Oral-History:Fran Berman

From ETHW

About Fran Berman

Dr. Francine Berman was born in 1951 in Glendale, California. She attended the University of California, Los Angeles for her undergraduate degree, and later the University of Washington for her graduate and postgraduate degrees in Mathematics. Berman’s Ph.D. thesis focused on nonstandard models of propositional dynamic logic, and her work has been primarily focused on theoretical computer science. She has worked as a Computer Science Professor at the University of California, San Diego and Director of the San Diego Supercomputer Center. Berman was also the inaugural recipient of the IEEE/ACM-CS Ken Kennedy Award in 2009.

In this interview, Fran Berman reflects on both her education and career. Specifically, she considers how she came to be interested in computer science, the results of her research, and how being a woman has affected her experiences.

About the Interview

FRAN BERMAN: An Interview Conducted by Janet Abbate, IEEE History Center, 18 July 2002

Interview # 577 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 Oral History Program, IEEE History Center, 445 Hoes Lane, Piscataway, NJ 08854 USA or ieee-history@ieee.org. It should include identification of the specific passages to be quoted, anticipated use of the passages, and identification of the user.

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

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

Interview

INTERVIEW: Fran Berman
INTERVIEWER: Janet Abbate
DATE: 18 July 2002
PLACE: UC San Diego, San Diego, California

Background and Education

Abbate:

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

Berman:

I was born in 1951, in Glendale, California. I grew up in Los Angeles and lived there until I went to college at UCLA. Then I moved to Seattle for graduate school; I went to graduate school at the University of Washington. My first job after that was in the Midwest at Purdue University, so I moved to the Midwest.

Abbate:

I’m going to back up a little bit. You grew up in L.A. What did your parents do for a living?

Berman:

My dad was a general contractor, and my mom was a homemaker.

Abbate:

Did you have any brothers or sisters?

Berman:

Yes, I had two brothers. I’m the oldest. My youngest brother ended up being an entrepreneurial computer guy, so he’s been involved with a couple of computer companies. My middle brother is an artist, and he works on “The Simpsons”—the TV Show.

Abbate:

Had you been interested in math or science at an early age?

Berman:

Yes, I was always interested in math; so I came to computer science through the math end of things. I was a math major in college and then went to graduate school in math, and I actually got into theoretical computer science in graduate school.

Abbate:

You got your B.A. in ‘73.

Berman:

Yes: B.A. was in ‘73; M.A. was in ‘76; and my Ph.D., which was in math, was in ‘79.

Abbate:

Had you used a computer before college?

Berman:

No. I didn’t use a computer during college, either!

Abbate:

When was the first time?

Berman:

In graduate school.

Abbate:

So until then you’d been doing more pure math? What area of math?

Berman:

You know, the general mathematics curriculum: topology, and real analysis, and algebra, and that kind of stuff.

Abbate:

What did you think you were going to do with that?

Berman:

I thought I was going to be a math major. The reason why I got into computer science is that, as I looked around for a thesis topic in graduate school, I got very interested in programming logics, which was an area at the interface of mathematics and computer science and philosophy. I ended up working with people in the Computer Science Department and doing a theoretical computer science thesis. So I was in the Math Department, but my real technical advisors were computer scientists, and I really ended up with a degree essentially in computer science, even though formally it was in mathematics.

Abbate:

And you went straight from college to grad school?

Berman:

I did.

Abbate:

So you knew you were going to go on in the academics. That wasn’t a question.

Berman:

After college, yes. I knew I was going to go to graduate school. I had no idea what I was going to do after that.

Abbate:

Had your parents encouraged you to have a career? Or to go into math or science?

Berman:

They didn’t discourage me. I don’t know that they particularly encouraged me to go in this particular direction, but they didn’t discourage me.

Abbate:

Did you think you would be a professor?

Berman:

You know, I don’t know if I really knew what I was going to be. I knew I was going to go to graduate school. By the time I got out of college I thought I would go for my Ph.D.; I wasn’t quite sure, really, what I would do beyond that. And during graduate school I had a lot of questions about what I wanted. I thought maybe I wanted to go to a teaching school, and then I got a taste of doing research and going to conferences and set my sights a little bit higher at the end of graduate school. I got a job at a research place—I got a job at Purdue University—and so then I got on the research track.

Abbate:

Were there a lot of women doing computing at U.W.?

Berman:

There were some. There weren’t very many. We had an informal group in Math—I knew some women in Computer Science, but mostly I lived in Mathematics; I was a T.A. there, and that was my peer group for the first few years. The women in Mathematics: there were maybe a dozen of us, and we used to have pot lucks every quarter. We got to know each other well, and what actually got a lot of us through was just knowing each other and that kind of informal bonding. There were a few people on the faculty who were pretty supportive, but for the most part, it was kind of benign neglect. There was one woman on the faculty, and she didn’t really care; and then, towards the end of my time there, a younger woman came on the faculty, and I’m not sure how much she cared, but she cared more than the woman who was there. But there was a guy, a Professor Emeritus, Bill Birnbaum. He was Emeritus then, so I’m sure he’s been gone for many years; but he was really supportive of a bunch of us, in our early years. He sort of took us under his wing, and it really helped.

Using a Computer for the First Time

Abbate:

So you actually used a computer for the first time in grad school?

Berman:

Yes.

Abbate:

What was that like? What were you doing?

Berman:

I decided that I needed to take programming, so I took a programming class and did what you do when you take programming classes: you stay up all night and try to figure things out!

