead of the students. But because he did this, and he was himself a very good teacher, and he could give me things to say and things to do, I was able to stumble through the second week without being quite such an embarrassment as I was the first week.

As a result of my inability to perform as a tutor in mathematics, they actually decided it would be useful to have a computing person, because they did have some computing exercises. Now, they were really simple: students would come in and write a few things in BASIC, just to get the flavor of how to program in BASIC and do some simple mathematical programming. But they decided it would be useful to have somebody who did that, as opposed to passing it around amongst the mathematicians. So thereafter I was able to sit and teach basic BASIC for two weeks! [laughs.] Subsequently, I discovered I didn’t have to do it in my faculty; I could do it in another faculty; so since then, I’ve done it in the Technology Faculty, which is basically an engineering faculty. For some of their summer schools, I’ve been their subject expert in IT, which is much more congenial to me, because it’s really something I don’t need to prepare a great deal for, and I feel I can really bring something to it—but when I was supposed to be this generalist mathematician, there was no hope! [laughs.] I mean, I can do differentiation; I can do integration; but I do them by rote. It’s a pattern, and I fill in things in the pattern. I don’t really understand what I’m doing, which doesn’t make you a very good teacher.

Abbate:

Well, I can’t imagine you had needed that much, working commercially.

Keller:

I’d never needed it. No, I’d never needed it. The only time I’ve ever needed it was actually in studying. It was not anything that I had ever done. Statistics, yes. For various reasons, that gets called on from time to time. At TRW, for instance, we developed some benchmark systems that we could use to test, whenever we changed the operating system. It was a kind of statistical selection of normal kinds of work, so we could simulate a normal working load on the system to see how it functioned before we actually turned it over to the real load, and that involved statistical principles to select those things. Performance tuning involves statistics, like, “Where are your hits occurring most frequently? And where do you put those to get the best use out of the machine?” So I felt comfortable with that; but that was not what was taught in the summer school! [laughs.]

Main Course Projects and Human-Computer Interaction

Abbate:

What were your main projects?

Keller:

Well, the actual postgraduate program that they had in mind didn’t start for a year after I came. They were still negotiating for the money and exactly what the content was going to be. So the first thing that I was given—I suspect it was just to keep me busy—was to take the little programming course that they had at undergraduate level; it was in BASIC originally, and I was told to take that and turn it into a “Teach Yourself COBOL” pack, which the university was then going to market. The first thing that happened was, I looked at this and I thought, “You can’t teach them in the same way.” So I had to write from scratch, basically, this COBOL course. Because [with] COBOL, you can’t very well write a little program that says, “Write ‘Hello world’” and get something out. Nothing happens! There’s all this background that has to go into it, and a structure that has to appear before you can do anything in a COBOL program. Anyway, I did that. The pack was put on sale; it hung around for quite some time. Quite frankly, I think it was terribly overpriced for what it was, since you could go down to the bookshop and buy yourself a book for about a quarter of the price and get about the same thing out of it. But anyway, it was an interesting exercise! [laughs.]

In the meantime, they’d firmed up this money that the university got from the government to carry out this industrial project. This was Margaret Thatcher’s big thing about industry, that British industry was backwards; it wasn’t applying digital techniques, and so on. It was manufacturing, all kinds of industrial processes. So the idea was to grab the manufacturing industry of Britain by the scruff and pull it into the late 20th century, and for that we got some millions of pounds to develop a series of courses in manufacturing and in what was called the “Industrial Applications of Computers” at the time, [which] included things like robotics, real-time monitoring, real-time control. My particular part was the Computer Architectures and Operating Systems, because that was my expertise, but I did do a bit of work on the real-time monitoring course. There was [also] Project Management, which, since I’d been a project manager at TRW, they sort of pointed at me and said, “You! Do Project Management.” [laughs.]

I worked on a number of courses in this. As I say, the Computer Architectures and Operating Systems was the first one; a bit of the Real-Time Monitoring; then the Project Management; and then the last one was initially called “Man-Machine Interface”—and it turned out to be an all-female course team, so the first thing we did was chuck out the word “man”! [both laugh.] “Human-computer interaction” was just beginning to be a term that was appearing in place of things like “man-machine interface,” so it became a Human-Computer Interaction course. Although I didn’t have a great deal of expertise in that area, I suppose I’ve had a long interest in this, because every time I had to change jobs, change machines, change languages, change this, change that, there was always this learning curve; and there were always things that were still wrong, that were difficult, and you’d sit there and think, “Surely we can do better.” For a long time I’d been feeling, “Surely we can do better at this,” and so the HCI course was going to address what we needed to do better. I worked on that initially as an editor, because by now I had quite a bit of writing experience, but in the end I made some written contributions to it, too, just based on my experience. The other thing was, I was the only foreigner on the team, and you’d find people saying things like, “All light switches work this way,” and I was able to say, “No, they don’t! In America, the light switches are the other way up.” [both laugh.] From that point of view, there was a lot of experience that I could bring to the course, and I really enjoyed working on that one. And it was one of the first human-computer interaction courses around, I think. It wasn’t the first, but certainly in this country it was the first.

I also got lots of journal papers out of it—with Jenny Preece, who’s now at the University of Maryland at Baltimore County—writing about engineering education and bringing HCI to the forefront of computing education. They were basically engineering education–type papers. We got quite a few out, [and] we got out a collection as a book. We got out another book, which the British Department of Trade and Industry commissioned, called “Usability Now.” They paid us to put this together from material that we’d gathered for the course. We didn’t use all the material, but we had this large body of material about usability, so we did that. Then [Addison Wesley] decided it was such a good thing that they would produce it as a [commercial book]. The DTI gave it out free to companies; [Addison Wesley] actually, I suppose, made a fair income out of selling it, because they could sell it on a worldwide market, not just within Britain. So there were a number of outgrowths of that particular one.

Abbate:

Did you have a philosophy of human-computer interaction? Or principles?

Keller:

Well, we tried to keep it away from a sort of cookbook approach [and toward] getting people aware of the underlying concepts and the underlying necessary knowledge. For instance, it included quite a bit about cognitive psychology. It was trying very hard, and I think it succeeded in being a multi-disciplinary thing. One of the team members was in fact a cognitive psychologist. We brought in a number of things to do with the social systems that surround computing systems. It was really to get people to think about their computing systems in a completely different way, which had to do with not just their function, but how usable they were, what kinds of usability problems you can get, [and] what kind of disasters you can get from usability problems—like Three Mile Island, for instance, or the plane crash here on the M1 in Leicestershire, where the controls weren’t terribly clear. The pilots knew what to do, but they were doing it with the wrong engine; they were actually trying to use the engine that had failed, because the display that told them what was happening was unclear as to [which engine had failed]. They knew one of their engines had failed, in a two-engine aircraft, and they shut what they thought was the failed one down, and it was the other one. So there were a number of things to draw on there that I think, at that time in Britain, really were at the forefront of thinking about how disaster avoidance is actually an important aspect of human-computer interaction. From what I’ve said, you should be able to work out that we weren’t just strictly talking about computers, but [also] systems within which computers might be embedded, like control systems for nuclear power plants.

Abbate:

And previous to that, it had been a bit more narrowly technical?

Keller:

Yes; and I think Man-Machine Interface had possibly focused a lot more on things like machine tools, and interactions with a worker manipulating a machine tool, and safety considerations like having a gate drop down before the hammer hits, so you can’t accidentally put your hand underneath; dead-man’s controls and stuff like that. So it was terribly nineteenth-century industrial, I think. At least in this country, [that] was the focus, and there wasn’t a great deal of focus on the kinds of systems that were beginning to appear here, more and more, without somebody stepping back and taking that same kind of safety and ergonomic view that had been happening with, say, machine tools or robots.

[telephone rings; recording pauses]

Abbate:

You said the class was all women, the HCI class.

Keller:

Oh, not the class; it was the team that did the [course].

Abbate:

Oh, the team that was writing it! I thought it was the students. How did that come about?

Keller:

Almost entirely accidental, actually. Jenny Preece had joined the department then, and that was, I suppose, closest to her particular area of expertise. She’d done her Ph.D. looking at how physics teachers can use computer displays to teach physics, so it was kind of in that general area. The OU itself has, for a long time, been obviously very interested in educational technologies, and she’d come to us from our Institute of Educational Technology, and so she came here specifically to work on that particular course; that was the reason that she was moved from IET to the computing department. We got the cognitive psychologist, Yvonne Rogers, on a fixed-term contract. She has subsequently gone to University of Sussex.

Initially, the editor was supposed to be a man; he was somebody I shared an office with for a time. There was another course team I was supposed to be the editor for, which was the database course team, which was all-male in the end; but anyway, I was supposed to be theirs, and he was supposed to be the HCI one, and we decided I was more interested in the HCI and he was more interested in the database, so we swapped. And it just turned out that the team was then all-female: the one was all-female, the other one was all-male. And we used to have a good chuckle, because as we were sitting there in our all-female team, we’d hear the other team meeting down the corridor, and these voices would be sort of shouting at each other: “SURELY YOU CAN’T THINK SUCH-AND-SUCH!” [both laugh.] We thought it would have been a very interesting thing to have an outside observer of the two teams, to see how they might have worked completely differently. I’m sure they did, because ours was much more egalitarian. Jenny was supposed to be in charge, but basically, everything was decided in a fairly democratic way, and there weren’t great arguments. We would discuss things for a while, and then somebody would say, “Okay, I’ll do that,” and that was it—which sounded completely different from what was coming down the corridor. I may be wrong; but that’s what it sounded like! [laughs.]

