Oral-History:Jean Bacon

About Jean Bacon

Jean Bacon was born in 1942 in Sheffield, England. She attended Penistone Grammar School and Royal Holloway College, University of London, graduating with a B.S. degree in mathematics. She first used a computer for a summer job and soon after decided to pursue a career in computing. Bacon worked at the National Physical Laboratory and the GEC Research Centre before she became a lecturer in 1968. After teaching at Watford College of Technology for five years, Bacon moved to Hatfield Technological College. There she participated in the design of a computer science degree program that was among the first in the new discipline in the United Kingdom. While working at Watford and Hatfield, she also earned a M.S. degree and a Ph.D. degree in computer science. In 1985, Bacon joined the faculty of the University of Cambridge Computer Laboratory, where she currently is Professor of Distributed Systems. She leads the Opera Research Group, whose research interests are in open, large-scale, widely distributed systems. In addition, she is the Director of Studies in Computer Science at Jesus College. Bacon is also a Fellow of the British Computer Society and the IEEE. She has served on the IEEE Computer Society's Board of Governors and was founding Editor-in-Chief of IEEE Distributed Systems Online.

In this interview, Bacon discusses her career in computing. She talks about her experiences as an academic computer scientist, sharing her opinions on the advantages and disadvantages of this role. As both an early computer science student and instructor, she offers a unique perspective on the evolution of the discipline. In discussing her involvement in computer science education since the 1960s, Bacon comments on various topics such as the relationship between academia and industry and the status of women in the field. She also talks about her contributions to the international distributed systems community.

About the Interview

JEAN BACON: An Interview Conducted by Janet Abbate for the IEEE History Center, 10 April 2001.

Interview #574 for the IEEE History Center, The Institute of Electrical and Electronic 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, 39 Union Street, New Brunswick, NJ 08901-8538 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:

Jean Bacon, an oral history conducted in 2001 by Janet Abbate, IEEE History Center, New Brunswick, NJ, USA.

Interview

INTERVIEW: Jean Bacon
INTERVIEWER: Janet Abbate
DATE: 10 April 2001
PLACE: Bacon's office at the University of Cambridge, Cambridge, United Kingdom

[Notes courtesy of interviewer Janet Abbate]

Background and Education

Abbate:

To start out with, when were you born, and where?

Bacon:

1942, in Sheffield, England.

Abbate:

What did your parents do for a living?

Bacon:

My parents . . . My mother was just a housewife, and my father was a steelworker.

Abbate:

Did you have brothers and sisters?

Bacon:

No, I was an only child.

Abbate:

Did they have any interest in math or science?

Bacon:

Oh, not at all. No, they . . . I come from a working-class background; the whole family were really very working-class.

Abbate:

What kinds of schools did you go to?

Bacon:

The North of England have very good grammar schools, and this has worked over the last hundred years or so; and it’s been a very good way of getting bright working-class children well-educated, and a step up towards university. So I went to one called Penistone Grammar School, and of course got good A-levels, and then went on to university.

Abbate:

Was that a girls’ school, or mixed?

Bacon:

No, it was a mixed school.

Abbate:

Did you know early on you were interested in maths or science or computers?

Bacon:

I was always very good at maths, and it was . . . I was very interested in English and the arts, and I found it very difficult to choose, really; I would have preferred a broadly based course. But there were a lot of choices all the way through; you had to select various subjects, and I suppose in the end, the fact I could do maths well . . . so I did maths. And then, if you do maths, you have to do a science with it, rather than—I would have liked to do something like double maths and English, like they do nowadays, but at that time you couldn’t do that; so I feel quite envious when they come up for interview with a very broad range of subjects.

Abbate:

What science did you end up with?

Bacon:

I did physics—just double maths and physics. That was all one could do, three A-levels, at that time.

Abbate:

So, pure and applied is the double maths?

Bacon:

No, it’s just like maths and further maths. Yes, you do a broader-based math course, then you do a second math course that’s deeper in every way. I think it’s still on offer, further maths at A-level. Whereas nowadays, you can do selected modules in statistics and different things: applied and pure, and so on, which is slightly different. It’s broader, but not necessarily deeper.

Abbate:

You ended up going to the University of London?

Bacon:

Yes.

Abbate:

How did you end up there? Or why did you choose that?

Bacon:

I don’t know, really. It’s difficult . . . It was just suggested at school. I was very young, and they said, “This place will take you when you’re seventeen,” and I did a scholarship exam. For some reason the school were very concerned that you did scholarship papers and got a scholarship to somewhere, rather than just going, applying. And so, they put me up for this and I got a scholarship and went. I didn’t even try for Cambridge, or Oxford, which is a pity, because retrospectively I think I’d have really enjoyed it.

Abbate:

Which college did you go to?

Bacon:

It’s called Royal Holloway College. Quite a long way out to the west, and it’s a women’s college, or it was at that time. It no longer is. And that was very different, because I was used to a northern grammar school—coed place—so that was very different.

Abbate:

What was it like?

Bacon:

Well, I regret going there rather than trying for Oxford or Cambridge; but it was only three years, and it comes and goes, and it was a good degree, really.

Abbate:

So that was maths again.

Bacon:

That was maths.

Using a Computer for the First Time

Abbate:

At what point did you first use a computer?

Bacon:

It was very briefly, during a vacation job [either] the year before I went up to university or the first year at university. It was in a place called Cyber House, which is a research center associated with United Steels in Sheffield. Somebody called Stafford Beer was there and he was very much into cybernetics and work study and such things; I think he was very forward-looking for that place. There were computers being introduced into United Steels at that time, and I was doing some numerical things, and people around me were starting to look at how to program computers, and they sent me on a short course at Sheffield University. So it was about ‘60, ‘61. That would be assembler kind of programming.

Abbate:

So they trained you then, and you did something for them?

Bacon:

That was only a vacation placement, just doing sort of dog’s body mathematical-type jobs, over the summer holiday.

Abbate:

But you learned assembler for that?

Bacon:

I went on a short course. I wasn’t really able—they didn’t allow me onto the few computers at Cyber House. It’s just that the culture was starting to be established there. And then when I first came to use computers was at the first job at National Physical Laboratory, after graduating.

Abbate:

Did you, after that summer, after that vacation job, did you think, “Well, I want to use computers?”

Bacon:

Yes, yes I did, really. I was in a very traditional maths degree. There was no statistics, nothing at all broadening it, and I was rather keen to trying something along those lines. So NPL seemed a very good place.

Working at the National Physical Laboratory

Abbate:

How did that work, getting that first job? Did you look for a lot of them? Or did you just hear about this one?

Bacon:

Well, I looked . . . I suppose there must be places where these things are advertised; maybe they’re sent out to students. I was looking for not one specific commercial company, but—the government research labs at that time seemed a very nice general thing to do, not go straight into teaching, and not go into commerce, but do—try government research. I think I applied for Central Electricity Generating Board and NPL, and NPL offered first, so I went there.

Abbate:

That was maybe ‘63 or so?

Bacon:

It was, yes. I can’t . . . What did I say? ‘63? Yes: ‘63, ‘64. In August or September of 1963, when I was age 20.

Abbate:

What did you actually do at the NPL?

Bacon:

NPL: I wasn’t there for very long. The first thing they did was make everyone do a desk machine course, to learn numerical methods. And so it was a Brunsviga course, where you turn the handle.

Abbate:

Mechanical calculators.

Bacon:

Mechanical, yes, that’s right. And I think it’s just the program for people going in; they have to do that. One interesting thing at that time was that they would not employ women as what they call Scientific Officers, so I was what’s called an Experimental Officer, or an Assistant Experimental Officer—and I think soon after I left, they changed the policy, and they started to allow equal entry into the scientific civil service grades.

Abbate:

What was the difference between Scientific and Experimental [Officers]?

Bacon:

I wasn’t really there long enough to say. I imagine that if I’d stayed, it would have been that the Scientific Officers are supposed to lead projects, and the Experimental Officers are probably supposed to write programs that they’re told to write. But I just thought I would go there anyway, because it sounded an interesting place.

So, the first thing was to do the desk machine course. When you’d gone through the induction course, it was then you’d start to write programs, and the first computer was the ACE computer. Way back in the late ‘40s they had the pilot ACE there, and that worked and was used in the early ‘50s, and then the DEUCE machine followed on from that, and that was still there when I was at NPL. The computer I was programming, the ACE computer, occupied an enormous hall. The instructions were stored in mercury delay lines, and it had a long instruction format, and one of the fields was the address of the next instruction. So as the mercury delay lines rotated, you could pick up the next instruction from wherever the current read place was, so if you programmed very carefully, you could make the program run faster by making the next instruction just in the right place for where you were currently reading. It was a skill that was relevant at the time! [laughs.] It soon became obsolete, of course.

Abbate:

So that was very intimately tied to the hardware.

Bacon:

Absolutely, yes!

Abbate:

And this was some sort of assembler or something?

Bacon:

Oh yes! Yes, that’s all, all you could do at that time. Close to machine code.