Abbate:

And this is the late ‘70’s, I guess?

Berman:

Yes: mid-to-late seventies .

Abbate:

So what would that be in terms of technology? Minicomputers?

Berman:

No, I think we had cards or something! I don’t know! [laughs.] I’m not sure. I’m trying to remember the language we used. I think it was PASCAL, or something like that, and I remember having some programming projects. I didn’t take a lot of programming classes in grad school, but I took some, and I took a computer architecture course and a lot of theory courses. I did this from the Math Department, so I was out there on my own, not with this group in Computer Science. Later on I actually bonded more to the people in the Computer Science Department, but I really made that transition late during graduate school. It was kind of a difficult time to make that transition.

Abbate:

Did you take to computers right away?

Berman:

Well, yes; I liked computers a lot. I think that often people come to computer science from two different ends of things: either the math end of things, where you have appreciation for the aesthetics—and theoretical computer science is a really great venue for that—and then people also come from the engineering side of things, where they’re hands-on and want to see how things are made and built and constructed. I came more from the math side of things, so for me it was the aesthetics of it that really appealed to me.

Abbate:

What was your dissertation on?

Berman:

Nonstandard models of propositional dynamic logic. So it was programming logic, and I basically did these mathematical models of it. It was a very theoretical thesis.

Abbate:

So that’s almost getting to the philosophy end, or at least the logic parts.

Berman:

Yes. It was really a mix between philosophy and math and computer science. But it was in a really hot area at that point in time: programming logics and propositional dynamic logics were pretty hot areas, so there was a lot of interest in theoretical computer science in these particular areas.

On Purdue University and UCSD

Abbate:

Did you have a lot of offers when you got out?

Berman:

I was lucky. I interviewed at maybe a half a dozen places, and I got offers pretty much everywhere I interviewed. I decided to go to Purdue University. I never would have expected that I would end up there, but I got there and it was a great place. It was a very solid department, and it was a good area for theory; they had some very well-known people there. And I had an adventurous spirit, so I thought I would go. I had grown up in Southern California and stayed on the West Coast my whole life.

Abbate:

So Indiana was something new.

Berman:

Yes, I thought it would be fun. And it was!

Abbate:

Were you there until you came to UCSD?

Berman:

I was there for six years. During the time I was there, I met my husband, who was a grad student in biochemistry, and we got married. I’m three years older than my husband, and when he graduated we were looking for a place where I could continue doing what I was doing and he could get a good postdoc. So we looked at everything that was good for him and everything that was good for me, and we looked at all the places in the intersection, and we just interviewed around and ended up here [at UC San Diego]. This was both of our second choice. My first choice would have been to stay at Purdue, because I liked it a lot, and his first choice would have been to go up to Stanford and take a postdoc up there; so we ended up here.

Abbate:

But you’ve been here eighteen years or something?

Berman:

Yes, I’ve been here a long time.

Abbate:

So it couldn’t have been too bad!

Berman:

No, it’s been great! I’m quite happy we ended up here.

Abbate:

Were you in Computer Science and Engineering when you came in ‘84?

Berman:

Yes. I came to Computer Science.

Abbate:

How developed was that department?

Berman:

It was a little department then. The computer scientists were a division in EECS, and shortly after I came, we actually became a department on our own. But when I first came here, it was between eight and twelve faculty in this little program, and now we’re almost forty faculty in the Computer Science Department; so it’s grown quite a lot in the intervening years.

Being a Woman at UCSD

Abbate:

Were you the first woman there, or did they have others in computer science?

Berman:

No, I was not the first woman in the department. The woman before me had kind of self-destructed and became an Amway sales person! [laughs.] It’s true! And so there was a lot of [skepticism]. I think women are often asked to represent their whole gender: so one woman flaked out on them, and they thought, “Okay, are women [in general] going to flake out on me?” So I was the person that was [supposed] to prove to them that women were not going to flake out on them. Now, guys had flaked out in the department, too; but no one had expected that it would be a problem with another man.

Abbate:

“Everyone knows men can do it!” [laughs.]

Berman:

Yes! [laughs.] Actually, people were very welcoming, and they were quite good. I was the first woman in the Division of Engineering to have a baby while I was on the job, and people were very supportive. I would say that for the most part—not every single minute of every single day, and not every single person—but for the most part, people were very supportive of me.

On Balancing Work and Family

Abbate:

Was it difficult balancing work and family?

Berman:

It is always difficult balancing work and family. It never gets easy. I think if you’re a serious parent and serious career person, that’s a lot for one life! I think it is always difficult.

Abbate:

Was the university helpful about that? I don’t know how many people were trying to do this on the faculty.

Berman:

The university was good about that, but my experience was that it really depended on how I approached them. So when I found out I was pregnant—I was quite happy to be pregnant, and it was planned—I approached my Department Head and I said, “There’s really good news: we’re going to have a baby. Here’s my plan.” And I had planned it out so I was going to be free the quarter I had the baby and teach one two-hour seminar the quarter after that; and I made sure that all my responsibilities were not on anybody else’s shoulders, that they all got worked out. So it wasn’t a problem for anybody, because I made sure that they were not doing my work: that they could be really happy for me, because my work was getting done elsewhere. So people were really supportive. And I treated it like it was one of life’s natural things—because it is one of life’s natural things, and we shouldn’t act like it’s some bizarre thing that happens to people! So that’s pretty much how I approached it, and everybody else went along with that, and it was really fine. That was for the first baby. With the second baby, I actually found out—which I didn’t know with the first baby—that the university has a modified duties policy: you’re still active, but you have a lot of flexibility in what you do. So for my second child, who was born two and a half years after my first child, I arranged with my next Department Head that I wouldn’t teach the quarter my son was born, and then I took off a sabbatical quarter after that, and then I taught the Spring quarter; so that gave me a good time with him. And again, I saw my graduate students—sometimes I’d bring the baby in and sometimes I wouldn’t—so my jobs got done without anybody else having an extra burden. So there was no reason for anyone have a problem with it, and so people went along with it just fine.