Thoughts on Women and HCI

Abbate:

That suggests one of the questions that I wonder about, which is whether men and women tend to concentrate in particular areas of computer science. Now, maybe it’s just coincidence that the men all ended up in databases and the women in HCI, but . . .

Keller:

Well, the thing was that we already had two men here who were database experts, and they’d worked together for years, so they were going on to this; and the only reason that the third person was male was because I swapped with him. So in a sense, that was accidental. But it also indicated something about interest, I think. I think women tend to view computers more as tools to an end, and they’re interested in getting the end right. I think men have a much more intrinsic interest in the technology—which doesn’t mean that women can’t handle the technology, or that men aren’t frustrated by things that don’t work at the other end, but I just think that there is this thing with women that . . . You know, I’m not really interested in my washing machine. If necessary, I probably could mend it, but all I want it to do is do a good job of washing my clothes. [In contrast,] I’ve got two men up the corridor here who are really keen on developing a course in robotics, and they’ve got a hold of this Lego robot kit, and you can tell that it’s the toy aspect that has actually attracted this—because this thing goes up and down the corridor every so often. [laughs.] It’s just something that doesn’t really interest me, and yet I’m very interested in programming. But the idea of trying to cajole people into programming by playing around with a little thing on wheels that learns how to follow a light around—or you can paint a stripe on the floor and it learns to follow the stripe—just really doesn’t do a great deal for me! And I don’t think I’m atypical in that. I think that a lot of women actually do have what you might call, I suppose, an instrumental view, that wants things that do a job and work properly. They want to get it there, rather than just play around with it.

Abbate:

So there’s some useful application at the end.

Keller:

There’s some useful application at the end. There’s lots of interesting things along the way, but the thing that’s important is the useful application at the end.

Abbate:

Which I guess could be true of databases as well?

Keller:

Oh yes, I’m sure it is. But it was just that, after doing my performance tuning of databases, it wasn’t really something that I felt deeply wedded to! [laughs.]

More on HCI

Abbate:

Right.

So it sounds like you had quite an impact in the HCI area.

Keller:

Well, yes; the group certainly did, I think. It was really the first course of its kind in this country, and as a result, Jenny and I, and to some extent Yvonne, did lots of traveling around: saying what we’d done, what use we thought it was, what we thought needed to follow from it, and so on and so forth. Interestingly enough, it’s now in its third rewrite, and it’s been not as popular as I would have hoped.

Abbate:

This is the course?

Keller:

The HCI course. But the interesting thing is, when people come to do their Master’s dissertation, the bulk of them who do the Master’s dissertation actually choose an HCI project, or one that has an element of HCI in it. So in the sense of people who actually go on to complete the whole course . . . I should say these courses—the top courses—are part of a postgraduate program, for which you end up with a diploma; but if you want to turn that into an M.Sc., then you have to do a 660-hour project and dissertation, and it’s only a minority of students who end up doing the M.Sc. At the start, if you ask people if they are going to do the M.Sc., about sixty percent of the nearly four thousand students will say Yes, but in the end it’s about a hundred and ninety who actually do it. But they almost always have an HCI element in their project, or specifically choose an HCI project—which says that somehow, for the people who do do the course, it has an enormous impact.

Abbate:

So beforehand, there’s not this awareness that it might be something they’re interested in . . .

Keller:

Yes.

Abbate:

. . . but afterwards . . .?

Keller:

They might just be doing it because they have to make up the credits, to get as far as the diploma; but it obviously has an impact that’s much greater than the numbers of students who go through it.

Abbate:

Maybe it has an image problem?

Keller:

Yes. I mean, we’ve changed the name now [more than once], because initially it was just called “Human-Computing Interaction,” and a lot of people didn’t really understand what that meant. They thought it meant designing interfaces, and of course it was much more than that: it was looking at computer systems within the social system of the office; it was looking at how computer systems change systems of work—for instance, in factories—and there was that element to it; there was the cognitive psychology aspect. It was funny, because students who came with a software development or hardware background came to this course, and they couldn’t cope with the psychology; and the psychologists would come, and they couldn’t cope with the software development parts of it very easily. [both laugh.] So there was obviously this gap of people coming from two different cultures and meeting in this course. But, as I say, [for] the students who’ve gone on to do their M.Sc. dissertations, it’s obviously been extremely influential, which is interesting.

We’re now calling it—I can’t remember the exact title, but it has the word “usability” in it, because that’s something companies are more likely to sponsor students on. If it says, “How to make your systems usable,” and they’ve just ended up paying for [computer?] stuff that gets left to get dusty, they’re probably interested in that as something they would be willing to sponsor their students on.

So, yes: I think it has a bit of an image problem. I think, again, the hard people sort of think, “Oh, gosh, all this wishy-washy stuff about people’s mental models of their heating systems at home, and stuff!” [laughs.] But, it does have an influence, and I think more and more people are beginning to realize that in the end, you can write the best system in the world, [but] if it isn’t usable, or if it isn’t properly usable, it’s nothing but trouble, or it’s been a waste of time.

Earning a Master of Science Degree

Abbate:

Now, you yourself got an M.Sc. at some point.

Keller:

Yes; in fact, it was just before we started on the HCI course. [I did it] because my first degree was English Language and Literature, and one of the questions I was constantly being asked was, “How can you possibly be teaching computer science in higher education, without having a computer science degree?” Well, the first thing is, computer science degrees didn’t exist when I graduated; but then, those people here who are about my same age came from backgrounds like physics or engineering. So I didn’t have this technical degree, and it just really got to irritating me! [laughs.] So I thought “Okay, I’ll go off and I’ll do an M.Sc.”—Master of Science degree—and I actually wanted to do what they call a “conversion” Master of Science degree, because [they are very general?, and] my knowledge is actually quite spotty in many ways. I know lots and lots about operating systems; I know lots and lots about computer architectures; I know less about compilers—I know almost nothing about theory of compilers, for instance; and I thought, “At least a Conversion Degree will give me this kind of spread.” But when I applied, they said, “No, you’re overqualified. We wouldn’t allow you on such a course, but you could take one of these specialist M.Sc.’s.” The one I ended up opting for was Data Communications, Networks, and Distributed Systems.

Now, I was interested in the distributed systems half of that more than I was the data communications networks. Unfortunately, it turned out to be about seven-eighths data communications and networks and one-¬eighth distributed systems! But that’s all right. I just learned an awful lot about telephony, and radios, and things like this that I didn’t really [want to know]. [laughs.] And designing waveguides: that was where I used my differentiation and my integration, was in the electronics part of that course—which I hasten to add, I passed by the skin of my teeth! [laughs.] So now I’m an informed buyer of hi-fi equipment, if I need to be. [laughs.]

As I say, I learned an awful lot about telephony that I didn’t particularly want to know and wasn’t interested in. I still think there’s a real gap there between the kind of people who set up data communications networks—who tend to come from the telephony background, even though everything has become digital—and people who are actually wanting to do things on those networks. There still is a gap of terminology and all sorts of things. If you’re developing an application and you’ve got a good network, you’re not going to be terribly concerned with noise, for instance; [and] now most networks have lots of redundancy in them, so if a part of the network breaks, you almost don’t notice, because it’s going by another path, because they’re self-routing. At one time you might have wanted to know this—[and the telephony people still focus on this?]—but if you’ve got a stable network, you really don’t worry about that. All that happens is you get really upset when your local area network goes down, and you’ve got to go find the little guy somewhere—the man with the screwdriver, who needs to go in and fix it—and even that doesn’t happen very much any more. So, as they get more and more secure and more and more reliable, these things recede further and further into the background; and yet there’s still a kind of culture gap between the two halves. It was quite clear in my course that they didn’t really know how to bring the two together; they were two distinct parts of the course that really didn’t have an awful lot to do with each other.

Abbate:

So you never really used that communications part?

Keller:

The only thing I use it for is, I work closely with the Technology Faculty, with what used to be the Electronics Department, and is now called Telematics; and there are a number of network people in there, and sometimes they need somebody to come along and sit on an exam board or something like that—and I can do it, because I’ve got the background. Conversely, I sometimes call them over when I’ve got projects from students that have a heavy element of data communications networks, and I want really the most up-to-date knowledge applied to it. But that’s about it. I have not wired up my home! [both laugh.]

Abbate:

So that was mainly useful to you as a credential, rather than . . .

Achieving Chartered Engineering Status

Keller:

It was useful as a credential; yes. It helped me do a number of things. One was that it helped me achieve Chartered Engineering status.

Abbate:

Now, what is the significance of that?