Abbate:

What was the application?

Bacon:

I wasn’t there very long; I can’t really remember what I was asked to do. I would just write programs that I was told to do; and I left after a year.

Abbate:

So you would be doing some mathematical function . . . ?

Bacon:

Some mathematical—solving some mathematical equations, yes; linear algebra or something like that. I think it was more like linear equations than differential equations, if I remember rightly; but I don’t really remember! [laughs.]

Abbate:

Was it interesting?

Bacon:

Well, it was interesting to learn what computers were like then, as one perceived it, and how to make them operate.

Numerical analysis was the focus of that laboratory, and so they were publishing papers in how to compute the eigenvalues of a set of linear equations, and that was one of the important things they were doing at that time. Solving differential equations; numerical methods. The desk machine course that I did was a throwback to the tables they used to publish, all that kind of thing; and they were just moving on to doing everything by computer.

Abbate:

But you only stayed for a year.

Bacon:

Oh, well, this was the two-body problem: I met my husband, and he was also an Experimental Officer, but had done his degree and got his degree and wanted to move up to do a more prestigious job, so he moved, and he got a job over at Dollis Hill.

Abbate:

What is that?

Bacon:

It’s the Post Office research center.

Abbate:

Oh, right, right! They did the Bletchley Park[1] stuff, isn’t that true?

Working at the GEC Hirst Research Centre

Bacon:

Yes, that’s right. I mean, he couldn’t really say what he was doing, because he was in that sort of area. I think he still can’t really say what he did! [laughter] Yes, so he went into that area, and so I was looking for a job in the same geographic area. So the next thing I did was GEC Hirst Research Centre at Wembley, close to Dollis Hill.

Abbate:

Which is what?

Bacon:

It was looking at theoretical telecommunications.

Abbate:

Who runs the Research Centre?

Bacon:

GEC?

Abbate:

Oh! The General Electric . . .

Bacon:

The General Electric Company, yes.

Let me see if I’ve said everything about NPL . . .

If you’re interested, I can kind of dredge up a lot of names from the NPL, but you may know them already.

Abbate:

I don’t know, actually. Were there a lot of other women there?

Bacon:

No, not very many at all. I probably off-line can think of names and people who were there.

Well, GEC Hirst Research Centre, that was, as I’ve said, theoretical telecommunications, and the leader of the group who employed me was George Medhurst. He got me working on designing aerials, and he was looking at directive aerials; it was called a super-directive aerial. And so it was a complete switch, really, from what I had done at the NPL. So it was just a matter of following your husband, getting a job, doing what they want you to do; and I’d got programming skills by then, so that’s what I did.

Abbate:

What sort of computers were they?

Bacon:

Now, what they did was hire time on the London University Atlas computer, which was a very famous computer; so we used to travel up to Gordon Square and put our card jobs over the counter and sit there until they were run, and then get the results back. So that was the way. [laughs] Very slow turnaround time!

Abbate:

That must have been . . . How long would that trip have been? Was that a couple hours you’d be traveling?

Bacon:

An hour or so. Wembley’s quite close to London, and you’d get very close to Gordon Square when you got on the Tube, so it’s not too bad.

And then I worked with someone called Mike Clayton, who’s still at GEC Research, and that was on fading in frequency—frequency modulation systems (analogue systems). Things like pulse-code modulation (PCM) were just being looked at, were just starting (digital stuff).

Abbate:

Oh, right. That must have been an exciting time.

Bacon:

Yes, it was, really. And they had various branches around there just starting to do optical fibers, but they couldn’t really make them go and make them commercial at that time, but later on they did.

Abbate:

Did you enjoy that?

Bacon:

Yes! I enjoyed it, yes! I went to my first conference from GEC—that was in Stresa, Italy; let’s see, it was a radio meeting—I don’t know what exactly what it was . . . But it wasn’t a natural area for me to be in at all, because I’d got a maths degree, and it was okay—it was probably more electrical engineering, communications, plus maths. So it was following my husband, really, rather than choosing what I would do next.

Abbate:

Did you get to use any of your physics?

Bacon:

Physics! [laughs.] I’ve forgotten all that! I’d not done it since A-level. But yes, it was electrical, electromagnetism, and I kind of dredged it up again. We’re thinking how electrical induction, at one part of the aerial, induces current in another part; and so the physics was—I wasn’t actually designing or anything, but I was doing analysis and writing programs. The digital analysis (PCM) was very interesting.

Abbate:

Were there a lot of other women at GEC, in computing?

Bacon:

I don’t think there were any. I can’t think of anybody. We had the occasional vacation student, but that was all.

Abbate:

Where were they based, the company?

Bacon:

North Wembley? GEC probably had commercial outlets all over the country making whatever communications equipment, but this was their research center: North Wembley—I think it was demolished recently. [laughs] But GEC did a lot of—[Arnold] Weinstock was the head of GEC, and he was very commercially orientated, and was buying companies and merging a lot, so it became GEC-Marconi. But I don’t know what happened afterwards.

Abbate:

How long were you there?

Bacon:

That was ‘64 to ‘68. So, at that stage—‘68—I was 25. I went to teach at the local tech. We were living in Watford, and I started thinking, “Well, I should have a family, maybe, and how to fit it in with working where you get a fortnight’s holiday a year?” And lecturing seemed an attractive proposition from that point of view. I was a bit more mature and more kind of inclined to go talk to people, and standing up in front of a class didn’t seem so intimidating as it would have done when I first graduated, so I thought I would do that. So I got a job lecturing at the local tech. So this was out of kind of mainstream computing research for quite a few years.

Lecturing at Watford and Hatfield

Abbate:

What were you lecturing on?

Bacon:

I was lecturing in what were called Higher National Certificates—Higher National Diplomas—the degree in Printing Technology, which was what happened in Watford, and I taught numerical methods and numerical analysis. So, all the stuff I’d done at the NPL on different ways of solving linear equations or differential equations or data interpolation, I started to teach on these courses.

And at that time, I thought, “Okay, I stopped after B.Sc.,” so I did a part-time M.Sc. at the local polytechnic. Hatfield Polytechnic was offering a part-time course. So, while I was lecturing at Watford, I did this part-time M.Sc. course, and then, after that, moved to be a lecturer at Hatfield, which was much more strongly into computing. This was in September 1973, when I was age 30. Also at Watford I taught BASIC programming, because BASIC was coming in as a programming language that students could use. And the method there was to put your tape on a van; the van went to Hatfield; they ran the program over there; the tape came back next day; and the students saw all their results! So again, it was one-day turnaround time. If the van went early, you missed your program run for the day.

Okay, so then—Hatfield is now a huge computer science department; [back] then—I can’t remember the exact numbers, but it must have grown from about 15 to about 30 when I was there.

Abbate:

Was that— The one you’d been teaching at before, Watford, was a kind of small local tech?

Bacon:

The College of Technology, it was called. Polytechnics were created around 1970. Hatfield had an emerging national reputation even then.

Abbate:

And this other one: Did it have a more sort of regional base?

Bacon:

Yes, it had, I suppose, a Polytechnic. It’s called the University of Hertfordshire now—so yes, it would have this regional base. And they were forward-looking. They bought a DEC—DEC machine—DEC System 10, running the TOPS-10 operating system . . .

Abbate:

Ah, TOPS-10. I’ve used TOPS-10.

Bacon:

Okay, right. So that was the machine that we were all using as our system.

Abbate:

So that was a time-sharing system . . .

Bacon:

Yes, that’s right.

Abbate:

Was that the first one that you had a chance to use?

Bacon:

Yes, that’s right. So the first main system I was using from day to day was a time-sharing system rather than a batch—well, the London University Atlas was a batch system, but we actually physically traveled there to use it.

And when I went to Hatfield, we still didn’t have a machine on the desk. There was always a big controversy about: if you give staff a computer on their desk, it’s not—it’s idle for all that time—because resources were very scarce in those days. So you’d have a computer room, and go and use a terminal there.

Abbate:

Because this was the mid-‘70s now?

Bacon:

Um, yes. ‘68 to ‘73, I was at Watford; ‘73 onwards I was at Hatfield. So ‘73 to ‘85. So I suppose it must have changed a lot during that period.

Abbate:

Were you helping to build the computer science course?

Bacon:

Oh yes! Yes, very much so. I was on one of the committees to design a modular degree, and we were obviously bringing in new topics all the time, because the subject was evolving.

And using new computers: I was—I taught the assembler for the DEC 10, Macro-10; wrote a manual for it. And then we decided that the students would be better off if they started to use a small machine hands-on, so we bought in a number of minis, called Digico Micro-16. That was an English company, but just—it soon died when, let’s say, Intel and Motorola and the like started producing chips in volume. But microprocessors then came in, and we then bought 68000s; and I wrote a book on the 68000 for educational use, showing how to use it, how to program it, how little programs are written; that’s Prentice-Hall. And it’s still—I still get a few royalties, though I’m sure it went out of print by now! [laughter]

Abbate:

Someone’s still buying it.