Abbate:

But it was up to you to make those arrangements, so that all the work was going to get done somehow?

Berman:

You know, it never occurred to me that it should be somebody else’s responsibility. I figure when you’re a Professor, it’s a very autonomous position. Your prominence as a researcher, your research career: I mean, you’re driving, your colleagues aren’t driving; so you can manage it in your own way. That happened later on: I had breast cancer as well, and there was a time when I went through a year of chemotherapy and radiation and stuff like that; and again, it was me driving, and I figured out what I needed to do—because that’s not something, even in the most liberal and progressive of atmospheres, that someone else can figure out. Again, I was pretty straightforward with my colleagues. I said, “This is what’s happening in my life, and I could use all the support I can get, but I’m not going to make you do my work.” And people were really great about it. I’ve had a lot of support for those kinds of things.

Abbate:

It’s good that it was the kind of environment where you were able to have that flexibility.

Berman:

Well, it wasn’t just then; it was all throughout my life. I mean, my husband was an academic, too, so we’ve always shared stuff down the middle. We shared babies, we shared illnesses, we shared everything. So I feel like part of what made having kids work, in addition to the flexibility of being a professor, is that I have a husband with a flexible job who thought of it as part of his responsibilities as well—and that made a huge difference! There were a lot of times when my husband spent a lot more time taking my daughter to ballet and picking up kids. We would change who took the morning and who took the afternoon. He did at least as much as I did.

Abbate:

What do your kids do now?

Berman:

They’re teenagers now, so they spend time actively being teenagers! My son’s just about to start high school, and my daughter will be a junior in high school next year.

Abbate:

So we don’t know yet.

Berman:

We don’t know what they’re going to do yet.

Abbate:

Do you think they might go into computing?

Berman:

Well, my son’s sort of interested in science and computing. My daughter’s a dancer, and she’s really interested in being a choreographer. Modern dance, ballet, fine arts, stuff like that: performing arts.

On Parallel and Grid Computing

Abbate:

Now, you founded a laboratory for parallel computing . . .

Berman:

I founded two laboratories.

Abbate:

Were those two separate labs? Parallel computing and something else?

Berman:

Grid computing.

Abbate:

Okay, I wasn’t sure.

Berman:

Parallel computing was my researcher area of interest when I first came here. I founded a laboratory, and I had a lot of students, and then other faculty came and we shared it together. Then, when I branched out in the ‘90s into grid computing, there started being enough concentration of activity that we founded a Grid Computing Laboratory. Part of the reason to start a laboratory in a university venue is resources: so you can get the space that’s appropriate, and give the activity a name and a focus, and stuff like that. The Parallel Computing Laboratory actually still goes on today, and there’s other faculty associated with it who are now more in that area. Now I’m more associated the Grid Computing Laboratory, because that’s the area that I’m working in now.

Abbate:

Okay, so the Parallel Computing Lab: Were you actually building hardware, or was that more software?

Berman:

Software.

Abbate:

And was that for a number of machines in parallel, or were you using parallel hardware like a Connection Machine or something? What was the approach there?

Berman:

My focus in parallel computing was to look at mapping and scheduling issues for parallel computers, in particular distributed memory parallel computers. In the early days, we were writing some software that helped do automatic mapping and load balancing and do that in a performance-efficient way. Then, using that as a theme, I moved into Grid Computing, where you’re trying to do mapping and scheduling of a distributed set of heterogeneous resources. So we wrote software for that venue, and we worked with application scientists as well.

Abbate:

What is the definition of grid computing?

Berman:

A grid is composed of a number of resources networked together, and grid computing allows you to use them individually or as an ensemble. Using them as an ensemble means you can actually use those computers together as a big virtual parallel computer; grid computing addresses how one might do this kind of distributed parallel computing. Grid computing is also an umbrella for ways to access resources of any kind: remote instruments, data archives, computers, et cetera,. You think of the resources as a big group, and you want to use whatever subset of the group is relevant to your problem.

Abbate:

Is that analogous to a power grid, where you have different generators, and you can connect them?

Berman:

Absolutely! It is analogous to a power grid. The difference is that, if you think about the power grid: there are generators, there are transformers, and you can plug in your laptop, or your hair dryer, or your toaster in any socket and you get power. You don’t think about which transformers are being used, and sometimes you don’t think about who pays for it; you just think, “Here’s power.” That’s roughly what you’d like to do with the computational grid: “Here’s computing”. But the difference between the power grid and a computational grid is: you might get different performance depending on which resources in the grid you use. It’s not like you’re going to plug your toaster in and you’re going to make better toast depending on which transformer you’re going to use [laughs], but in some sense, if you’re going to plug your program into the computational grid, you’re going to get better performance on some machines than others, depending on how busy they are, how fast they are, which libraries are there, where the data resides; et cetera.

Abbate:

So it’s heterogeneous, and I guess it’s also . . .

Berman:

Dynamic.