Keller:

In Britain, professional societies and professional society membership is much more important to someone’s career than in the States or in Europe. In Europe, it tends to be your degree that counts; and in the States, I think it’s what you’ve done, what’s on your resume, that makes the most difference. But here, it’s professional society membership; and the British Computer Society has long suffered from deep envy of the various older engineering professions, like the Society of Mechanical Engineers, Society of Industrial Engineers, and so on. They’ve wanted to promote computing as an engineering discipline, and about ‘87 or ‘88, they managed as a professional society to get the Engineering Council, which is the umbrella body for all these professional engineering societies, to accept BCS membership as a means to the Chartered Engineering status.

Now, Chartered Engineer would mean that I could practice as an Individual Consultant Engineer. It’s not a legal requirement.

Abbate:

It’s not like passing a bar exam?

Keller:

It’s a little bit like passing a bar exam, except it’s not a legal requirement in the sense that I can also go out [without the C.E.] and hang up a shingle that says “I’ll do this for you.” But the Chartered Engineering status, I suppose, says that you will adhere to certain ethical principles; you will be insured; [and] they are offering to your clients some kind of guarantee that someone has looked at your qualifications and decided that you’re qualified to practice. Britain is a funny society, in that very few things are regulated in the way that they are in American society. A lot of it is just by this funny little consensus, which I find quite strange. It’s sort of codes of practice, rather than regulations. But yes; that’s what it offers.

Of course, what it offered to me personally, in terms of my career here as an academic: well, two things. One is: as a woman, you tend not to be taken quite so seriously, and being able to put the initials after your name says, “Yes: I’m somebody to be taken seriously.” The other really is that it helps towards promotion.

Abbate:

We spoke a little bit about this, last time I spoke to you, about the promotion system. As I understand it, it’s kind of self-initiated, in the sense that there’s not a regular timetable for you to be considered for promotion, but you’ve got to put yourself forward?

Keller:

No, there is a timetable, but what happens is, the dean sends out a notice some months in advance, saying, “In so many months’ time, we will be considering people for promotion. You need to make a case and submit it by the first of February,” or something like this. But it has tended, by and large, to be individuals putting themselves forward for promotion, and the problem with that is that people who are self-effacing don’t tend to do so well under such a system as people who are self-advertising. So from that point of view, it doesn’t feel terribly fair. And the other thing is that this case that you make for yourself—this is peculiar to the Open University, now: it tends to have to have all kinds of superlatives in it. We joke about it. We say, “You have to say things like ‘catches bullets in bare teeth’; ‘leaps over tall buildings in a single bound’; ‘stops railway trains with a raised hand’”—these kinds of things! [laughs.] But it is actually that kind of superlative, because the criteria for promotion are excellence—and the word “sustained” appears there—so “sustained excellence,” and “innovation,” and things like this. So you really have to be jumping up and down in your case and saying, “Look what I did! Look what I did! Look what I did! Look how wonderful it is! Look at all the things that have come from what I’ve done.” And a lot of people—and I think particularly women—do not find this a very natural thing to do. I certainly found it extremely difficult when I went to do it.

What we’re trying to do now is to have senior people within the department sit down with people that we feel should be putting a case, and help them to develop that case. It’s a little bit different from what it was, say, two or three years ago. It was this rather awful “Hey, look at me” thing—and of course, in a way, you stood a better chance if you’d been standing there saying “Hey look at me!” for several years in an informal way, sort of advertising what you’ve done around your colleagues, rather than just getting on with the job.

Abbate:

Had there been any system of mentorship?

Keller:

That’s really come in more. I mean, when I came here, I was just shown a room and left to get on with it! [laughs.] And now, people coming in: I’ve got five people coming on Monday. They each get a probation supervisor; they get an induction supervisor; they get a mentor; they sit down and they negotiate specific objectives; the objectives are discussed; they’re revisited after two or three months to see how they’re getting on; there’s a training program. None of that existed when I came. So that’s changed a lot, over the past three years or so.

Mentors in the Field

Abbate:

Have you had people that you considered to be mentors?

Keller:

Well, I’ve had people I’ve considered to be mentors, I suppose, really throughout my working life—without necessarily putting the term “mentor” to them. When I worked for Computer Sciences Corporation, our boss was Larry Cannon [sp?], and we always used to refer to him as “Old Tab-Man Larry,” because he learned his computing at the end of the Second World War, using tabulating equipment to keep track of soldiers who had contracted venereal disease in Paris! [laughs.] Would you believe? He was stationed in Paris, doing this. Anyway, he was the age of my parents, and he was always a kind of father figure, and he gave me advice and stuff like this.

The boss that I talked about that I moved with—who was much closer in age to me, but senior—was, I would say, very much a mentor. But with him, that was almost his very nature. He was just that kind of a person; he was always very concerned with the people who worked for him, and to know what they were doing, what their problems were; to help them out when they had problems; to protect them when they’d done something stupid. He was very good at that; he would protect you. He would say, “The department takes responsibility.” He would never name an individual—which of course, needless to say, endeared him to those of us who occasionally did some really stupid things! [laughs.] I took a lot of what I hope is my management style from what I saw as his way of approaching people. Also, technically he was very good, and he was good at giving people chances to do things.

Here [at the OU], I didn’t really find somebody until I ran into Jenny Preece. She’s actually slightly younger than I am, and I’d been here longer than she had, but she had the research background and stuff that I didn’t have, and was very helpful in that and very supportive in that. In that sense, we’re still very close friends, and we still help each other out, and send people’s names back and forth, and say, “You should speak to so-and-so.” But I never had a formal mentor here. The one person who might have helped me out, when I went to ask him about starting research, he said, “My advice to you is, ‘Write lots of papers’”—which was not terribly useful.

Abbate:

There’s a novel idea! [laughs.]

Keller:

Well, it’s not a novel idea, but it’s also not particularly useful to somebody who’s never done academic research before, either.

Publishing in the Field

Abbate:

When did you actually write your first paper? Was that from the HCI work?

Keller:

No, actually, the first one I did was—would you believe—back to that COBOL Pack. I wrote a paper for—I can’t even remember now; it was a journal on education and training—about using the problem space, rather than the details of the language, to teach the skill. So, instead of sitting down and teaching people about COBOL, what I did when I wrote that pack was to teach them about solving problems, using COBOL as a language in which to solve those problems. I don’t know if it was a particularly unique approach, but they accepted my paper and published it! [laughs.]

Abbate:

It sounds like what you said about focusing on the usefulness of it, rather than the thing itself . . .

Keller:

Yes.

Abbate:

. . . which is possibly a gender thing; and if it is, then maybe they hadn’t seen it before.

Keller:

Well, I find that [in] an awful lot of programming classes—even when I was working at Continental Airlines; of course, I was teaching programming as well—the poor people who were sitting there would get all this detail dumped on them and not really understand what the heck they were supposed to be doing. Hopefully, eventually this would work itself it out in a skill that looked like programming. But I actually found it an awful lot easier to say, “Right: Now let’s take a problem”—and it can be a really stupid one like keeping a bank account up to date, or it can be something more exciting if you like—”and let’s talk about the problem and how we have to solve it, given that what we have to do is work it down to a step-by-step description of what to do.” And then, “Lookee lookee! Here’s a programming language. We can translate these step-by-step descriptions into statements in this programming language.” And then the programming language, the details of it, have meaning, in a way that I don’t think they do if you start with, “Every COBOL program has to start with a statement that says, ‘Identification Division.’” [laughs.] You know: why?

Abbate:

[mock excitement:] “Let’s talk about data types!”

Keller:

Yes: “Let’s talk about data types.” Much easier if you start talking about data in the round, and then you say, “To the machine these are only bits, so how do we tell the machine which kind of bits these are? Are they numbers? Are they letters of the alphabet? Is it part of a graphic?”

Reflection on Teaching

Abbate:

Have you been able to shape the teaching program here?

Keller:

I think so. Particularly the postgraduate program, because in ‘89 I took it over as the Director of the postgraduate program. At the time, the money from the government had run out; it was losing money; it had very few students and was having fewer all the time; and I got the university to underwrite three years in which to turn it around. In fact, it only took a year—because there was good stuff there; it was the delivery end that was the problem. There was good material; it was the way it was being delivered. We changed a lot of things. Oddly enough, this coincided with recession, and I think we got an awful lot of students out of the fact that it was a recession, because people were either worried about losing their jobs and thought, “Right: if I take a course and I’ve got a certificate, then that helps to recession-proof my job,” or people were looking to move, maybe from an industry that was a bit shaky into some other industry, and just didn’t feel confident to do it without some training. Anyway, that went from losing three hundred thousand a year to turning over about four million, of which it was turning a profit of about two hundred thousand a year. So that was one of my accomplishments that I feel very proud of!

Abbate:

You did that in a year?

Keller:

Well, it was three years by the time it went all the way to the top. The first year, we went from losing three hundred thousand to losing fifty [thousand]. The next year we went to just over breaking even, and then the third year it really took off.

A lot of the things that I did were: for instance, we looked at what courses were selling and which courses weren’t, and we got rid of the ones that weren’t, and we replaced them with ones that would. We did a lot of change into the way they were marketed, and so on; and in the end it worked out very well. But I think I pretty well set the shape for that, for the period from ‘89 to about now. We’re just again rethinking what it looks like. We even changed the name of it from “Industrial Applications of Computers” to “Computing For Commerce and Industry,” because we saw that the market of commerce was actually a big market; the industrial market is actually quite small.