Bacon:

It still—someone still sells a few copies all this time later!

Abbate:

Was there a particular direction you wanted to see things go, as far as what students should be learning?

Bacon:

Well, I was very interested at that time in operating systems. I mean, it was numerical analysis from the early days, and when I did the M.Sc., operating systems and how they work were the thing that I became really involved in. I was interested in [the] hardware-software boundary—so, the operating system’s driving the hardware—and teaching that. So, very much how you have the software drive the hardware. That was the kind of area that I was interested in.

Abbate:

Were you doing research in that area, too, or just teaching?

Getting a Ph.D.

Bacon:

Later on I started to do research. I hadn’t at first got a Ph.D., so . . . What happened? My son was born in ‘76, and I was sort of not going out very much, and I decided that I needed to do something; and when he was about two, I started a Ph.D. So, not going out and sitting there after he’d gone to bed, I decided I would do a Ph.D. part-time. So I did a part-time Ph.D. while I was at Hatfield, while I was lecturing at Hatfield. This was very hard in that you didn’t get time off for doing this. On the other hand, the subject (Computer Science) had become an established subject, as opposed to . . . There were lots of courses in it by then, whereas when I first started, there were no C.S. degrees.

Abbate:

In computing in general?

Bacon:

In computing in general. And so people had started—the young had started to move through from doing their degrees, and most of them had Ph.D.s—and I’d sort of missed out on that, doing a maths degree very early and then switching to computing—so I thought I should do a Ph.D. It was an obvious thing to catch up. So, I did that in three years part-time, and the topic was kernel design and interprocess communication in distributed systems; I got interested in distributed computing. And the way I managed to do it was to get students doing alternative designs; it was all in assembler at this time. So one would do a message-passing kernel, an asynchronous message-passing kernel; another did a synchronous message-passing kernel; the others were trying to do a sort of more procedural model; and I got someone doing a little filing system design. So I was trying to look at operating systems and their components, but not trying to do it all myself, out of my own time; somehow bring in what I had to do—which was supervise projects—anyway. So at least I got them to concede that I could do more project supervision and less lecturing—and so I managed to do the Ph.D. in that way!

Abbate:

That was clever! [laughter.]

Bacon:

So I think that I was the first to do this; and after I finished, I kind of made the case that people doing Ph.D.s in this new area should get time off for doing it, and they began to do that after that.

Abbate:

At Hatfield.

Bacon:

At Hatfield, yes.

On Women at Hatfield

Abbate:

Were there other women on the . . .

Bacon:

On the staff?

Abbate:

. . . staff?

Bacon:

Yes, there were. A woman taught me operating systems. She’d done a course at Manchester. She was maybe working—I don’t think she ever did do a Ph.D., but she must have worked quite close to the hardware, and she may have done a research job or something at Manchester before she came to Hatfield. She knew about the MU-5 operating system, or its predecessor—I can’t quite remember, but certainly she was obviously up in her subject, so that appealed to me. And it was probably why I got into operating systems, because there was someone who’d actually hands-on worked in them . . . whereas maybe some of the other courses could have been people just reading it in books and passing it on, rather than doing it. And it does make a big difference, I think, when you’re being taught, whether you’re being taught by someone who’s done it.

Abbate:

Do you remember her name?

Bacon:

Kathy Levine.

Yes, and other lecturers: There was someone called Rita Munslow, who is retired now, and she was there before me. There weren’t very many. I would have thought, I don’t know, a fifth of the total staff: so three out of fifteen, something like that.

Abbate:

That’s not so bad. Were there a lot of women students?

Bacon:

Again, not very many. I would have thought ten to twenty percent, something like that. But . . .we were not asking for A-level maths, so it was an area where people could come in from a broad background, and they didn’t have to have wonderful maths, and we’d teach the maths that they needed. It, being a polytechnic, of course didn’t have so much theory as we have here [at Cambridge], so that the theory they did need to do we could teach them on the spot anyway, if they didn’t have it.

Comparing Programs at Hatfield and Cambridge

Bacon:

But yes, it was a good course, and it went modular very early, so I had that experience of designing a modular degree—which we do not do here! [laughs]

Abbate:

How does it work here?

Bacon:

It is very strange. It works the way it’s always worked, really, in that the model seems to be that a member of staff is very specialist in an area, and does research in the area, has graduate students in the area, and offers courses—this is one model, anyway, for how courses get done. And so I would teach a course in concurrent systems, another one in distributed systems; I don’t teach computer architecture now, but I did teach it; I taught the MIPS assembler, because that was an area I was very happy with and familiar with, and wanted to keep up with the new architectures. So that’s the way—one way it works. And the final year is mostly like that. And then other courses, I suppose people—probably, again, people who are doing research in programming language designs would be the ones who are teaching the programming languages, so they offer courses.

It’s a strange degree course, in that nothing, as I say—nothing is compulsory, or everything can be attended, is the way I tell all the people who are coming up for interview: that you never have to choose between A and B; you can always do everything if you want to.

Abbate:

Is there a required core that everyone has, or do the students just do whatever they want?

Bacon:

In practice, there is, because of the prerequisites that you—you really should do the fundamental courses.

The exam papers are mixed, in that you don’t take a module and do an exam on it like most places do. You have an exam paper where there may be 12 to 15 questions, and you do five in three hours. And so, the students can look at what courses are examined on which papers, and decide what they’re going to specialize in. Typically, I think, students would take the opportunity to sit in on at least some of all the courses, and they may sit in on all of many courses, but they may decide to revise, say, eight courses to do five questions, rather than try and revise the whole lot.

So it’s very flexible. Most places make you choose streams, so you can’t do certain courses. Here they can do absolutely anything they want.

Abbate:

Interesting.

Bacon:

It’s very unusual! When I get asked for transcripts for people to go and do American Ph.D.s, it’s very difficult. You just have to say, “No, nothing’s compulsory. No, they can do anything they like from this list!” [laughs]

Coming to Cambridge

Abbate:

You came here from Hatfield . . . you came to here at Cambridge . . .

Bacon:

I came here in ‘85, yes. Roger Needham was head of department, and he said I was the first woman who applied—and so they’d appointed every woman who’d ever applied here! [laughs]

Abbate:

Applied to be a lecturer?

Bacon:

Applied, yes, made an application to be a lecturer.

Abbate:

I don’t know how many there’ve been since then.

Bacon:

Not very many. There was Karen Sparck-Jones in 1988, Janet Efstathion for a few months circa 1990, Anne Copestake circa 2000 I think, and Simone Teufel circa 2001.

Abbate:

Was that your initiative to come here?

Bacon:

Yes. I got the Ph.D., and obviously Cambridge and Xerox PARC were world centers in the area I was working in. I’d also got a Research Council grant towards the end of the time I was at Hatfield, and that was very much in an area the lab here were real experts in and many people were working in.

And I . . . It was just by chance I opened the Guardian one day and saw an advertisement for a lecturer, and just applied for it, and it was just an absolute fluke, really! I don’t know why. [laughs] And then I thought I’d better apply for somewhere else, as a comparison, and just to see how things were elsewhere; so I was offered University College London and Cambridge, more or less at the same time, so I decided to come here.

Abbate:

It sounds like you were a very employable person! Is that because your field . . . Was it a very hot area at the time?

Bacon:

Probably. It was and still is. I think distributed computing was taking off. They’d gotten the Cambridge Distributed System here, which was very early. So we had the Cambridge Ring—we got linked Cambridge Rings—and for all through the ‘80s it was used as the main research environment here. Gradually, things like UNIX and Ethernet came in, but when I first came, we were at the end of the Cambridge Ring–based systems.

Abbate:

Did your husband get a job here as well?

Bacon:

No, he didn’t. He stayed at Hatfield—he was a lecturer at Hatfield, too, by that time—and the marriage broke up about four years after I came here. We sort of grew apart, through commuting back to Watford, and interest divergence.

Abbate:

I think it was his turn to follow you!

Bacon:

Yes, he could have done, yes!

Abbate:

Sorry, it’s none of my business.

Bacon:

We could have lived half-way, but in the end, we didn’t; it was too big a wrench to move the Watford house. My son was nine when I came here, so that was very difficult, really: commuting back to see him as much as I could, weekends and vacations.

Abbate:

So he was living with your husband . . .

Bacon:

He stayed with my husband, because he was at school there, and his friends were there, so it seemed the most sensible thing to do, for him to stay there. He’s living with me at the moment. He’s now 25! [laughter.] He’s just changed his job and come to work in the Cambridge area, so he moved back in at the end of January.

Abbate:

Is he also in computing?

Bacon:

He is now. He did a history degree at York for three years, and he really didn’t like it very much, and he took a conversion course in computing. The conversion courses are really valuable, and the government’s stopping them, which is a very sad thing.

Abbate:

Someone else mentioned that. So, that allows you to convert to a different degree?