Abbate:

Yes. I was going to say “context-dependent,” but I guess more dynamic. Now, that’s so fascinating, because my first book was a history of the Internet, and at the very beginning of the ARPANET, they had this vision of distributed computation that didn’t actually work at all in the early seventies, because the obstacles were just too big.

Berman:

Yes.

Abbate:

I mean, it turned into basically being an email thing. And it’s just so interesting that now that the technology is more mature, that’s more of a reality than it was going to be thirty years ago.

Berman:

If you think about it, this is an amazing [time]. It’s funny: I’ve just been preparing a talk that I have to give next week, sort of a “State of the Grid” talk, to the Global Grid Forum in Scotland; and one of the things I was thinking about in preparing my talk is that there are extremely sophisticated applications out there that do not, in fact, use grid services at this point. Walmart inventory control: every time you go to Walmart and you buy something at Walmart, there’s a little bar code on that object, and it gets scanned in and gets beamed through their satellite network to a centralized data repository that does real-time inventory control. If they’re running out of red Christmas trees, the order is in. They’re predicting the number of red Christmas trees versus the number of green or blue Christmas trees instantaneously, to make sure that they don’t run out. This is an extremely sophisticated inventory control system that’s distributed; it’s dynamic; it’s coordinated; and people use it all the time. My son plays Everquest, the multi-role-player game. They’re doing distributed data management; communication; their avatars are interacting with each other in real time. There are some very sophisticated distributed applications out there. The grid should make it easier to build applications like these, and I think one of the interesting things that’ll happen is that we’ll see the commercial world and the grid world come together to create the new generation of applications. So all these things, starting with the Internet—where your browser, of course, is another sophisticated application, but there’s more that you can do with the Internet than just have browsers. So I think things are really happening.

Working as Head of the NPACI and SDSC

Abbate:

The National Partnership for Advanced Computational Infrastructure: what is that, exactly, and when did that happen?

Berman:

About five years ago, NSF—I mean, it all started a little bit earlier than that, obviously—but NSF observed, quite correctly, that computational infrastructure has become really a first-class tool for disciplinary science. If you think about the Grand Challenge collaborations of the eighties, and some of the gigabit testbed applications of the early nineties, and the rise of grid computing, and all of that: all of that points to the fact that some of the most stunning results in disciplinary science came from the use of data management, computational devices, networking to remote instruments: all kinds of things. So the NSF wanted to look at really building wide-scale infrastructure for the scientific community, in order to catalyze more of those kinds of results—results that lie at the interface of science and technology. They developed the idea of Partnerships for Advanced Computational Infrastructure to bring a lot of people in to interact around building infrastructure: to get applications people in, computational scientists in, computer scientists in, technologists in; to have large-scale resources of big machines, but also to have this notion of the right kinds of infrastructures: portals, and tools, and libraries, et cetera. So the idea behind the PACI Program—Partnerships for Advanced Computational Infrastructure—is to build these large-scale partnerships. They had a competition. At that time there were four supercomputer centers, and each of the supercomputer centers submitted a partnership proposal, and NSF chose two. SDSC [San Diego Supercomputer Center] is leading-edge site for one of them, called NPACI: National Partnership for Advanced Computational Infrastructure; and NCSA [National Center for Supercomputer Applications] in Illinois is the leading-edge site for the Alliance, which is the other Partnership for Advanced Computational Infrastructure. The Alliance and NPACI both get large-scale funding, and we’ve both been getting this for about five years now—it was initially visualized as a ten-year program. I think other things may happen in the next five years as a result of current evaluation that NSF is doing about what it wants to do about large-scale infrastructure in the next few years; but these two partnerships get a lot of funding, have a lot of partners, and lead a lot of activities around building computational infrastructure. In the meantime, Pittsburgh, which was a former one of the four supercomputer centers, won the first terascale facility proposal that NSF put on; so they also have now what’s called PACI resources, but they don’t have a PACI partnership.

Abbate:

You’re head of the supercomputer center . . .

Berman:

I’m head of NPACI and SDSC.

Abbate:

So they’re two separate things.

Berman:

They’re separate, but highly related.

Abbate:

So SDSC is the physical computers?

Berman:

The San Diego Supercomputer Center, which is the building we are in right now, is an organized research unit on the UCSD campus. We’re a large-scale facility. We’re leading-edge site for the National Partnership for Advanced Computational Infrastructure. There’s a large-scale cooperative agreement, so we’re kind of the point of the spear for NPACI. There are actually other things that go on at SDSC other than NPACI, and there are other partners in NPACI other than SDSC.

Abbate:

Right. But NPACI doesn’t own physical infrastructure?

Berman:

NPACI builds software and has large-scale computing resources: there are computational resources at SDSC, at Texas, at Michigan, at Cal Tech; there are some resources at Berkeley; it’s varied a bit over the five years. So we have hardware resources; we also build software infrastructure; and we have projects that help propel the next generation of application results as well. So there’s a mix.

Abbate:

Okay, I think I see.

Berman:

And educational programs as well. So there are educational programs associated with NPACI—and the Alliance as well.

Abbate:

You must be collaborating with scientists from all of these other schools, and other disciplines.

Berman:

Absolutely!

Abbate:

So they could be in biology or something.

Berman:

Yes. A lot of biologists, actually.

Abbate:

That sounds incredibly stimulating!

Berman:

It’s a great job! It’s the world’s greatest job!

Abbate:

What is that like?