Abbate:

[The former name] sounds like you’re just making machine tools or something.

Keller:

Yes. Well, there’s a companion program called “Manufacturing, Management, and Technology,” which is about one-third the size that we are, that started up at the same time; and they’ve never been able to get over that—because basically, this is no longer an economy driven by industry. There’s still industry; there are still people out there; but it’s not nearly as big as people who are pushing paper and money around—or, these days, electronic equivalents of paper and money.

Becoming Department Chair

Abbate:

Let’s see. So you were head of that program, and I know you’re the [Department] Chair now. Were there other [administrative positions]?

Keller:

No, after that I sort of went back to being a normal member of staff! [laughs.] Let’s see, I left that in ‘95, so between ‘95 and 2000, I was really just being “one of the boys” [laughs] doing teaching work. A number of things came up. I had an opportunity to work on a course: there was originally a course called “Women’s Studies,” and then they wanted to replace it with a course which became, I think, “Issues in Women’s Studies”; and the first Women’s Studies was largely sociology, and the second one they really wanted to expand this, and they wanted to have certainly some technology and science in it. I got a chance to work on that—well, I volunteered to work on that—and that was enormously interesting. It was a personal interest of mine, obviously, having spent a long time working in—well, you can’t even call it a “man’s world,” but a specifically male environment, which was engineering; quasi-engineering. And I thought, “I’ve got lots of experience with this, and I would like to abstract that, to see what it’s like for other people elsewhere.” So I worked on that course, and that’s the only undergraduate course I’ve worked on up until [now]; I’ve just started a new one here. That was great fun, and a book came out of that, which is called Inventing Women by Gill Kirkup and myself, which we had great fun doing. It’s partly collected papers, partly stuff we wrote ourselves, and partly commissioned articles, and we looked at things like the technology of war and women, the technology of medicine and women, scientific views and women, biology and women. So: quite broad-brush, and we wanted it to have a really worldwide feel to it, so for instance we addressed Third World technologies and women, and development and women. It was a real eye-opener. The Open University has a long tradition of “If you don’t know something about it, and you’re interested in it, go and teach it!” [laughs.] Which means that you have to do all the background scholarship to get up to speed, and that was extremely interesting. I did that; and rewrote the Computer Architectures and Operating Systems course; rewrote the Project Management course; made an attempt to replace the real-time courses, which we had thrown out, with a general course on writing real-time systems. That was never quite successful, because we really couldn’t get together a team. There just aren’t enough people with expertise in this department. The expertise is over in Telematics, but they’ve been busy on their own thing, so we’ve never been able to do that. It’s still something I wouldn’t mind being involved with.

Right now, of course: I’ve been Head of Department since March of 2000, and when I took it over, it was not a very happy department. There was a war going on between the department and the faculty. The Head of Department was—well, let’s put it this way: his solution to problems was to ignore them. He was gone a lot; he typically was gone two to three days a week, every week. A lot of people were unhappy, and I just thought, “Either I have to leave here, if it stays like this, or I have to do something.” We have elected Heads, so I put myself forward as Head; got it; and here I am! [laughs.]

So now my influence is much more . . . We’re spending a lot of time looking at curriculum, both at undergraduate and postgraduate level, because up to now we’ve had this collection of courses that don’t quite mesh up. [This is] because of the way the OU degree used to be, and still is in some cases: it was a general degree, so people could just come here, and as long as they did so much at Level One, so much at Level Two, so much at Level Three, it didn’t matter what. I mean, they could mix and match to their heart’s content: it was a degree. Most people did tend to specialize in something, but there wasn’t a great deal of guidance, and the degrees didn’t say anything after them. They just said, “Bachelor of Arts, Open University.” And then the first [change] was, students—particularly in the sciences area—wanted a Bachelor of Science degree; so it became a general degree that could be either a B.A. or a B.S., but if it was a B.S., then it had to have a goodly chunk of things that were qualified as B.S. But it still was very general, so somebody could do a computing course; they could do a physics course; they could do a biology course; they could do a chemistry course; they could do maybe one in the social sciences, which might be something like the Sociology of Computer Systems course; and they could say, “That’s a Bachelor of Science degree,” and it could be just this real mix and match. Or they could come in, as people did, for instance: they might do a degree that really was basically an electronics degree, but it still didn’t say “Electronics”; it just said “Bachelor of Science.” Now, due to student demand, we have what they call in Britain “named degrees,” which means that it’ll now say “B.A.” and then in brackets “English Literature,” or “B.Sc.” brackets “Materials Science,” or something like this. [This] means that all these course where we didn’t really care that people came and went, we now actually have to have a progression through them. And it has to make sense; there has to be some kind of starting point and end point—because the other thing about the old degree was, people could do things in any order; as long as they felt they could cope, they could do it in any order. Now, we’re in a completely different business. But it takes us about three years to develop new courses, or even to make major changes to a course, so we’re in that transition process, sitting here having to rethink what we’re doing, what we want in two years’, three years’, four years’, five years’ time.

Abbate:

Do you think it’s better for faculty morale to have it more structured?

Keller:

I think it’s better for quality, because previously, the check on the quality was entirely internal to the module. Somebody would write the module; somebody else would look at it and say, “Yes, I think that looks okay,” and that was it. That was the sole check. Now—it seems bureaucratic in some ways, but now you have to sit there and say, “What are actually the learning outcomes of this course? Do they make sense at the level that I’m pitching this course at? And what am I expecting? Since I’ve got these now detailed outputs, what am I expecting as inputs?” So, you need to know what the courses are that might precede you, and what their learning outcomes are. So it’s much more structured, and I think it’s much better for quality.

Abbate:

It sounds like it might also force the faculty to work together more, have some common vision . . .

Keller:

Yes.

Abbate:

. . . which could be good or bad for [faculty] relations, depending.

Keller:

Yes, I think it will [be good]. I think in the end it will. I mean, there’s still a lot of old-timers around here who like it the old way and get terribly upset when you start talking about changing things. For instance, the Mathematical Sciences Degree, as it’s called, got slated because it has no path through it, so there is no clear progression; and “progression” is now one of the buzz-words of the named degrees: you have to show that people start at a certain point, and they progress through, and they end at a certain point. And they’d put theirs together just out of a collection of courses that had no progression. They’re very unhappy about this. They were perfectly happy with the way it was, but obviously they’re going to have to rethink this, because it just won’t wash now, in the outside world, anymore. You can’t just say, “Well, it doesn’t matter. As long as they can cope and they can pass it, who cares?” So, new thinking is having to blow in to these people. They’re good teachers, but I think they’re still teaching exactly the way they taught twenty-five years ago. Nothing much has changed. They’ve changed the courses; they’ve updated them; [on the] TV programs they’ve done away with the guys in kipper ties and flares, and they’re now balding and dressed in pinstripe suits [both laugh]; but nonetheless, the basic methodology that they’ve been using to teach has stayed the same. The notion that anybody can do anything in any order is going to die hard, I think.

[DISC 3]

Recent Research and Interests on Gender and Computing

Abbate:

What kind of research have you been doing lately?

Keller:

Well, actually, I’ve really stuck with the women-in-professional-computing area. I initially started looking at error detection in complex systems, because, having been a debugger for so long, I was interested in that; but I never really made very much headway in that. I did a large literature search—which I then turned over to somebody who used it in his work, which was “why systems fail”—but I never did anything with it. It was actually while I was out visiting with him—he was at the University of Western Australia, and he and his wife, after my second husband died, offered me a place to come to, and arranged with UWA for a visiting research fellowship; and I went out there, and the only thing that UWA asked of me was that I give a seminar to their engineering department on why women didn’t take up engineering. That was one of the reasons I got really interested in this, because I was supposed to be out there looking at how people detect problems in complex systems (which I did), but the much more interesting background reading was actually putting together something to give this seminar. In the end I gave the seminar to the combined engineering departments of three universities in western Australia, which was a bit daunting, because initially I thought I was going to have this rather small group, and I sort of ended up with two hundred people! [laughs.] And I hope it was thought-provoking when I gave it; because they’d been trying various things to get women to come into engineering, which included computing—I mean, computer science actually sat in the Engineering Faculty at UWA; they’d been trying various things and just not having any success, or any noticeable success. So they asked me to give this seminar, and I hope I was very thought-provoking, because I was really kind of nasty to the engineering profession! [laughs.]

Abbate:

What did you say?

Keller:

I just said that it was that it was not conducive to women coming in. For a long time they’d been excluded for various reasons, both from the education and from actually practicing. There were all kinds of philosophical antecedents, particularly in science—Bacon’s emphasis on science as a masculine activity and nature as a feminine figure to be investigated, to be unveiled, and so on and so forth—that led women to just feel “this isn’t really the place for me”; and now you couldn’t suddenly stand there and say, “Hey girls, the door is open; come on in,” and expect people to come rushing through, because it wasn’t happening. So that was basically what I said! [laughs.] So yes; I was a bit nasty about the way it’s practiced: about the fact that often projects in engineering schools are highly competitive, which is a problem, [and] how sometimes male instructors don’t really understand that women’s tentativeness is not their inability, it’s just a kind of gendered approach to things, where you don’t put yourself forward as positive of what the answer is. So it was that sort of thing that I talked about.