Bacon:

Yes. He’d got good science and maths at A-level; he did maths, physics, and history, and he’d got As in the maths, physics, and history, so he was able to pick up those things that he’d dropped three years earlier and do the conversion course. I mean, quite often you can’t really do them without some science/maths background. So he did very well, and then got a job in computing.

Abbate:

An inspiration!

Bacon:

Yes, yes.

Abbate:

You’ve been here at Cambridge ever since, in various capacities?

Bacon:

Yes, yes, right yes, that’s right.

Procuring Research Grants

Abbate:

What are some of the main projects you’ve worked on here?

Bacon:

Research grants. First of all in an operating system called Mayflower for distributed computing, so it was OS, and distributed OS, and communication between OS, and it involved a remote procedure-call system, and . . . You can think of it as a predecessor to things like CORBA[2], the object-oriented middleware: sort of remote procedure call, remote objects, invocation. So that went on for a while.

My . . . . The person whose job I took was Andrew Herbert, and he’d sort of started a lot of the distributed computing work on the Cambridge Ring, although he did his Ph.D. on the CAP machine; so he moved out to lead the Alvey NSA project. He took remote procedure call and standardized it—he was involved in the standardization of . . . He moved heterogeneous—that is, different languages were combined in this middleware that allows communication between distributed components—whereas in the lab, he’d worked on, and I continued working on, single-language remote procedure call. We took a language called CLU from MIT, that Barbara Liskov designed, and we added concurrency, monitors, semaphores, and we added remote procedure call to it—and we had a lot of research projects based on that. So it was a very good environment for Ph.D. students to pick up and do different components in.

Abbate:

I interviewed Barbara Liskov a couple months ago.

Bacon:

Oh did you? Right! I know her quite well.

Abbate:

She’ll be glad to know!

Bacon:

I sent her a whole load of publications that were based on CLU, because she was gathering together, you know, the influence CLU had had, I suppose. I think she was quite pleased we’d done so much with it!

So that was the first thing. Then storage services: I’d done a little filing system thing at Hatfield, so we got a couple of grants on storage services, looking at architectures, anticipating multimedia storage. So, what are the different—what architecture could be quite generic and easy to use, yet still have tuned components for video and voice and things like that, as opposed to normal flat files or structured files? So we’re doing that sort of work. And we made it open and distributed, and that led on to the requirement for access control in designing . . . You’re probably hearing about peer-to-peer and things like that these days, so it’s very much that kind of thing, like, “Where do you put the components? What’s in the client? What’s in the server? Where does the communication take place to minimize how much is communicated? Are the interfaces public, or private and trusted? How do you prove you are who you say you are?”—and all of those sort of things. So it’s really system design, and how the components of systems interact.

So, that was that! What did we go on to after that? [Looks at CV] It’s all in here, I’m sure, these research grants.

Abbate:

Did you spend a lot of time writing grants?

Bacon:

Oh, yes! It takes forever to write grant proposals! I’ve spent a lot of time doing that.

Yes, so storage services: [I] got involved with Pandora multimedia workstation a little—storage service hierarchy—then, a platform for Interactive Multimedia Presentation . . . and . . . yes, multimedia; that’s right, the IMP grant. This led on to— It was imagining that you’ve got multimedia and all sorts of things stored, and you’ve also stored the structure of a presentation, and it is very much event-driven. So this is the first work that led us to looking at event-driven systems. Say you have, oh, if you’ve perhaps got an authored film of the lab, and someone can click on a face when it appears on the film, and you’ve authored the presentation management to say, “If the user clicks on this marked area of the screen after the event ‘Janet’ appears, which is Frame 234, and before the event ‘Janet’ disappears, which is Frame 968, then stop the presentation, pop up a window, and put some information about ‘Janet’.” So it’s event-driven, interactive presentation. That was background . . .

That made us become very interested in events in general. So it led on to various, what we called “active systems” grants. We started to look at . . . At that time, the evolution of middleware had gone from things like remote procedure call to CORBA, the object invocation model, the synchronous model, in that nothing happens until you invoke it, you poll it. And then, for all of these things like event-driven systems, for things like finding faults in telecoms’ networks; like active houses—make something happen when someone moves from room to a room; and we’re now going on to active cities—like the bus has a badge or tag, the bus stops have a sensor, the sensor detects the bus, the bus driver sends back the number of passengers, or whatever it is; you have a control center . . . All of these things are event-driven, and you need to know immediately something happens, not when you choose to test; and so we’ve been looking very much at that sort of middleware.

So we called it the Cambridge Event Architecture. We’ve published quite a few things on that. There was a paper in IEEE Computer, which is a very nice magazine, in the sense that they have a hundred thousand members and it goes to all of them.

Abbate:

I get it [the magazine].

Bacon:

You get it! [laughs] They’re good at copy-editing at that level. The reviewers are wonderful for Computer, really very good. It’s a magazine, but the reviewers were better than I’ve had for a lot of conferences! And so, you get this very easy to understand and well-polished publication out of it; so that worked very well. That’s what you can give to all your contacts, just as a first, you know: “This is an overview of what we do.” So, it’s easier than diving in . . . Most of the papers are on this little bit or that little bit, and you’ve got to read the whole lot to get the details.

Abbate:

I should actually look for that IEEE article.

I was looking at this list of research you did at Hatfield. Let’s see here: You had a student whose Ph.D. was on “Traffic Monitoring in an Operational Service Network”. So this is kind of more of a networking topic?

Bacon:

Yes, this is what a Ph.D. student wanted to do. I supervised three Ph.D. students at Hatfield.

Working on Standards

Abbate:

You were doing some networking research, and I see you’re on the Open Systems [committee].[3]

Bacon:

Yes.

Abbate:

That started out in the mid-seventies . . .

Bacon:

I was doing distributed systems management. Yes, ISO . . . I don’t quite remember how . . . I think I was just invited to be on the British Standard Committee for network management. Operating systems are very much management, in that they’re managing devices and managing resources, but this was much—when I got to it, I found it was rather more high-level sort of—the network management module is sort of monitoring the network, and for com[munication]s rather than for operating systems. So it’s one aspect of what I did, but not at all a central thing—but it was fun to see how standards worked and to . . . I went to one ISO meeting, in Paris, and then I—that was just before I came here.

Abbate:

What was that like?

Bacon:

It’s very much text-run—well, I found it so. I don’t know; I’m not a standards expert; I’m not one of the kind of high-flying drivers of the standards process at all! But if you want to propose something, you must put text forward, and then it’s discussed, and then it may become a proposal or a draft comment, and then it goes through various phases, and if it’s—if all the members agree that the text is okay—then it becomes a Draft Standard, and then eventually it becomes a Standard. It is a very, very long process, and at the stage that . . . I was looking at new structures for distributed systems, and it wasn’t the way they were thinking at all, because they were still in the mainframe mode, I think, to be fair; so I couldn’t really make very much contribution, and when I came here, Cambridge didn’t want me to be involved in standards anyway.

Abbate:

Really?

Bacon:

No, I got a little involved in it, but it’s not very research-oriented. It’s good for them to have a bit of research input, but they mostly want commercial input.

The other thing with standards is if you have different factions on the committees, the standard tends to be the union of what they all want, and then anything can conform to the standard, and it’s not very helpful—so you have to guard against that happening. In a research environment you can really do much more what it should be like. Ideally, if you were researching this new area, this is how you would go about it, how you think about it.

On Commercial Applications of Research

Abbate:

Have any of your research projects been commercialized?

Bacon:

No. . . I think some could be; some could have been. I think what’s happened is that the Ph.D. students who’ve worked in the group have gone out. One went and wrote OmniORB, which is free software, and went to AT&T Research, and . . . It’s a CORBA.

Abbate:

It’s called Orbitis?

Bacon:

It’s called “OmniORB.”

Abbate:

OmniORB—I hadn’t heard of that.

Bacon:

It’s just a free version of CORBA, and it is C and C++ -based. There are various free CORBAs you can get from all over.

So that’s one I was sort of close to.

Abbate:

Sort of second-generation distributed systems . . .

Bacon:

Yes, that’s right.

I’ve never . . . I think I’m just a bit too old to be in the startup generation. Ten years later I may have thought, “Oh, I should commercialize that,” but I haven’t.

Abbate:

You haven’t gotten funding directly from industry?

Bacon:

We have had input, but not major funding. I’ve had grants from ICL, and from HP Palo Alto, and from Nortel—for grants. So, you know, it’s been enough to support a research assistant, or something like that. So yes, they’ve put money in.

Quite a few are funding Ph.D. students at the moment. Agilent are, and DERA (Qinetiq); ICL. So that’s quite a lot of overhead, in that you are invited to give seminars at all these places, and go and visit there, and talk to them; traveling.

Abbate:

To industry.

Bacon:

Yes.

Abbate:

Do you get anything out of that, in terms of learning things, or . . .?

Bacon:

The Ph.D. student has got the funding, and I go and talk to the funder. But I don’t get anything; no, it’s voluntary.

Abbate:

I mean they learn from you, and do you sort of pick up interesting ideas?