Berman:

It’s wonderful. I firmly believe that the whole purpose of all of this computational infrastructure and all of these resources is to make science go to some place you couldn’t get to before. This is like working at Toys R Us for scientists! Every day I come in and we make galaxies collide; we look at inner space through understanding the passage of information from inside to outside of molecules; there are just so many things we do here at the Center. We help the National Archives with our database technologies; we’re archiving electronic files from the government. If you think about presidential libraries: your presidential library of today could fit on a few CDs as a bunch of electronic files. What do you do with that if you want to understand all that information fifty years from now? What kind of mechanism are you going to use for communication fifty years from now? What’s going to happen to your tapes? Now, you’re going to transcribe this [interview] into stuff that we can keep on electronic media, but say you wanted to keep it on tape: what would you do? All of those are issues that people look at, and SDSC is large enough scale and is synergistic enough to look at these high-end problems and really crack them. It’s a really exciting place to be!

Abbate:

How did you get to be the head of this? Did that grow naturally out of your work in parallel and grid computing?

Berman:

A combination of things. My predecessor is Sid Karin, who is a professor on this campus. He was at the Center for fifteen or sixteen years, and it’s a pretty stressful, demanding job. When Sid started the center, essentially it was forty people, and it’s grown to four hundred people; the budgets have grown, the expectations have grown, the world that the Center has lived in has grown, and it’s a very stressful, demanding job. When Sid decided to step down and the campus had to decide, “What are we going to do about Director of SDSC?” they looked inward. They thought about doing an outward search as well. I was approached about what my interest might be about being Director of the Center, and it was a good time, in terms of my life and my career, to consider doing something like this—which is a real departure from being a professor and running a large-scale lab, but I’ve loved the Center forever, and I’ve been interacting with the Center since before there was a building. It’s really tempting if somebody comes to you and says, “Here’s a huge entity. Can you take it into the future?” It was a very interesting opportunity for me, and I decided to do it.

Abbate:

Do you have a vision for it?

Berman:

Absolutely! [laughs.] That’s my job. For the last couple of decades SDSC was primarily about the Big Machines, what kinds of software do we need to build for the big machine, how do we make users able to use the big machine profitably. We’re in the Internet age today: everything is about distribution and coordination; big machines are important components of the mix, but it’s much more coordinated and dynamic. It is incumbent on me to take the Supercomputer Center into the Internet Age from the Supercomputer Age. [Drawing on] my background with grid computing, and my work with applications scientists, and my own national and international contacts, we’re focusing on things that will take us into modern cyber-infrastructure. It’s a different thing, but it’s the kind of thing that happens when organizations evolve, and you have to evolve towards the future.

Abbate:

What skills does it take to run a center like this, that are different from being a research professor?

Berman:

My job is really focused on strategic thinking. It’s very important to be looking ahead. What should the Center be doing? What’s the right thing? You need political skills, because it’s a highly political job. You need a good sense of humor—because a lot of weird stuff happens! [laughs.]

Abbate:

Do you have an example of those things?

Berman:

None I’d like to put on tape! [both laugh.] These jobs are very demanding and highly political, and I think at the uppermost level of organizations, things happen because of a mix of people’s vision, of people’s energy, of political environments; of what you promise and what you deliver, and how you can balance all of the competing things to think about. So I think this job requires someone who’s a real leader, who can balance all of those different things. It’s an immensely stimulating job and a really demanding job.

Abbate:

Did those skills come naturally to you? Did you have to work at them? Did you have some mentors helping you?

Berman:

I think, like all good jobs, you grow on the job. I think that there are things that I brought to the job: I think I’m a good leader; I think I’m pretty independent; I think I have good political skills and I’m a fairly sensitive person. You know, I had no idea how difficult this job would be! [laughs.] I had no idea how political this level of organizational structure would be. And a lot of stuff you just learn by doing. You make a lot of mistakes; you hope that none of them are fatal; and you just keep doing the best you can. I’m still learning on the job. I’ve been on the job about a year and a half now. I think of the first year as: I was in the deep end, and I was breathing! [laughs.] Now I’m in the deep end and I’m swimming, and I hope to swim even better next year. I mean, these are just enormously difficult jobs. You just do the best you can.

On Commercializing Research

Abbate:

Do you have to go to Congress every year and convince them to fund you?

Berman:

Not directly but we talk to a lot of people in leadership and in Government. You have to talk to Congressional folks; you have to talk to government agencies; you have to talk to CEOs; you have to talk to university Presidents, Governors, all kinds of people.

Abbate:

You mentioned CEOs: this is also a business-academic partnership? Where do they come in?

Berman:

There are opportunities for strategic collaboration with large-scale centers, and so you have to figure out what your industrial partnerships are going to be with various and sundry companies—people for whom it makes sense.

Abbate:

Have some of your ideas been spun off into products, or worked their way into commercial software?

Berman:

The Center has a number of projects that have ended up going out in the commercial sphere in some way, shape, or form: either through a company, or through a product, or through an individual tech transfer. The University has pretty stringent rules about what’s okay and what’s not okay, and there’s a large office on campus, a large-scale infrastructure there, that helps us figure out what makes sense and what doesn’t. But we’re funny, because we’re hybrid: we’re halfway between academia and industry, and so rules of one place don’t necessarily apply to another.

Abbate:

I was thinking about your own previous research. It sounded very practical; it sounded like someone could really use it. I wondered if you, or even someone else, had commercialized it.

Berman:

In my own research, I’ve never spun off into commercial venues.

Abbate:

I was just wondering if you ever, for example, discovered some software and said, “Hey, I know where that idea came from.”