And I’ve really kept that up, working with two Ph.D. students who were working in the area of computer science education: one specifically on gender issues in secondary schools, and one was not specifically gender issues, but we kept finding interesting gender things sort of popping up out of the research. And I ran a group for not just my two, but Ph.D. students elsewhere in this university and also University of Hartfordshire’s Department of Occupational Psychology, all of whom were looking at some aspect of gender and occupation, and most of that impinged on computing or computers in some way. [For example,] one woman who was at University of Hartfordshire was looking at nonstandard gender choices in training and employment, and to do this she was looking at males who went into radiotherapy. In this country that’s almost a completely female [occupation].

Abbate:

Radio therapy?

Keller:

Radiotherapy, yes. For cancer treatments and so on, you have radiotherapists, who are the ones who—you might call them “radiologists,” I think.

Abbate:

[In the U.S., yes.]

Keller:

Anyway, [she was looking at men] going into radiology, which is almost a completely female occupation here, but you have some men going into it; and she was looking at women who went into programming. So it was that kind of thing.

Unfortunately that has kind of ended, and it ended with a paper that we jointly published on the theme, again, of “opening the door and saying ‘Hey girls, welcome’ is not enough”; but we actually then had research that went all the way from small childhood (age three) on up to career choices on leaving university. So we had a real range of research in this paper, right from the fact that three-year-olds have already picked out “Boys do this and girls do that,” on to people who have to make the choice of going into an unusual area, and the problems that they run into: maybe with parents, maybe with peers, maybe with potential employers, going on a course at university where they might be the only one of their sex in a class. Very interesting! So that’s what I’ve kept up with.

But since I’ve been Head of Department, I’m afraid I’ve done [nothing in that area?]. No, that’s not quite true. I’ve also been looking at: we now have had almost four hundred students pass through our M.Sc. project and dissertation, and it’s quite a rigorous one—it’s actually probably more rigorous than most in this country—and one of the things I became curious about was, what made a successful student? Because we found that we’d have a cohort of students going through, and only a minority would pass on a first attempt.

Abbate:

Would pass what?

Keller:

Pass the M.Sc. In other words, their dissertation was acceptable. That was the only thing that counts.

Abbate:

So they’d done all the coursework fine, but . . .

Keller:

They did all the coursework fine; they may have done everything fine; but the dissertation was what counts. The dissertation was a description of their project: the research they put into it and what they did, the methods that they applied, their justification for the use of the methods, and so on. Only a small minority passes on the first attempt. Then there’s a larger minority that gets minor corrections. Now, these aren’t just simply two or three minor corrections; they tend to be what I would call “major minor corrections.” [laughs.] They can’t write, is usually the reason that they get minor corrections. And then there were people who essentially failed and had to do it again. We called that “major corrections,” although that’s a bit of a misnomer, because the whole thing was so flawed that there was nothing that they could do to rescue it without redoing a lot of the work. And then we had some outright fails. So I got interested in why it is that some students manage to make the hurdle and others don’t, because I’m interested in getting more of them past that hurdle.

So, I had this large database of information about these people—everything from all their proposals, their interim work, their dissertations, remarks by their supervisors, remarks by the examiners, right back to all kinds of information like age, occupation, whether or not they were being supported by their employer, whether they were married, how many kids they might have: I’ve got all of that. We turned the statistical handle on all of this to find out what correlations might exist. What their previous educational attainment was: now, you would think [that would be an indicator.] We’re open as to people coming in, so people do not have to have a first degree in order to study for the postgraduate diploma; but they do have to be working in the area for five to seven years before they’ll have enough experience, so you could kind of call [the degree] “post-experience” as well as postgraduate. So we get people who come with almost no qualifications, because they left school before taking what used to be called “O-Levels,” and are now called GCSEs; or they might have a couple of A-levels; or you’ve got people who did the Higher National Diploma, which was a kind of sub-degree—it was like the first two years of a three-year degree, and it tended to be more practically oriented and less theoretical; or you might have people with first degrees. So: does their previous educational attainment tell you anything about their chances of success? And the answer is no, not at all! No correlation whatever. [laughs.] What did have a correlation was whether, in the taught part of the program, they had studied the human-computer interaction course, interestingly enough, and the project management course. I don’t know quite why. My two guesses, which I still have to follow up, are that, [first,] in both of them, the assessment is project-based—in other words, they have to follow through the steps of a project in order to do their assessment; and the second is that their work is described discursively, so they have to write; it gives them practice in writing, and it gives them feedback about their writing at an earlier stage. There’s a very good correlation between doing those two courses and doing well on the M.Sc. There’s also a perfect correlation between doing perfectly on the taught part and passing the first time on the dissertation. But once you break that “perfect”—so there’s eight courses; if you do seven of them with “distinctions” and one with [only] a “merit”—the correlation falls apart totally. Don’t ask me why; I don’t know; but if you do everything perfectly, your chances of success on the M.Sc. are just about one hundred percent. Just a slight deviation from that, and it goes right back to the same as the people who struggle. So it’s really strange. Somewhere here I’ve got a big envelope with all that stuff in it; I haven’t written it up yet.

Abbate:

So you might end up requiring one or both of those courses that seem to prepare people for the M.Sc.?

Keller:

Yes. Well, certainly we’ve now passed the word to students that if you intend to do the M.Sc., you will greatly help yourself by doing these two courses. So that’s useful, because again, the program is a little bit of an open-ended thing; there’s only a restricted set of courses, but they could avoid one or both of those if they wanted to. And what we’re saying is . . .

Abbate:

“You don’t want to.”

Keller:

“If you really want to do the Master’s Degree, then we recommend it extremely strongly that you do these two.”

Abbate:

One thing that has struck me in the course of this interview is that there seems to be a strong identification of computer science with engineering. We talked about the British Computer Society sort of wanting to be like the other “real” engineers; and maybe also this group in Australia, where you talked about women in engineering. And I wonder how this association is made. I mean, it doesn’t seem inevitable. At a minimum, there seems to be an argument about whether computing is mathematics or engineering, and clearly there are other alternatives.

Keller:

Yes.

On Computer Science, Engineering, and Mathematics

Abbate:

Do you think that’s accurate, that there’s a sense that it’s more aligned with engineering?

Keller:

I think that good software development practice draws on good engineering practice and can learn from it. On the other hand, they’re not really the same, in that with software, you develop a product once—which you may then duplicate and send out, but the duplication is a perfect process, so once you’ve got the product working, you can just copy it and send it out. Whereas if you’re designing an aircraft, you design the aircraft, you build the prototype, you refine the prototype, and when that’s done, all you have is plans for building more aircraft. The product isn’t the end of the development process, whereas the product is the end of the software development process. So that aspect of it is, I think, quite different. But good engineering practice—I mean things like, for instance, getting your requirements right initially; understanding the materials you’re working with¬: the programming language you’re working with, the underlying computing system, the underlying network system, and so on—is important. So that aspect of it [is similar].

On the other hand, it’s got lots and lots of elements to it that really don’t have a lot to do with classical engineering, if you think about it. For instance, Web designers: I think you would put them into the category of people who work closely with computers and make computers useful for other people, and yet they need know very little about the underlying system to develop a decent Web page. A lot of them do, but some of them don’t. I mean, we’ve got four of them sitting down the hall, and one of them is really a graphical designer who then learned some computing, and the other three are computing people who’ve learned some graphical design—and yet they’re all doing basically the same job, having come at it from very different [directions]. Interestingly, the one who started as a graphical designer is female, and the other three are male. [laughs.]

Abbate:

Why am I not surprised?

Keller:

I just throw that in! I just throw that in, but yes.

Abbate:

Do you think it’s perceived as a branch of engineering?

Keller:

I think particularly people like the British Computer Society and a lot of my colleagues here like to see it that way. Years ago I used to talk about biology suffering from what I called “physics envy”: because physics is, in a way—it’s not really, if you get deeply into it, but in a way it’s this sort of deeply mathematically oriented and precise science, and biology is always kind of worried about the fact that they can’t come up with formulae which describe the behavior of foxes, or something like that, and there’s always been a little bit of an envy there. And I think in computing, there’s always been a mathematics envy, or an engineering envy, where there’s a rejection of it as an art and an embracing of this—because mathematics is hard; mathematics is difficult; engineering is one of those things that’s hard. It’s hard in both senses—hard in the hardware-y type sense and in the sense of being more difficult for a lot of people—and I think when you do things that are perceived as more difficult, there’s a kind of admiration from other people that goes with that. I certainly found that when I was first working: people would say, “What do you do for a living?” and I’d say “I’m a programmer,” and they’d say, “Oh! You must be terribly mathematical!” It was with this kind of awe. On the other hand, they’d also very quickly find some reason to go and speak to someone else! [laughs.] So it’s a two-edged sword! But I think there is this kind of envy that drives a lot of that. I think rather than just say, “There are things that we can borrow from engineering which would be useful in computing,” we’re really saying, “Gee we’d really like to be engineers.”