Bacon:

Yes: exchange of ideas. Yes, yes, you find out things. Also, if you’re doing something like generic middleware, it’s essential to have application areas, and to . . . For example, the one from Agilent, we really need some notion of volume of data, and bandwidth. If we’re doing event monitoring of faults in networks, or just monitoring networks, we need to know, you know, at what rate the events are coming at, and how you filter them. There are very practical issues that you have to look at, rather than just, “This is a good idea to have event-driven middleware so that you can respond instantaneously.” One angle of it is the large volume of events. When we move to very large numbers of devices scattered around the world and carried by mobile people, in articles of clothing as well as phones and laptops, the volume of events can get colossal. So the filtering, and management, and all of the distributed systems issues are very much up for grabs again, with mobility becoming so prominent. Naming, and location, and access control. So those are the sorts of areas we’ve moved on to at the moment.

Abbate:

So, do you have sort of pilot experimental projects that are supposed to prototype . . .?

Bacon:

Yes.

[START TAPE 1, SIDE 2]

Abbate:

I was just asking about applications . . .

Bacon:

About the applications, and about the software that’s developed here. Well, all Ph.D. students in systems have to build experiments, build prototypes, so that always happens. We’ve had an awful lot of Ph.D. students, so a lot of software gets built. Not all of this is in a state where it’s like a product, because they come for three years, four years, and then they go away. Sometimes they build on each other’s work, but you can’t start a Ph.D. off saying “Oh, he will do that in two years’ time, and then you will do that afterwards,” because it just wouldn’t work. We’ve had a lot of software, and—I mean the product problem is that: I was using DEC Alphas for a long time, and the operating systems—we were always pushing the edges of the services they give, rather than being right in the middle. We tried to do multimedia work where the streams followed the people as they—as their badges indicated they moved around Tower One in the computer lab, for example; but new releases of the operating systems wouldn’t keep up with this, or at least what our support were able to give us, the computer officers maintaining the systems. So, things kind of rust if they’re in a university environment, rather than if they’re in a product environment, [where] there are huge numbers of people keeping them up to date with new releases of the operating systems they run on, and things like this. So it’s a very different area, if you’re doing prototype, just proof-of-concept software. We have put—what did we put on the Web?—we’ve got the Event Architecture, and it can be applied to message-oriented middleware or CORBA—event- and object-oriented middleware. What we did was get OmniORB from—the one that the graduate student wrote when he went to what’s now AT&T Research—and we’ve added events to it, and that’s now free on the web for anyone to take, under the same conditions that OmniORB is released. So we can do the sort of free software route, so people can take it and use it and play with it as they wish.

Bacon:

Applications: We haven’t talked about the role-based access control thread. The event stuff is one major thread in what this group does; the other is role-based access control. It’s in a very practical environment, and it is very much implementation-driven. There is a lot of work on role-based access control that’s very theoretical, and they don’t . . .

Abbate:

I’m not sure what role-based access control is.

Bacon:

Well, it’s a nice scaleable way of saying who can do what in a computer system, in that, if you have to say all the user I.D.s who can access this object, this file, then it becomes a very long list. If you wanted to build, for example, a national health record service for the U.K., it is much better to design it in terms of, “Doctors can do this; nurses can do this . . . ” And so, in a hospital, you might have an accident and emergency team, and they would have roles; and when they come on shift, they enter the role “Doctor” or “Nurse” or “Receptionist” or whatever it is. So you don’t have to keep track of who’s employed as receptionist today or whatever, and so they’d have to give their credentials to access, to enter the role—and then they can do things: get medical records of the patients they’re treating, or something like that.

So that’s another main thread of the stuff we’ve been doing: the role-based access control. So we have a grant to look at applications of that. We might have gone for looking at digital libraries with somebody called Henry Gladney from IBM Almaden, because he saw a paper we published. He was on the program committee, came and talked about it, got very keen on it, and said he wanted—he was going to do digital libraries, and he wanted to use this scheme; but it fell through because his software had a flaw of some sort—I don’t know what—so it didn’t happen. So that was one possible application: role-based access control for digital libraries, so a student, or librarian, administrator, professor, whatever it is—the roles would indicate what you could read.

Okay, so instead of that we went into health care. This also arose through contacts—just happened to make contacts, at college or whatever, just whom you meet. So we started to look at a system for electronic health records, and had a workshop which was half computer scientists and half medics, about how you might go about building such a system; we put a grant proposal in a year ago, and it’s sort of gradually creeping on. They didn’t fund it directly; they said it was too broad, because we were trying to look at a national system top-down, whereas their strategy had been rather local, because the Thatcherite model was to have everyone competing against everyone else and making decisions locally as to what they bought, and implement them locally. And so, if you want, say, a national system, you’ve somehow got to integrate all of these local components. And we were trying to look at what sort of naming, what sort access control, what sort of indexing and location—how do you do it—in a very broad, national system. So we’ve worked on that; that’s a major application area. So it is electronic records, rather than bioinformatics—but we would need it to interact with medics to do that. So I’ll have to keep you posted on how that goes!

Then there’s supposed to be going to be another workshop in July, I think?—June, July, I’m not sure. But the record of the workshop we held is on the Web, I think it was September 1999, to get all of this area going.

On the International Distributed Systems Community: Editing IEEE Magazines

Abbate:

I guess you’ve done some . . . you visited MIT, and . . .

Bacon:

Oh yes. That’s right. That was a good term, yes.

Abbate:

What sorts of things were you doing?

Bacon:

We went to work with David Gifford and his group, and talked to graduate students. It was really just a term there. We gave seminars, and interacted with graduate students, and helped people to write up—I’m still writing references for some of them—and just generally interacting. It was just fun to talk to people, and to go their seminars, and to give them seminars. It was very, very good.

Abbate:

You must have a lot of contact, through conferences and things, with scholars everywhere . . .

Bacon:

Yes.

Abbate:

So you’re very tied in to the . . .

Bacon:

Yes, a sort of international distributed system community: yes. And I’ve been editing IEEE Concurrency for two years, and then Distributed Systems Online —which is their first attempt at a Web-based magazine, a community project.

Abbate:

The IEEE’s?

Bacon:

The IEEE’s, yes. So if you go to the top page of the Computer Society’s [web site], you can see it there; then you get to the top page of Distributed Systems Online. And it’s trying to be a sort of information service, with a lot of subareas with experts—we have about, I suppose, 35 volunteers from the different areas in distributed systems—and they are doing a sort of value-adding of through what you might call a portal: lots of references; links; talking about conferences; educational things; tutorials; journal papers—just any information you choose to put up, “you” being a subarea editor on your area.

Abbate:

Interesting . . .

Bacon:

Yes, I’d say it’s very interesting to do this. And they’ve learned a lot, I think, and all the people I’ve started working with seem to have been promoted very quickly, because of this work they’ve done.

Abbate:

So it’s kind of a distributed system itself.

Bacon:

Yes. So yes, I work very closely with Los Alamitos [IEEE Computer Society]—from five o’clock onwards, and . . .

[Displays web site on her computer.] Okay, so this is the current issue [volume 2, number 3].

Abbate:

We’re looking at this web page . . .

Do you have to be a member of the [society to access the web page]?

Bacon:

No, it’s free. It’s free to everybody. They should advertise it more. But they’re rather—They seem reluctant to use mailing lists—email lists—to push it to people. They just sit there with it advertised on the page, and hope you find it.

[Going through web page.] So, all of these areas are just selected . . . We’re in distributed systems, so this is real time and embedded, conferences, journals, books, technical committees, contacting the editors . . .

Abbate:

That looks great!

Bacon:

Yes! It’s been interesting to do. We publish, at the moment, just one paper per issue, so that’s “Strategies for CORBA Middleware-Based Load Balancing.” Then there are news items, and the other thing we do is advertise selected things from IEEE publications, so it’s a little taste of things that people can subscribe to, but they get it free when they read it here.

Abbate:

Was this your idea?

Bacon:

Yes. They asked me to be editor-in chief of Concurrency, and when I took it over, they were at the point of deciding that the subscription numbers were so low that they weren’t going to run it any more—which was pretty silly, really, because they should have thought about this six months earlier; then they needn’t have recruited me! So, they asked me just to finish it for two years, and then in the meantime, what could I do online? Because I’d started to interact with the staff and editorial groups, and was getting on very well with them, and IEEE-CS asked me if I’d propose something like an online journal—so I did that. So I’ve been going to the Publications Board meetings three times a year, and I’ve kept on proposing things, and they thought this was a good idea, and set it up as what they called a “communities project”—a part of their Communities Projects—so it’s serving the distributed systems community: That’s the theory of it. They funded it for six months or something, so they need a funding model for it to continue.

Abbate:

That’s a good question.

Bacon:

The thing that we will move to is that it will be a sort of community support for Internet Computing and another magazine they’re going to start on pervasive computing. Internet Computing is just a paper magazine with a digital version with nothing extra, and this is going to be the extra stuff. So part of the funding for Internet Computing and the new magazine will be diverted to keep us going, so I don’t have to keep asking for money for another six months to see how it goes!