Berman:

Well there is [that possibility], but I think people make choices in life. I think I have done things as a researcher that were commercially viable, but you can only have so many full-time jobs in your life, and between being a full-time academic and a full-time parent . . . Starting a company would be another full-time job, so I think being a parent and an academic were enough! [laughs.] Although other people do it.

Abbate:

I didn’t even mean you, necessarily. I’ve talked to a couple of other people who said that someone else had—and they were just thrilled that someone was using their stuff, and I’m just curious if that ever happened to you.

Berman:

Yes. It’s exciting. Our stuff was always out there at the bleeding edge, so it would have taken a lot of work— reasonable work, but a lot of work—to get it into the more robust commercial sphere.

Abbate:

So maybe it would be a couple of generations later?

Berman:

Yes. I mean, research computer scientists often work with domain scientists who are willing to put up with out-there technology. Things often break when you’re doing some really interesting stuff—but in the commercial arena, you really don’t want things breaking. You want things to be very robust, and that last mile in developing usable, robust software—the documentation, the making sure that everything that can go wrong is not going to go wrong—there’s a lot of work that goes into that. That’s really a lot of what companies do: there are some very cool ideas where things start, but there’s a huge amount of work that goes into making software ready for prime time.

Abbate:

Do you think there might be two different personalities: people who want to do the cutting-edge research and people who want to actually turn it into something?

Berman:

Not necessarily, because I know a lot of people who have gone from the academic environment to start companies. It’s really exciting—very entrepreneurial, very high-risk—and some people succeed and some people don’t. There’s also a timing issue: there are times in your life when I think you’re ready for the next big challenge, and maybe that’s a good time for people to move to another university, or become an administrator, or start a company. So I think there are lots of things that come into play on individual decisions to do that.

The Treatment of Women in Computer Science

Abbate:

Now, you’ve obviously been very successful in your career. Did you ever feel that being a woman was any obstacle in terms of not getting equal pay or promotion or attention, or overt discrimination? Was it ever problematic for you?

Berman:

I’ll give you two answers. The first is that I think that it’s been all kinds of things: there were times when being a woman was a real impediment; there were times when being a woman got me a lot of attention that I might not have gotten otherwise; and there was a lot of time when being a woman was a real no-op : you know, nothing happened for or against. And so if you put it all together, it kind of all washes out.

Abbate:

It averages out.

Berman:

Yes. So for the most part, it hasn’t held me back from the things that I have wanted to do. For the most part! [laughs.] I’m the first woman Supercomputer Center Director ever for the NSF, and I have a different leadership style than my colleagues, and I think that people expect that they’re going to deal with the Director of the supercomputer in a particular way, and they expect you’re going to act a certain way, and they expect you’re going to lead a certain way. If people know you’re a reasonably nice person, a team player, they don’t think you’re going to make hard decisions; they don’t think you’re necessarily going to have vision; they don’t think you’re necessarily credible. One of the things that I’ve talked about with some of my colleagues is that a lot of times, being a professor, it’s like people assume there’s some different job for you—the girl job. [They think] that you don’t have to publish as much as anyone else, and your ideas don’t have to be as good as anyone else’s; somehow there are different rules that apply to you, because you’re nice, or something like that. And in this job, I think that’s a bit true, too. Because you’re not what people expect a Director to look like or be like or whatever, they don’t know how to deal with that. So, like any position, there’s this credibility-establishing phase—which lasts from two picoseconds to days or hours or months, depending on your audience! But once people are over the hump, they’re pretty much going to treat you just like anyone else. And this is a very high-stakes position, and so no one is going to do stuff for you just because they like you, or just because you’re nice, or just because you deserve it. Typically there’s a lot of risk; there are a lot of resources; very high stakes; and you have to perform just like anyone else. One thing that I think is really true is that what drives me is not necessarily the same as what drives my colleagues. We get in competitive environments all the time, and it’s my job as Director of the Center to make sure that SDSC is doing the right thing and being successful in these competitive environments. Sometimes I think my colleagues go out there, and their mental image is: “Kill the enemy!” [laughs.]

Abbate:

The “enemy” being the other supercomputer centers?

Berman:

Well, just whomever you’re competing against in these large-scale venues. And I always go out there and I think, “I must protect my cubs!” [both laugh.] You know, you have to motivate yourself somehow, and I think, “I can’t go back to the Center not having done this for us!” So I do think I take maybe a different approach than my colleagues, but we still have to fight it out in the same venues, and there are still the same rewards. So I do think it’s a little bit different.

Abbate:

Do you think computing in general has gotten more open to women since you started, or less?

Berman:

I’d like to think that; but it seems like we went through a bump in the eighties where there was a lot more diversity, and a lot more women colleagues; and it seems like now it’s closed up again, and I don’t know why.

Abbate:

Do you have any speculations on that?

Berman:

I don’t know. I watch my own kids, and I watch how they use technology, and my son and his friends: they get together and they kill things together—you know, Klingons etc.; and that’s fun for them. And my daughter never does this stuff. Sometimes my daughter does Neopets online. A lot of girls [do]. It’s been an interesting thing. We’ve had girls come through the Center, and they’re doing these Neopets. I don’t know fully what happens, but they’re like . . .

Abbate:

It’s some kind of virtual pet?

Berman:

Something you take care of, and you can have them play games with each other, and you feed them. My daughter does instant messaging, so she uses the computer for communication. She’s comfortable with it, but she’s not using it quite in the way that my son does. He’s managing his little warrior-weapon database, and he’s talking to his friend; his avatar is shooting somebody else’s avatar. The only thing I can think of is, there’s not as easy a venue for girls to get interested in technology, perhaps; but that’s just a seat-of-the-pants comment. I really don’t know.