Abbate:

They want that status.

Keller:

And we’d like real engineers to believe that we’re engineers—to accept us as engineers. That was why it was so important, I think, for the British Computer Society to achieve the necessary status so that their members could apply for Chartered Engineering status.

Abbate:

But at the same time, I remember you once told me an anecdote about people not wanting to be too associated with the hardware. I think you said you were the only person in the department who had a screwdriver, or something!

Keller:

Yes, yes! This department is funny that way, because I am the bits-and-bytes person, or as my colleague Darrell used to introduce me: “Laurie is our bits-and-bytes man!” [laughs.] I suppose [it was] because I was the only one who ever had any hands-on experience with the equipment. Maybe the anecdote comes from the fact that Darrell used to have several machines in his office. This was at a time when not everybody had machines in his office, but his looked like the deck of the Starship Enterprise; and of course things wouldn’t work, and he was in the next office, and I’d hear this tremendous THUMP and a loud and naughty word! [laughs.] I’d go in, and usually something had knocked loose off the back—because he was always thumping things [laughs]—and I’d get my screwdriver out and tighten it.

But it was funny. I think it was more like they wanted to be at the mathematical end of it, the very abstract end, and somehow, actually understanding how the machine worked was just not there! [laughs.] You know, knowing how the machine works: that’s not interesting. It was at a time when formal methods were the coming things, and a lot of these people were very interested in formal methods. We still have a formal methods group here, but it kind of died the death, because it’s not eminently practicable in most circumstances: it was over-sold. But at that time, I think there was very much the thing of “Oh, discrete mathematics, formal methods: that’s the way of the future. We need not get our hands dirty on these machines.” So the fact that I was the one who could understand how the operating system was working; could, if necessary, get out my screwdriver and tighten things up so that the connections weren’t [loose]—I actually understood that connections need to be tight!—was quite funny.

Abbate:

Yes. It just seems ironic to me, because I think of that as something that would be more stereotypically male, to be the one doing that.

Keller:

Yes.

I think I’m still the only person here who has a set of screwdrivers in my desk drawer [laughs], except for now we have support people downstairs who have screwdrivers.

Abbate:

I’m wondering if that’s a cultural thing. Do you think that’s different between the U.S. and the U.K.?

Keller:

It might be. It might very well be. Engineering, oddly enough, in this country has a funny image, in that the word “engineer” is used where Americans would use “technician.” So sometimes when you say “engineering” to people, what they picture is the guy with the oily rag in his pocket, coming along to do dirty things underneath a piece of equipment. There’s also the professional engineer, but professional engineers will always say “professional engineering,” just to differentiate themselves from this. The word “technician” is not greatly used here. So I suspect that may be one element of it: “I don’t want to be seen as a technician.” And yet oddly enough, there is this thing about computing becoming an engineering discipline, and people now calling the software development process “software engineering.”

Abbate:

Are there other differences you’ve noticed between computing culture in the U.S. and the U.K.?

Keller:

Not a great deal. I suppose one thing that does come out is, I think Americans are slightly more inclined to just jump in and try it and see if they can do it than Britons are—although there’s a certain subgroup of British men working in computing who are very much the [type who say], “Well, let’s go in and hack a solution really quickly and see what we get.” But then they generally tend to lose interest in it, and it passes to someone else to get it working right! [laughs.] So there is a bit of a geeky culture here; but that’s much more recent, I think. I wouldn’t have said that was true twenty years ago. I think twenty years ago it was much more: it’s a different—it’s a British way of doing things. I mean, Americans come to a river and they think, “We need a bridge,” and so they go looking for logs and throw logs across the thing. The British get there and they look at this and they think, “Right: let’s get out the paper and pencil, and measure the distance across the river, and start . . .” [both laugh.] I think they both end up in the same place, oddly enough, but I think the approach is [different]. They were more different than they are now.

Abbate:

Hmm. I guess the timing [of that change in approach] sounds like it’s post–personal computer. Is there a correlation with technology?

Keller:

Probably. I mean, there’s probably a correlation there, but—correlation not being causation—I won’t say! [laughs.]

Abbate:

Who can say?

Keller:

Yes, who can say? But I think that’s where it came in.

Abbate:

Can I ask a few general questions?

Keller:

Sure.

Changes in Computing Over Time

Abbate:

I probably asked about changes in computing over time, which obviously are legion, but what do you think are the most striking changes since you started?

Keller:

Well, the most obvious one is the size factor. When I started, even a modest-sized computer required a room—and a good-sized room at that—to put it in. I remember the first one I worked on at Computer Sciences Corporation; they had just bought this, and it had 128K memory, and I remember everybody sort of went down and peered through the window at this. Of course, it looked just like the one next to it, which only had 64K memory, but [people were saying], “Gosh! It’s got 128K memory!” And that one [points to personal computer on desk] has sixteen megabytes; that one [points to another] has even more.

This shrinkage in a sense makes a lot more things possible. Things were always possible; for instance, you could always do text processing on a system like a mainframe; but the fact that you’ve got lots of memory now means that you can do text processing on very large documents, because you’ve got the whole of memory that you can load these things into, and you can do things that would have been very, very difficult. You could have done them, but they would have taken a long time, and they would have been difficult before. Weather forecasting is an[other] outstanding example. At one time, sure, you could use computers to do weather forecasting; the trouble was, it took twenty-six hours to forecast the weather twenty-four hours from now, which meant it was nothing more than a curiosity. Now these things happen very [fast]. Fancy graphics is another thing. You could have done them before, but it would have been very, very slow: too slow to make it worth your effort; you can do it faster by hand. But now you can do amazing things, because you’ve got the speed, and you’ve got this size now, which means that you don’t need to give over whole buildings to these kinds of things. That’s probably the most noticeable [change], in terms of what would strike you visually, over the time, as well as just changes in culture.

I think the other thing I’ve noticed is the real change from what you might call “static” computing to “dynamic” computing. With static computing, you had files, and you might process them once a day or once a month, and you’d start at the beginning and you went through to the end: batch processing, basically. And although that’s still around, it’s been increasingly replaced by transaction-type processing. Your banking now: when I first had my bank account and I went into the bank and I said, “What’s my balance?”, they went to printout from yesterday’s computer and told me what it was at close of business yesterday, but they couldn’t tell me how many checks might have come in today. Now I can know pretty well up to the minute what my balance is, and I don’t have to go in to the branch anymore; I can go to a hole in the wall at my supermarket or the gas station or something.

And the ubiquity of [computing], now that it’s moved into all kinds of things. I mean, my washing machine now does things like weighs the wash and decides how much water should go into it! [laughs.] It’s a nice water-saving thing; and it’s got a delay button on it, so if I don’t want to listen to the noise, I can have it start in the middle of the night. It’s got all these sorts of things that are now hidden. Your car, when you go along: your braking system, your fuel system and stuff, is all controlled by computers. God help you if you break down in the middle of nowhere, because you’re no longer able to get out and fix it yourself with a piece of tape or some chewing gum! But it’s control at a degree that you couldn’t possibly have hoped to control it before, and it’s made things safer and more reliable—most of the time.

I guess the other really big change is the fact that when I started, I think there was something like not quite a quarter of a million programmers in the United States, and people used to do things like walk around with punch cards in their pockets to let everybody know that they knew about these things! And that’s gone from that to computers in a lot of homes—not that people program them, but there are computers in a lot of homes—and certainly the use has gone from this little coterie of specialists out to virtually anybody. Because again, even people who don’t have a computer probably use an automatic telling machine, or an automatic pump at the gas station, or something like this. They may not think of it in terms of “There’s a computer embedded in there somewhere that’s doing all this,” but there is. I don’t know if this happens in the States now, but here, for instance, you can swipe your card through the pump when the station is closed . . .

Abbate:

Oh, right. Yes.

Keller:

. . . and key in your little PIN number, and get your gas, and it actually debits your account; so basically, your gas pump is wired to the bank through a network. It’s a very wired world. And I think the change that is coming . . .

[Interrupted by phone ringing.]

Anyway, I suppose the last thing that’s happening now is, not only are these things now hidden in things like your washing machine and so on, but things like your telephone and your television are beginning to take over functions that formerly a desktop computer would have done. So you can send and receive email through your television, and shop on the Web through your television, and you can get Web-enabled phones to—I don’t know what you’d want one for, frankly, but you can get one! [laughs.] I see that as being more and more the case, and I think what’s going to happen is: for a lot of people, the machine that’s sitting there on the desk at home, or even at work, is going to disappear in favor of other smart appliances that you’re going to be able to use to get what you want. Now we’ve got desks that are designed specifically to hold computers; maybe in ten years we’ll be back to having a phone, for instance, with a screen and a little keyboard, and you can do everything through that, and you won’t need something like this [indicates desktop computer].

Reflection on Women in Computing

Abbate:

How do you think the field has changed for women over time, in terms of opportunities?