Abbate:

So they did discontinue the paper version of the Concurrency magazine.

Bacon:

They did, yes: Yes.

Abbate:

And substituted this web site, which is now . . .

Bacon:

That’s right. The title Concurrency didn’t seem to mean anything to people.

Abbate:

I’m not sure what it means either.

Bacon:

No, and when I looked at it, it started off as Parallel and Distributed Technology, and then they changed the name to Concurrency, which seems to have been a bad move; but it still had a subtitle, about “distributed, mobile, and parallel systems.” So my editorial policy was to do all of these mobile computing issues, and I wanted a work flow issue, and so on, all the way through. But it wasn’t in time to save it, really, because they decided to stop it, almost at the time I took over. But I did the eight issues, and that was fun. I’ve learned a lot from doing it—wrote a lot of editorials . . .

On Computer Science Education

Abbate:

Have you done other types of professional service activities?

Bacon:

For the British Computer Society, I’m on what’s called the Academic Accreditation Panel. So I visit institutions—I’m supposed to do three a year—to accredit their courses as exempt from having to do the British Computer Society examinations, and be a Chartered Engineer. Does this ring bells?

Abbate:

Some of it does. I wasn’t sure about . . . So the accreditation—if your school is not accredited, you need to take . . .

Bacon:

Yes, if you want to stay a professional member, a member of the British Computer Society or member of the Institute of Electrical Engineers, if your course is not accredited, then you have to do their exams.

Abbate:

I see.

Bacon:

I don’t think you have those in the ACM or the IEEE.

Abbate:

No. Okay, so I realized you could take the exams, but I thought that was for people who hadn’t taken a computer science course, but wanted to be qualified.

Bacon:

Yes. Well, it is used for that as well. You . . . you could just study, I suppose, if somebody would teach you, for the BCS exams. But it is also [for] if you’ve gone to somewhere whose course is not accredited. If you want to be a professional member. But, I mean, in this day and age, when there’s such a demand for computing people, it’s not obvious that the professional qualification is demanded by employers—so I’m not sure what proportion of people in industry are professionally qualified in that sense.

Abbate:

In terms of being BCS members?

Bacon:

Yes.

Abbate:

Now, being a Chartered Engineer: I’ve met some people who were, but I’m not sure what . . . Is that just honorary? Or does it allow you to do things you couldn’t do otherwise?

Bacon:

I think if you’re in a “real” engineering discipline where you’re building bridges, like civil engineering, then you would need it to practice. In practice, in computing, you don’t need it to practice, so it doesn’t really allow me to do anything I couldn’t do otherwise as a Lecturer—Reader— in Cambridge. But in theory, it is saying you are professionally qualified.

Abbate:

But how do you get that? Is that . . . take an exam? Or is it . . .

Bacon:

How did I get it? I—there is the educational part, which is your degree, and then you put your professional profile in: say what you’ve done over the years, what jobs you’ve done, what educational things you’ve done—and so you apply for it. I became a fellow of the British Computer Society in the same way—said all of these things that I’d done, and got that. I think you can do that for the ACM and the IEEE, as well. You can apply for senior membership . . .

Abbate:

They have Fellows . . .

Bacon:

Fellows, and senior members. I’ve just never gone into that.

Abbate:

It’s honorary.

Bacon:

Yes.

Abbate:

I don’t think it qualifies you to do anything you couldn’t do otherwise.

Bacon:

No! It’s just something to go up on your web page, or on your wall or something!

Abbate:

I know for some professions it’s much more of a necessity for practice.

Bacon:

Yes, that’s true.

But I think it’s not even essential to have a computer science degree to get a programming job. A maths degree, physics, or almost anything , and they’ll train them in place. So it is very worrying, really, that there are all those people out there who haven’t ever really been taught the discipline correctly, and what there is to know! Because even Lecturers in the subject are always running to keep up with the way it’s evolving—so if they don’t have a starting point . . .

I mean, my son’s going to do a graphical information system database, and it is colossally data-intensive. And you know, his course would stop at traditional databases: transaction processing, records of parts of airplanes, or whatever—traditional stuff!—and if you get these massive maps, it’s a completely new area. And how people who haven’t even got a computer science background, who just don’t know anything until they’re taught—I don’t know how they cope, really, or how their companies have the resources to teach them—and who does keep up!

Abbate:

I don’t know if they hire them at that level—they’re probably doing something more basic, without the degree.

Bacon:

Yes, maybe. Maybe the higher people . . . It’s just the issue of people keeping up with developments in the subject, which I think the media and politicians have no concept of. They think you get your degree and you know the discipline!

Abbate:

Do you have an agenda for computer science education, or a way that you think it should be going? If you were “king of computer science”?

Bacon:

Yes, I know what you mean. Whether . . . I don’t know—whether I wish it or not, it will broaden. The national bodies, the quality assurance agencies, are pushing everything in that way, and the British Computer Society have followed, in that there used to be a compulsory core that everybody with a computer science degree that’s accredited must have done: some hardware, some networks—just a little, so that you’ve got the basics; some maths, some theoretical underpinnings, and so on. But the movement is rather away from that single intensive core, towards saying, No, you may do this whole area that is, say, artificial intelligence, and for this maybe you didn’t come in with a lot of maths, or maybe you did—I don’t know, human-computer interfaces with psychology—you know, a quite, quite different degree that’s still under the computer science umbrella. And also, you’re seeing the trend away from what you might call “hard” systems, in that it is very difficult to recruit staff in that area, because they’re so much in demand in industry—you know, anyone who can actually build systems.

Abbate:

You mean hardware? Or operating systems?

Bacon:

The operating systems. The real hands-on, make-things-work, building systems. It is . . . It’s difficult. Because there are so many opportunities, and the young lecturers are paid very little—here in the U.K., anyway, particularly in Cambridge. So they naturally—they’ve got brilliant Ph.D.s and would naturally perhaps do startups, and have half of their time or half of their energy diverted towards their company. And this is a new model . . .

On Computer Science at Cambridge

Bacon:

So, yes, education: An issue in Cambridge is whether we continue to have college supervisions. Do you know about this system in Cambridge? The collegiate university?

Abbate:

Why don’t you explain. I know some of it, but . . .

Bacon:

Oxford and Cambridge are different from other universities in that, as—I should say just Cambridge if I start to say how we are, because Oxford is a little different. In Cambridge, you’re appointed as a Lecturer in the department, or whatever it is—I’m a Reader—and you give your lectures and you do your research, and you supervise your graduate students; and your duty is that, for the department. There are also colleges—and you need not, in Cambridge, have anything to do with a college if you don’t want to. It is the colleges that admit students, so that if you are what’s called a “director of studies” in computer science in a college, then you interview the students, decide who you’re going to take, make the offers and then they come—and then they all converge and do the computer science part of the course in the computing laboratory. And so, our numbers are very much controlled by who the colleges will admit, and also who the colleges can persuade to do college jobs in computing. So that’s the recruitment. Directing studies involves looking after a given student throughout their three years here.

Abbate:

So there are tutors?

Bacon:

Sort of, yes, like tutors. And so you said, “How do they choose which courses to do?” Well, they have to talk to the Director of Studies twice a term at least, and so you would advise them which are compulsory, which they might choose between, depending on their interests and their strengths—and so you look at their program of study. Now, we don’t have so many tutorial classes as other universities, because we have supervisions, and Cambridge students are supervised for all the courses, and they’re typically supervised in pairs, so this is very labor-intensive. So, if I lecture on distributed systems to 80 students, then all of those students will be supervised in it, so we’re looking at finding supervisors for forty pairs of students.

Abbate:

And the supervisors . . .

Bacon:

The supervisors set work, and mark it, and talk to the students about it. And so, in theory it’s bringing on the very best intellectually, by having them supervised by very bright people and letting them explore the direction that they want to go in, and all the issues leading from the course. So it’s not just clarifying what the lecturer said; it’s anything you might wish to talk about on that subject. And this is the Cambridge ideal, and it’s brought forward a lot of very great scientists over the years.

So, we have to find a lot of supervisors in computer science—and again, the payment issue comes to the fore, because we pay the same amount for computer science supervisors as for any other subject, and the people who might supervise can earn a lot more doing other things. So a graduate student—I mean, they’re very good when they do supervise, because they feel it’s putting something back into the system, often, that produced them—but if they wanted to earn money, they’d be better to go write a piece of software for a month or so, to take time out, get some money, and then come back.

So that’s the way the Cambridge system works: The Lecturers lecture and do research, and the department and the colleges are responsible for directing studies, admitting students, and arranging supervisions. And the colleges are finding it increasingly difficult to find anybody who will do this work, because it is sort of virtually unpaid. It’s supposed to be for the honor of being a Fellow of a college, and you’re paid a small amount [laughs]—but really very, very little.