Why Computer Science Is Interesting

Abbate:

Do you think it’s gotten more competitive to get into computer science?

Berman:

Possibly; I’m not sure. You know, I think you get kids interested by making things fun for them. So if you’re interested in math and science, there’s the “geek factor”: as kids go through junior high and high school, and they’re interested in looking cool to their peer group, so the geek factor is much harder on girls than it is on boys. You really have to like it to be in it, because otherwise the geek factor will maybe push you out. And I think that also, computers are not so user-friendly—stereotypically—to girls as they are to guys. Now, the Internet’s maybe changing things a bit. Everybody uses the Internet: young and old, and girls and boys, in all kinds of ways; but I’m not sure.

Abbate:

What have you found most satisfying about working with computers?

Berman:

Well, I’m not a big programmer. That’s not what interests me about computers, the big programming task. I use computers as a tool. So when I want to know what movies are close to my house, I use the computer; and when I want to give a snazzy presentation, I use the computer; and when I want to communicate with people, I use the computer; and writing on the computer is much more fun than writing with pencil and paper. But I don’t spend a lot of time programming, and I’m not interested in being a sys admin. I’m not so inclined to use it as an end to itself. I’m not interested in tools as an end to themselves; I’m using it as a venue to accomplish something.

Abbate:

I guess I meant in a broader sense, not necessarily programming. I mean, your work, on the theoretical or on the software level, is certainly about computing, or at least some of it is. So what interests you on that level, or what’s interesting about doing work that’s involved with computers on that level?

Berman:

In terms of my own research?

Abbate:

Yes. I assume there’s something appealing for you about not doing just completely abstract math, but applying it using the computer.

Berman:

Oh, okay: I just love where computers can take us. I mean, the fact that we can simulate the workings of the body, or how the universe began: I think that’s a tremendously wonderful thing. I think the fact that we can pull together all of this information and really understand; just from random time series from the sensors on the bridge, figure out when to get off the bridge because there’s too many stresses and strains and it’s dangerous for people. The fact that computational and data management infrastructure provides a way to get to big societal problems: for me, it’s just incredibly compelling. [The computer] is a really important tool for these societal problems. You think about integration of all the database of information that allows you to identify a terrorist, or figure out the solution to any hard problem: I think that’s a really exciting use for computers.

On Bill Birnbaum and Mentorship

Abbate:

You had mentioned a professor—I guess it was Bill Birnbaum—who had been a sort of mentor to you. Did you have other mentors along the way?

Berman:

Oh, absolutely! I’ve had a load of people who were really helpful in my career. I’m a pretty independent person, and there’s just been a number of people all the way through that I have talked to. When I was at Purdue, I spent a lot of time talking to Larry Snyder; he was a professor there. He was a really wonderful mentor. Here at UCSD, Sid Karin, my predecessor, has been one of my mentors; also Jeanne Ferrante and Larry Carter, two folks in the department. There are also Professors who have just been wonderful people to talk to. Some of the women I know in the community: Maria Klawe, and Nancy Levison, and Jan Cuny. A lot of people have really been just wonderful mentors. Rick Stevens, at Argonne National Lab, is a great colleague, and I talk to him about a lot of stuff.

Abbate:

Do you try to mentor other women?

Berman:

Absolutely. I think it’s pretty important that we pay attention.

Abbate:

I was noticing: I picked up this little recruiting brochure, and it’s all women, at least on the first couple of pages. [Shows recruiting brochure for the San Diego Supercomputer Center, titled “IT Professionals: The Best Minds in the Best Environment.”]

Berman:

What is this? [Looking at brochure.]

Abbate:

I don’t know if you had anything to do with this. It has all these pictures of women on the cover, and I wondered if that was deliberate.

Berman:

I’m sure it was! [laughs.] Yes: it says, “Career opportunities,” so I think the message here is that it’s really open to everybody at SDSC—and it really is. So I guess that’s the message here. Your External Relations Department always tries to figure out how to make that message in the best possible way.

Abbate:

Well, it caught my eye!

Berman:

Yes. That was good.

Abbate:

Do you find that women end up in certain sub-specialties of computing?

Berman:

Well, it would be interesting [to know]. I’ve always thought that there’s rich material for someone from the social sciences, to look at where women fall within computer science. Arguably, women are attracted to computer science—they’ve all gone through the geek challenge and all these kinds of things to get there; but I think the interesting thing is—and I just have anecdotal information here; I don’t know if this is true or not—but my sense is that more women follow the mathematical side of things than the engineering side of things. You typically see more women in theory and fewer women in computer architecture, for example. I think it would be interesting to look at that, and I think there may be some interesting answers there, but I don’t know what they are.

Abbate:

Yes, it’s an interesting question. Other than the obvious hardware/software thing, I don’t see an obvious pattern, and I don’t know if anyone’s counted up the numbers of women in different specialties. Maybe I’ll have to do it! [laughs.]

Berman:

Yes. Maybe this is your [task].

How Computer Science Has Changed

Abbate:

I’m curious if there’s anything there. How do you see the field of computing as having changed? I mean, obviously it’s changed a lot since you started; but what really stands out for you?