Keller:

I think that’s funny, because when I started in it, and certainly people I’ve talked to who started in the ‘50’s and even into the ‘40’s: it wasn’t common for women to go into it, but if you could, you were sort of welcome. It was basically, “Anybody who can do this is welcome.” You might not be paid the same; you might not have the same terms and conditions of employment; but you could get in very easily. There were not a lot of women around, but a fair bit. I would have said at some places I worked it went up to, say, twenty-five percent, occasionally more. It was interesting, because when I moved into systems programming, I moved from an area where there were a number of women, which was the applications area, into an area that was almost all male. And yet when I worked for TRW, they made a point of promoting women into that area, so again it tended to be—I couldn’t do the figures in my head, but I’d say almost fifty-fifty male and female, in the systems programming area. And then you’d go to something like SHARE and GUIDE and discover, “I’m here alone!” [laughs.] So it was very spotty, but you could get in, and you certainly could make your way. It depended a lot more, I think, on personal aptitude and personal interest than anything.

And then there was this decline that happened in the ‘80s—particularly in this country, but I think it has happened in the States: what they refer to as the “shrinking pipeline.” For instance, here, in ‘81, approximately twenty-seven percent of the places in university computing courses was taken up by women; and that just went right down to next to nothing over a very short period of time. By ‘84, you were really lucky if you could pull eight or nine percent, and it was still dropping. Some effort was put in by the government in the ‘90s, throughout the ‘90s; but just recently I went to a thing put on by the government about getting more women into computing, and they were saying exactly the same things that they’d said in the early ‘90s, and this obviously isn’t working. Partly it’s that wonderful British business of, “We’ll all agree to do this, won’t we?”[laughs], rather than any kind of regulation or attempt by the government to push things any harder than this kind of gentle “This would be a good idea” type of approach. So it’s still low. I mean, this university has probably one of the highest proportions of women on computing courses of any British university, and it’s not high. This department is unusual, though, because it has the highest proportion of women of any computing department in the U.K.

Abbate:

On the faculty, or the students?

Keller:

As a faculty. It’s about 40 percent female, 60 percent male, which is quite high. And we’re now starting to get women working through to senior posts: so we’ve got a female Professor; we’ve got me as Head of Department and Senior Lecturer; I’ve got another female Senior Lecturer joining next week—she used to work here as a Lecturer, left, and is coming back as a Senior Lecturing post; I’ve got a female Senior Research Fellow, who is one of three Senior Research Fellows. So we’re beginning to bubble up; but when I first joined the faculty, the faculty as a whole had eight percent of the senior faculty [who] were female—no Professors—and thirteen percent of the faculty as a whole was female. So it’s now much greater, and there are more women in positions of at least some power. For instance, I’m not the only female Head of Department in this faculty; there’s another one. Out of six, there’s two of us.

Abbate:

Not too bad.

Keller:

Not too bad. We’re getting there.

Abbate:

Do you have ideas about either why there’s been a decline in women, or what would increase the numbers?

Keller:

Well again, going back to the things that I said to the combined engineering faculties of western Australia: there’s still a lot of emphasis on trying to push computing as an engineering discipline, and that word gets used a lot. That could be one thing that’s off-putting. Oddly enough, women have taken up mathematics. When I was in school, that was quite rare, and if they did, they became mathematics teachers. Now a much greater proportion of mathematics students are female than was the case in the past. So it’s not a math phobia thing that’s operating here.

Abbate:

Right.

Keller:

There is, I think, a kind of perception that . . . I certainly find this with school girls: they’ll say, “Oh, women can do this,” but then they’ll turn around and say, “but I don’t want to.” So there’s the “We can, but I don’t want to” syndrome going on. I think that’s partly looking at brothers or other boys and seeing these people sort of sitting there day and night, glued to the screen, playing dumb games or hacking out programs and stuff, and not really having a rounded life. That puts people off—or it puts women off; girls—because I think they sort of feel, “Well, I can do this, but there’s no human content in it. There’s no feeling that I would have a social life around this. This is all just boring, sitting there staring at screens and tapping away on keyboards and playing with joysticks and stuff”—which isn’t true at all, but that’s the impression I think that girls get.

One of the women who did research that contributed to this particular paper noted, that although it’s now the case in Britain that girls probably have about equal access to computers in the home—she was talking about young children, up to the age of thirteen—it’s most often sited either in a brother’s room, or in a room that’s shared by the whole family, rather than in the girl’s room. So there aren’t that many girls that have their own machine in their own room; they have to bother somebody—you know, there’s this whole disturbance. So, for instance, if it’s in a public space within the house, they might feel that it’s okay for them to go and do their homework on it, but not to sit and fool around with it, because it’s a family thing; and it’s there, and they don’t want to be seen, maybe, fooling around.

And I think there’s still a lot of peer pressure with girls, particularly in their teens, to do the sorts of things that girls are supposed to do, despite interests. It’s still very common—sadly common—in this country for girls, even bright girls, when you ask them at fifteen what they think they’d like to go into, they say “Hairdressing,” which is probably one of the lowest-paid jobs in this country! And I think that’s partly peer pressure. I know I felt it when I was that sort of age. I was never interested in fashion or anything like that, but everybody else was, and you sort of feel like you’ve got to go along with it, even if you’re not especially interested in it. It takes a while before you get old enough to realize, “I can be what I want to be, and if I don’t want to wear lipstick, I don’t have to wear lipstick!” [laughs.] I think that kind of thing is operating just at the point when girls are actually making the choices that are going to be important to them later in life. In this country, I think you do find that there are more mature women coming to computing than there are girls coming straight out of school into computing degrees. Because mature women are sitting there: they’re a little bit more confident about themselves; they’ve got over the thing that they have to spend a lot of time with fashion and things; and what they’re looking for is something that pays well.

Abbate:

Do you think the education system here tracks people at an earlier age, in the sense that you have to decide at sixteen which exams you’re going to take?

Keller:

Yes. Yes, it does. This is the great British compromise! [laughs.] Some years ago, they thought, “We’re going to do away with this business [of Eleven Plus exams].” Kids used to sit something called the “Eleven Plus,” which decided whether you went to a grammar school, which was an academic school, or you went to a secondary modern school, which was the sort where they taught you enough to be a decent plumber. I mean, deciding that at the age of eleven, on the basis of an examination, is [unreasonable?].

Abbate:

Was this O Levels?

Keller:

No, that was called the Eleven Plus. O Levels happen at sixteen.

Abbate:

And then A Levels are at eighteen?

Keller:

A Levels are at eighteen. So they go on for a little bit longer.

Abbate:

They got rid of the O Levels, didn’t they?

Keller:

The O Levels has been changed to something called “GCSEs”—General Certificate of Secondary Education. But “O” stood for “Ordinary” and “A” for “Advanced.”

Abbate:

I never knew that!

Keller:

It makes a little bit more sense when you know that; but I must have lived here for fifteen years before I found out! [laughs.]

There are still places—this county, for instance, is one—where there is still the Eleven Plus, but they now have it mixed in with the comprehensive school system. The old system, as I said, was [either] grammar school, which was the academic one—you’d go to grammar school and you’d study Latin and Greek and mathematics and science—[or] you’d go the secondary modern, and the mathematics would be add, subtract, multiply, and divide; English might be just learning how to write a business letter; so it was fairly low-level. And, at least when that system first existed in the forties (or even earlier), to go to a secondary modern was to be condemned for the rest of your life to work in a factory. It later changed, and that wasn’t quite true. For instance, my partner went to grammar school—he passed the Eleven Plus and went to grammar school; neither of his younger brothers passed it, and both went to secondary modern. The middle one went at a time when going to secondary modern just killed you off forever. I mean, he works in computers, but he’s come up through the sort of oily rag, good-with-your-hands route; but the younger brother who failed it nevertheless went on to Cambridge and got a Ph.D. So at some point it stopped condemning you to this life in the factories, but it wasn’t . . . . Anyway, the British system changed partly to what they call the “comprehensive” system, which is much closer to the American system in that everybody, regardless of ability, comes to a single school. The school then teaches a wide range of subjects, which range from the vocational through to the academic. But there’s still this mix: this county still has grammar schools and the Eleven Plus, but also has comprehensives. So you get that in a few counties. A lot of them have comprehensives.

They’re now looking at changing the system yet again—and these just get put on as sort of additional layers, so that they have these magnet school–type things, where they specialize in, for instance, theater, or they might specialize in science or something like that. And these are just layers on top of a system that I don’t think anybody ever designed in the first place anyway! [both laugh.] But even at that [type of school], you do your GCSEs.

At fifteen to sixteen, you would sit your GCSEs, and you’d probably do nine or ten, and they would cover a wide range of subjects; so they might cover geography, history, foreign language, general science: anyway, a wide range of things. But then you have to select your A Levels, and you normally would do three A Levels, and in order to get into a good university, you really need to do three related ones. So people will do things like physics, chemistry, and mathematics; or they might do history, geography, and economics. They stop taking foreign language; they stop taking mathematics, if they’re not in a mathematical stream, as it were; so there at sixteen, suddenly you’re cut off from that breadth, and you start specializing. And then when you get to university degrees, in this country they are very specialized—much more specialized than they are in the U.S. You don’t have options.