So over the next decade it will be interesting to see how Cambridge evolves, really. And, I mean one hopes that the management—the Vice Chancellor—are aware of these issues in this subject. What’s happened recently is that the government has stopped giving the supervision payment money—the student support money—to the colleges directly; they’re giving it all centrally to the university, and they’re going to gradually decrease it, and the university have to pay the colleges—so it’s a way of controlling it, and managing it, and monitoring it, and seeing exactly what’s happening. So it’s interesting times, in that the country—the local government and the country—are very wary of Oxford and Cambridge costing more. They say it’s elite . . .

Abbate:

Do they get more money than other English universities, per capita?

Bacon:

Well they have to, because of the supervision payments, basically. There is a college fee, as well as the tuition fee—and so this is controversial: Why should these two places get more money?

Abbate:

Why should they? Wouldn’t all the others have supervisions if they could?

Bacon:

They probably would if they could. Yes. It’s just a tradition. Yes.

Abbate:

Well, that must be difficult.

Bacon:

Yes, it is. It’s not obvious; there are arguments on both sides. And the colleges, they’re full of, not just university Lecturers, but people who have been appointed by the college to do scholarly research—so Trinity College, it is said, has more Nobel Prizes than the whole of France—there are these sort of things that are quoted! And they’re really centers of scholarship, and across disciplines, so a given college should have Fellows in all the areas that the university teaches.

Abbate:

Is it hard to get computer science Fellows?

Bacon:

It is extremely hard. And the colleges don’t often—don’t always perceive this; they don’t perceive what it’s like for technology subjects—so you really have to try and tell them.

Abbate:

That’s very interesting. So there’s quite a sort of decentralized . . .

Bacon:

It’s a very strange model.

Abbate:

And you, as a faculty member trying to have a coherent program, need to juggle all of this . . .

Working with Women Computer Science Students

Bacon:

It’s a nightmare, really! I just work all the hours there are—because I feel I ought to look after the women’s Colleges, which don’t have many students. So I’ve directed studies for Newnham and New Hall and Lucy Cavendish for years and years and years, and I’ve passed Newnham on to someone else but still arrange the supervisions, and . . .

Abbate:

And this is just for the computer science students?

Bacon:

Yes, just for the computer science students. You know, there are dribs and drabs across all the different colleges, so Lucy Cavendish may have had a student a year, because it’s a mature students’ college—a mature women’s college; Newnham may have sort of five students at a time across all years; and New Hall have got nine at the moment—and I’ve looked after New Hall for all the time I’ve been here, apart from when I’ve been on sabbatical. And I became a Fellow of Jesus four years ago, and I’ve got twenty students there across all classes. So it is a lot of work, because some of these people drop in and see you about their supervisions, and . . .

Abbate:

So you’re supervising all of those students?

Bacon:

No, I have to arrange . . . I arrange supervisions. The colleges don’t really understand that a given Lecturer cannot supervise the whole of computer science—particularly as it’s a subject that changes rapidly! But you have to keep trying to tell them, explain as best you can, and gradually it sinks in.

Abbate:

Do you try to encourage the women students to take computer science? Or is it too late by the time they’re here?

Bacon:

Yes, I have—we have tried to do this. As [Karen Sparck-Jones] may have said, we’ve had conferences of woman teachers, and—to hope that they’d go back to the schools and encourage their girls to apply.

There are a lot of access initiatives, not just for women but for working-class areas, for state schools, to make sure that we don’t get the balance weighted towards the private schools, the independent schools.

The Downside of Being an Academic Computer Scientist

Abbate:

Have you ever encountered any gender discrimination in any of your jobs?

Bacon:

It’s never been explicit. I think most of the people I’ve worked with would perceive themselves as trying to be equal, you know, trying to give equal weight. I think there’s implicit gender and class bias everywhere. And people just aren’t aware that they have it, and you can’t really do much about it.

Abbate:

But things in terms of pay or promotions . . .

Bacon:

Promotion: Cambridge is just ghastly for career structure; it’s got better recently, but it is just so bad that gender didn’t really come into it; it was just appallingly bad for everybody.

Abbate:

What’s so bad about it?

Bacon:

It was bad in that we had our first established Professor about—let’s see, four . . . it must be five or six years ago—so before that we had no established chair in computer science.

Abbate:

And that’s the university who chooses where those [established chairs] go?

Bacon:

Yes. The chairs we did have were personal chairs, so that if the person retired or moved away, then it reverted to the Lectureship that they came in with. So we’ve had a number of Professors, but not established chairs; so that’s gradually just starting to happen.

So when I came to Cambridge in 1985, I really came for the honor, because it was a good research place, and I took something like a thousand pounds a year cut in salary to come from Hatfield polytechnic to Cambridge University, to the top of the Lecturer’s scale! At that time [there were no Senior Lectureships]—Senior Lectureships have only just started last year, and Readerships and chairs were extremely hard to get. They’re trying to liberalize gradually, but it’s just been extremely difficult. And I think in this area, they don’t understand the extra pressures on technologists—there’s grant applications, and interacting with industry, and all of these things add to the burdens; it’s not like simply having to do computing, build your experiments—and it’s not like sitting and thinking and writing, and probably doing research with documents.

Abbate:

Why don’t they have more professorships? Is it a financial thing?

Bacon:

It’s a financial thing. It’s also a strange university in that it’s very democratic, and everything has to be balloted. So if you want to make changes: Other universities would say, “Computer science is an up-and-coming area; we must have three times more students and we must have three times more staff. We’ll do it now!” And so they make the department bigger, they make more senior posts, and they recruit more students. Cambridge can’t do that, because the colleges have to admit the students, and they can only admit those they can teach; so they have to have a Fellow who’s director of studies. And there’s pressure within the colleges, because they may have five historians and four classicists or something, and those people want to keep the numbers up in their subjects, and unless you decrease some areas, you can’t increase others, because of the accommodation and government quotas—and so it is difficult to be a new subject in a very rapidly expanding area. And, because the university is democratic, you can’t enforce changes by management decisions—not quickly, anyway.

Abbate:

That’s interesting.

Bacon:

That’s the way it works!

Abbate:

Well, I’m impressed you have as many people as you do!

Bacon:

You know, Jesus College—which I went to partly because they’ve just networked the whole college; every student has a port in their room for their own computer—they showed interest in increasing the student numbers, and so I’ve increased from two a year to six-to-eight a year since I’ve been there, so that’s fine. But they don’t really see the need for more computer scientists. It’s also so new that you don’t have all the people who’ve gone through and are now old and emeritus professors, and they’re . . So that if you look at the university prospectus for Jesus College, college entry, it has me for computer science, and that’s it! You know, whereas for some other subjects, there is the director of studies, say, for a small number of students like classics, but there are a senior tutor, and a couple of professors, and a couple of emeritus professors are still around, making a community in classics—and this has not happened yet [in computer science], because the subject’s new—so it’s difficult.

Abbate:

The people might not stay around.

Bacon:

Oh yes. There’s so much demand for good people in computing, it’s difficult to keep them. Particularly for no pay! [laughter.] Housing in Cambridge is now—it’s an area where it is just appallingly expensive, and the young university Lecturers can’t afford houses, because of all the startups. We’re in a sort of Silicon Fen. Yes

Abbate:

Like Stanford.

Bacon:

Yes. Well, it’s like the whole of Silicon Valley, I think. Yes. I have an ex-Ph.D. student who’s telecommuting. She works now for Cisco, [but] she’s still living where she lived near New York—when she worked for Telcordia, that was Bellcore—but she’s working remotely for Cisco, rather than trying to move west, because she knows she couldn’t afford to!

Abbate:

“Don’t do it!”

Bacon:

“Don’t do it!” That’s right! [laughs.]

Abbate:

The airplane tickets will be cheaper.

Bacon:

That’s right.

Balancing Work and Family

Abbate:

How have you managed to balance work and family?

Bacon:

I came here when Thomas was nine, and it was really hard till he went to university; and it’s really easy now in that he’s 25, so . . .

Abbate:

He didn’t live with you until . . . ?

Bacon:

No, he didn’t live with me until he went to university, and then he’d come back for vacations, as much to me as to his father. So I’d been commuting there weekends, and fetching Thomas at weekends when we split up.

Abbate:

He was back at Hatfield, you were at home.

Bacon:

Yes, one reason I, you know, moved into education was to be able to . . . I just had the one child, in fact, but that was the idea, really, to have something I could balance with a family, because I knew I wouldn’t want to stop, and I couldn’t see how else to do it, really! And Hatfield were just starting a nursery at the very time Thomas was born, and it had something like—the baby room had about five babies, and then the toddlers’ room, up to school age, had about fifteen toddlers, I suppose. So Tom . . . It was very nice, in that I could pop up there when I didn’t have lectures, because it really close by, and he was there from baby to school.

Abbate:

That was very forward-looking.

Bacon:

That was really good. People who were a little ahead of me had put their children with childminders for very long hours, and I didn’t like that so much, because they’d be on their own with one person! In the nursery they were with friends, and I was very close by—so was his Dad—you know, both of us could pop up.

Abbate:

Was that unusual at the time?