Abbate:

So at sixteen, you’re making a decision: are you going to take any more maths or not?

Keller:

Yes.

Abbate:

It’s always struck me that for sixteen-year-old girls, who are at a very vulnerable point in terms of identifying with the sort of ultra-feminine kind of image, that’s the last time you should be asking them to make that choice.

Keller:

Absolutely! Absolutely. And interestingly enough, there’s a lot of research here that shows that girls who went to single-sex schools actually were more likely to go into the sciences and things than girls who went to mixed schools.

Abbate:

Oh, it’s the same in the U.S.

Keller:

And I suspect it’s the influence of the notion of, “Gosh, I’m just in puberty; I’ve got to attract the opposite sex,” and you get very wound up into this sort of femininity/masculinity divide. I think boys suffer equally, perhaps, but just less obviously—because the things that they end up doing tend to be better paid and more prestigious! [both laugh.] But it’s just exactly at that point when you shouldn’t be asking them to make a choice, and that’s where you get these girls saying things like, “Oh, I’d love to be hairdresser.”

Abbate:

Or they want to be a Spice Girl.

Keller:

Yes! Yes, you get a lot of that. And people who want to go into theater or music or something like that—not because they’re going to be any good at it, but just because they think that this is something to do. I’m afraid schools here quite often get people who don’t have that degree of talent, who come along and say, “Oh yes, I really want to do theater studies,” and they couldn’t act their way out of a paper bag—which is sad, I think.

They do get some guidance—I mean, schools do try to guide people—but then you run into the problem that sometimes the guidance isn’t wonderful, either. Sometimes it’s very old-fashioned. I know when I was at school, they had a program whereby the bright students could go and study at UCLA for one hour a day, I think it was; and I applied for that, and they told me that it was only open to boys! [laughs.] And I remember asking my counselor whether I should go on and do mathematics, and he said, “Oh, you’ll never need it.” And it was just this assumption: “Girls don’t do mathematics, so you’ll never need it.” I don’t know what he thought I was going to end up being, but I could have used some more dispassionate advice, I think, at the time! [laughs.] Because unfortunately, it reflected my prejudices against mathematics. I was not bad at it, but I hated it, so somebody sitting there saying, “Oh, you’ll never need it again” was just exactly what I wanted to hear. And I think you still get some of that, although schools, I think, do try quite hard to foster people into the right direction. But you’re working against peer pressure, which of course is enormous, and sometimes parental pressure as well.

Abbate:

One of the other issues, of course, is women looking ahead and thinking, “Do I want to have a family, and is this career going to be compatible with that?”

Keller:

Yes.

Abbate:

I don’t know if that was ever an issue for you, in terms of either having kids or maybe taking care of other family members.

Keller:

I don’t think I really thought . . . Well, I suppose it must have crossed my mind, because . . . Well, it did; but I think at seventeen I decided I would never get married. [laughs.] And here I am on Man Number Three! [both laugh.] You know, I decided I would never get married, and it was a decision that I made at seventeen. So I sort of said to myself, “As a result of that, I will never have children.” So I’d thought through it that way, I think.

I think then later, when I did get married and at various points thought maybe a family should be coming up, I had the model of my mother, who as far as I knew had always worked—well, she had. So it was just one of those things that, if it happened, you had to super-organized, and as my mother did, you hired other people to step in for you. So for instance, when I was really little, the lady next door took care of me while my mother was out at work, and when we moved away from there, she had a woman come in at three o’clock in the afternoon, so when my brother and I came home from school there was somebody there until she got home at five-thirty or six. So to me, that was [normal?]; you would do that. But I would imagine for a lot of people that’s not the case, and it can be inhibiting when you’re thinking about your future—especially if you’re sitting there at sixteen, and if you’ve been used to a mother who’s been at home, or at least was at home when you were small, and you sort of think, “Oh, gosh. If you have children, you’ve got to at least take out time until they start school.” It must actually be quite difficult to think what those coping mechanisms might be.

Abbate:

Do you think computing is, and/or seems like, a field that can accommodate family responsibilities?

Keller:

Yes. After I’d been in it for a while, I began to realize that that was true.

Abbate:

That it could?

Keller:

That it could accommodate things like this. But I remember at one point, I’d left one job, which was—I can’t remember—but anyway, I was traveling for a few months, and I stopped in for some reason at the American consulate in Dusseldorf to have some change made to my passport, and Time magazine was there. I hadn’t seen any news for a month or more, and I picked up Time magazine, and it said “IBM announces the System 370.” I’d been working on 360s, and I remember sitting there thinking, “Gosh! When I get back, I won’t be able to get a job at all!” [laughs.] In fact, they were very similar; there were lots of similarities, and anyway, all the 360s didn’t suddenly disappear the day that the 370 was announced. So I used to think, “It’s such a fast-moving field that if you’re not right up to date, you’re going to have real problems,” [but] I don’t think that’s actually true if you get a good grasp of the fundamentals; you can always teach [yourself]. You might be a little bit behind if you stay away for a while, but you should be able to pick things up, because you’ve got the basic skills and you’ve got the basic conceptual knowledge.

So I do think it’s actually good that way, and certainly there are lots of opportunities—not necessarily there, but could be there if employers were willing—for part-time work. In fact, I’ve got a couple of mothers in this department who work shorter contracts. They work eighty-percent contracts, because they want to be home with their children. So it enables them to come in and do that, and yet I can still get very useful things out of them in that eighty percent. We had one who’s recently been put on a full contract, but she was fifty percent. So that has worked very well, and I don’t see why it wouldn’t work elsewhere.

Reflection on Working with Computers

Abbate:

What have you found most satisfying about computers, or working with computers?

Keller:

The sense of power and control that you can get! [laughs.] When I worked for TRW, I went in, one of these long weekends, to do some testing on the operating system. I was there all by myself—there were no operators; nobody—and I went in there, and there’s millions of dollars' worth of equipment. I turned on the lights, and I started turning on this equipment and listening to the motors start up and the air conditioning start up, and I remember having this enormous sense of power: just standing there and thinking, “This is all mine! I can make it do anything I want. I’m here by myself; there’s nobody to tell me anything.” And it just gave me such an enormous sense of power. It was fairly momentary, but it was just so real. I can still—you know, [the way that] years afterwards you can taste a really good meal: I can still taste that sense of going in there and turning on tens of millions of dollars' worth of equipment; turning that all on and realizing I was in sole charge of this. You know, just this enormous sense of pride. It isn’t quite as good with these little machines! [Indicates desktop computer; laughs.] Even though this has probably got more memory, and is faster and everything else, it’s a desktop machine. But again, I’m a sole charge: I can make it my machine, make it do what I want. So there’s an enormous sense of power and control that’s there. I know that doesn’t sound like a terribly feminine thing to say [laughs], but that’s one of the things I have really, really enjoyed.

The other thing I think I’ve got out of it is, it’s given me enormous confidence in dealing with non-computing systems. At one point, believe it or not, my toilet broke down, and it was at a time when I had very little or no money, and I just thought, “I cannot afford to have a plumber come in and look at this.” You can always flush a toilet by just pouring a bucket of water in it, but I couldn’t see myself doing that indefinitely, and I was really uptight about this and feeling really depressed. Then suddenly I thought, “Well, wait a minute. If I can make a great big mainframe do what I want; if I can write a program to do what I want, why can’t I work out how a toilet works, and work out what’s wrong with it and fix it?” So I took the top off, and I started jiggling the handle and pulling on pieces, and I eventually realized what it was, and I went down and got the little fifteen-cent part and I put it in, and it worked. And ever since then, actually, I have felt much more confident with electrical or mechanical systems.

Abbate:

[You would think,] “I can debug this.”

Keller:

Yes: I can debug it. You know, I may not be able to design one; I may not be able to build one; I may only be able to debug it and then have to take it to an expert to fix, because it requires special tools or I just haven’t quite got the skills; but they don’t faze me anymore. So that sense of confidence has really been something that I’ve enjoyed a lot and got a lot from.

Advice for Women in Computing

Abbate:

Well, just to wrap up: Do you have any advice for young women thinking about going into computing?

Keller:

Yes. Go for it! [laughs.] I mean, no two ways about it: it can be really, really satisfying; it pays very well; I have been able to work on both sides of the Atlantic. If I’d had really good [language skills]—I mean, I do speak German and I speak Spanish, but if I’d had really good foreign language skills, I could just as easily have gone to Germany or Spain or somewhere else to work, because in that sense it’s wide open. It pays well. It’s not difficult to find positions; you can’t always find them first go, like you used to be able to, but there are still more openings than there are people to fill them. There’s a lot of interesting things going on, and that’s not going to stop: it’s just going to continue, I think, being really, really worthwhile.

The other thing I would say is: because these systems are so often designed by men coming from a male point of view, I think there is a different, female point of view; and technology now controls an awful lot of our lives, and I would like to see more of that female point of view there in the design stages—there and saying, “What do we actually want to do with this? What end are we trying to achieve with this?” Because I think it would be a different perspective, and would make the technology different, and make it far more interesting, and maybe more useful.

Abbate:

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

Keller:

Okay!