Bacon:

Yes. When was this? ‘76, Thomas was born. Yes, so ‘76 to ‘80. Yes, it was only just starting, and it’s unusual. The university here doesn’t have very many facilities. There is a university nursery that people can pay to take their children to, but I don’t think it’s very big. I don’t really know much about it, because I moved, and got too old to be interested. [laughs.]

On Mentoring

Abbate:

Did you have any role models or mentors when you were starting out?

Bacon:

No women.

I got to know James Wilkinson at NPL very well, and he was a sort of academic father figure, because I didn’t have any academic influence at all in my family. So he was definitely one. And he worked, with Alan Turing, on the pilot ACE—so he was one of that team. There were a lot of people still over from that time, when I went to NPL. Yes, so you know, the fact he’d been at Cambridge, and did extremely well in the Maths Tripos, then went and did this—that was always a role model, I think.

I suppose all the people in Cambridge, really: I’d known about their work, and it was an honor to be part of all of that! So you put up with the appalling career structure, just to be part of it! The graduate students are wonderful. They’re really, really good, and so that’s another thing that keeps you here. We’ve got a team, my colleague and I who lead a research group, with something like sixteen Ph.D. students in a group, and it really is very good.

The Upside of Being an Academic Computer Scientist

Abbate:

What have you found to be the most satisfying aspects of your computer career?

Bacon:

I suppose it’s working with graduate students, and the research, getting the research moving. There’s so much to do that you can’t get it moving very fast—there are so many college things, teaching things. So, yes: working as part of the research group here has been very satisfying.

But I wouldn’t admit it! [laughs.] Wouldn’t admit it to anyone who’s not paying me! [laughter.]

Abbate:

It’s interesting that computing has this sort of stereotype that it’s very isolating, and you’re alone with the machine, and yet it seems much more—I mean, that wasn’t your experience.

Bacon:

Not my experience at all: absolutely not. Yes, they have to go away and write their big pieces of software, which take a long time, but the enjoyable part is the brain-to-brain sitting in a room, working things out and, you know, “How do we think about this?” That’s the essence of it, really. And it’s one thing you can keep passing on, to some extent, [and] hope it’s useful for them, in that all the technical stuff becomes obsolete very quickly, but the sort of experience of system architecture, building systems, all these sort of things, stay relevant. So we’re now moving to mobile systems, and it’s entirely new, but at least you’ve thought about big systems, or whatever, before, so you’re somewhere into it. So you can give a start, and then your brains can add a lot of details, add a lot of suggestions, and change things, and you can monitor the changes, and it’s very good. So it is a sort of community of scholars. That’s the enjoyable thing.

Abbate:

It sounds like they also learn something about teamwork in terms of . . .

Bacon:

Yes.

Abbate:

. . . developing a project.

Bacon:

Yes, yes. I think they do. They’re having regular meetings on events, on the “active city,” so that’s a fun thing: you talk a lot about how you might build an active city.

Abbate:

Is that going to be prototyped anywhere?

Bacon:

I think it would be too huge.

Abbate:

You’re not going to set up the buses, or something?

Bacon:

Well, you know, we could do [that]. We’re waiting to see what— They’re going to radio-network Cambridge, Vodaphone; they’re putting a network in in due course. So I think when that comes online, we’ll see what kind of mobile technology plus network technology will be around, and at that stage you might look at how you might have sensors. So yes, we’d certainly like to do that.

ICL, they’re doing similar things, and they’ve got some gadgets that—but they’re very crude, not very interesting at the moment, but . . . for the moment, what are we doing? They’re getting simulators of these palmtops, and playing with them, to see what they can do with the current technology—but it will get much better so quickly that you almost have to wait for it.

On Changes in Computing: The Growing Expectations of the Public

Abbate:

How would you say computing has changed since you started?

Bacon:

Just in every way, hasn’t it? [laugh.] In every single way you can think of! The scale—I mean, it’s technology-driven—and the scale has just been amazing.

Abbate:

In terms of processing power, or . . .

Bacon:

Yes. Growth.

Abbate:

Do you think the culture of computing has changed?

Bacon:

It’s has to have changed, because so many more people are involved in it.

The culture of patenting and making startups has obviously come recently. And the new Vice Chancellor, Alec Broers, is encouraging this, I think because he feels it’s very good for academics to put their stuff out there in the world. So that’s a change within the university.

Obviously, the Web’s made a huge difference, as an interface.

I think it’s the ubiquitous computing that has changed a lot, and over the years it . . . First of all it was a specialism for only the academics in that area; and then maybe the other academics had computers and began to feel they knew what computer science was! [laughs] And so, if you go to college at the moment, you still have people asking you about Microsoft Word or some such thing, and I don’t use it, so they think, “You can’t be a computer scientist! You don’t know what Microsoft Word does!” So there’s this mis—because they used to be sort of perceived as specialists doing their own thing, and then everyone was using computers, so everyone felt they knew what it was all about. So that became very different, culturally. And I think the expectation, the social expectation that there will be huge systems—counter to all the evidence of the massive failures that there have been—you know, such as the U.S. online tax system. I think they’ve had billions of pounds on different projects that had to be stopped because they were not getting anywhere. I think you’re on the third system now.

Abbate:

For . . . ?

Bacon:

Online tax: filling in the tax forms and . . .

Abbate:

Really? I didn’t actually know that. I’ve done it online for the first time this year.

Bacon:

Okay. All right.

Abbate:

It seemed to work.

Bacon:

Yes.

Abbate:

I guess I missed the beta versions!

Bacon:

At one of the conferences I went to, Jerry Saltzer from MIT gave all of these projects that failed: when they were abandoned, how much had been spent—and there were just enormous amounts of money . . .

Abbate:

I know that’s certainly true in general.

Bacon:

. . . on nonworking systems.

In the U.K. we’ve had terrible problems with just the passport system, which is a centralized database—it’s not distributed; and driving license system, centralized again, in Swansea; social services, and immigration: all these things, all single systems in one place, and they can’t get it right! And it’s behind schedule, and it costs millions more than it’s supposed to. And yet, the public perception is that, [because] we all use computers; politicians, people in general would think, “Yes! Of course you can transmit electronic health records around the country. Of course you can. You know, build it!” And, so there’s a big mismatch in what can be done, and what people perceive should be done.

Abbate:

Do you think that’s more than it used to be, or do you think that’s always been?

Bacon:

No, it is much more, because of the universal use of computers—and particularly the universal use of the Web, which they call the Internet—and the fact that everything is networked: they suddenly see this and think, “Oh! It should be easy!” And there is an awful lot of hard work. And you know, I just feel very worried about anyone attempting—you know, perhaps commissioning—a large system, and the knowledge in government—because they’re rather old and senior people—and is there enough knowledge there to drive this? And are they getting the right advice?

I went to an EPSRC meeting, and it was starting off a distributed information management program, and there was a government person there; he was something like the cabinet advisor on I.T. And he stood up, and he was a very good speaker, and he said, “Don’t ask me any technical questions; I’ve got a Ph.D. in” sociology or politics and economics or something like that; so he was a brilliant Oxford Ph.D., but not in computing—so he said, “I can’t answer any technical questions.” So then you think, “This is the person advising the Cabinet, and it’s social policy rather than real technical knowledge of what can be done.” And that’s a very isolated example, but you just hope it’s better than that.

Abbate:

Hmm. So you think the technology’s gone ahead of where the policy . . .

Bacon:

Technology has gone ahead, and because computers are very widely used, there is expectation, social expectation as well. So the policy makers, and the people who are affected—the population as a whole—both expect—and neither group knows—they both expect there will be large distributed systems. And, as a person who has worked in distributed systems for a long time, I know it’s hard, and I don’t know if there’s a perception of how hard it is, where it matters. So, it will be interesting to see . . .

Abbate:

Yes . . .

On the Status of Women in Computing

Abbate:

Do you think the field has gotten more open to women since you started? Or about the same, or less?

Bacon:

We would like to have more women, but they don’t apply! I thought it was a problem with the schools, and that when there was only one or two computers in the schools, it seemed that the boys would monopolize them; but as computers have become cheaper, you’d think that that wouldn’t apply anymore, so it’s something else. I think girls are not perceiving it is for them, in some way? I mean, I can see they’d think that games are mindless, but surely using the Web, and getting your information from a computer: surely that would be interesting. So you’d hope that a broader use of information from computers would encourage more women to come in, but we’ve not really seen that here.

Abbate:

Do you have any advice for young women contemplating a career in computers?

Bacon:

I should say, do it! Because there are wonderful opportunities, and it is well-paid, and if you want to be independent, and make your own way, then it is a very good way of doing it. Yes. Go for it!

Abbate:

I think I’ve probably covered my questions. I don’t know if I’ve missed anything important.

Bacon:

I can’t think of anything.

Notes

1. Bletchley Park was a code-breaking center during the Second World War; the Post Office supplied equipment for its code-breaking computers.

2. Common Object Request Broker Architecture.

3. Part of the Open Systems Interconnection effort organized by ISO, the international standards body.