Berman:

You know, I think it’s like life itself: it’s just gotten much more complex, the environment. I think that in the fifties, sixties, seventies, eighties, it was about one machine and algorithms for one machine, even if it was an increasingly complicated machine. I think with the eighties and the nineties, we started looking in a much more multidisciplinary, multi-scale, multi-information way, and then you get the complexity of coordination. How do you coordinate all of these different resources? You lose control; you lose centralization of focus; and I think that makes dealing with the potential of all of these resources as an ensemble, as a computational infrastructure, much more difficult. I think that computer science as a discipline is grappling with, not just networks or data or computers, but how you pull all of those together and make the tradeoffs you need to make. You want your program to run as fast as possible: what are the tradeoffs you make between transferring your data, and using the network, and how many processors you use, and all those kinds of issues? So I think there’s been just increasing complexity. One of the other things that has made things more interesting is that computers have gone out of the realm of academicians and a bigger focus for the private sector. Everybody uses computers today. There are more IP addresses per person than there are just about anything else. We all have cell phones, and PDAs, and computers, and cars with computers—and we expect them to work together; I expect my cell phone to show me my email. And as you start looking at pulling together all these devices—and not just doing it for the biologist or the physicist, but doing it for the average person, who doesn’t necessarily know what the operating system in their cell phone is—then you have to deal with usability. You have to deal with coordination of these devices. You have to deal with the commercial realm, where there are many vendors, and there are different rules and different data and different formats; and that gets you into policy issues, economic issues, security issues, multi-country issues: my cell phone comes from Germany and my PDA comes from Japan. So there’s a lot more things in the mix than there used to be. Technology becomes more important, but I think it’s also harder to think about design from first principles and make sure that the design is comprehensive and makes sense. It all has to come together.

Abbate:

So it’s much more heterogeneous, in a way. It’s not just technology.

Berman:

It’s a fact of life: heterogeneity is a fact of life.

Abbate:

It’s very social and economic and political, right from the beginning.

Berman:

Right from the start, yes.

Abbate:

So when you said computing was like life, you mean . . .

Berman:

I mean life is heterogeneous. For any individual one of us as a person, we have all these circles: we have our work friends; and we have other parents we know through our kids; if your hobby is looking for UFOs or whatever, you have your group of people that you hang out with doing that; and your life becomes this mix of all these different circles interacting in all these different ways, and you have to somehow coordinate them on the individual level to make it work. It’s challenging!

Advice for Young Women Entering Computer Science

Abbate:

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

Berman:

I think it’s a great field, and I think there are some of the world’s most compelling problems, and I think it would be great to have more women in this field. I guess my advice is just, “Hang in there and stay the course.” Don’t get discouraged; don’t let discouragements knock you out of there; and just do it the way you think you ought to do it. Learn how to play the game; learn when you don’t have to play the game. Have some confidence in yourself, and just do it!

Abbate:

I was struck about what you just said about “Learn when you don’t have to play the game.” Can you elaborate on that?

Berman:

I think every professional figures out that there’s this trade-off between establishing your credibility and playing by the rules, and doing things in the way that benefits you and helps you accomplish your goals; and I think everybody figures out how to do that in their own way. Here’s a good example: often for early career folks to be taken seriously as an academic—whether you’re female or male— your colleagues want to think you’re always working, that you’re not spending a lot of your time reading books, going to movies, and picking your kids up from ballet lessons; they want to think you’re writing papers every single minute. It just seems to be the nature of establishing credibility within academic culture, and people have strategies for handling these expectations. Some people say up front, “I’m picking up my kids from ballet,” or “I’m going to the movies,” or whatever; but some people say, “I’m going to an off-campus meeting,” and that’s what they do.

Abbate:

And they’re meeting their daughter at the ballet lesson.

Berman:

And they’re meeting their daughter at the ballet lesson! I think that everybody kind of decides for themselves; but if you’re in your credibility-establishing stage, and you want your colleagues to think you’re credible, and if you choose to handle it that way, then you’re doing it your way and you’re also working the system. I think people do stuff like that all the time.

Abbate:

Do you think that’s more important for women, in terms of maybe having to integrate some less conventional things into their career?

Berman:

I don’t know that it’s more important for women. I think that oftentimes, depending on the family, a disproportional amount of child rearing, social organizing, house cleaning, whatever, falls on women. But that’s not true for all families. I think that when men pick up their daughters from ballet lessons, they can suffer loss of credibility too, depending on where they are in their career; and I think they don’t get as much support for doing that as women do for being in the workplace. So I’m not sure that I think this is just a women’s issue. I think the issue is that in some venues—a lot of venues—there’s intolerance for balance in life: good lip service, but general intolerance at the fundamental level. I think that we’re in a culture, in America, where we’re in this “excellence-at-all-costs” place, and I think that’s hard for professionals at every level. You get a lot of kudos from your colleagues for working 24/7, Christmas, holidays, birthdays . . . People think you’re serious if you’re doing that, and they kind of expect you to be doing that. I think everybody finds ways to get something else in their life at some level, but I think that we are almost encouraged to be secretive about that, as a general rule in our society. When you’re in a more powerful place I think you can say, “My daughter’s going to be a spoon today in the school play, and that’s where I’m going to be.” So I think that it’s really important—it’s incumbent [on people] in positions of power—to say, “I get it that you want to spend this time with your daughter—or your son, or your mother, or whatever—rather than be at work.” But I do think it’s a relatively intolerant society for the balanced lifestyle.

Abbate:

Have you been able to make any headway in that direction, as far as the people who work under you are concerned?

Berman:

I think everybody who works under me knows that I think that family stuff is important, and they know that I think they should be excellent at what they do; and if they’re grown-ups, I assume that they’re going to manage that mix well. If things aren’t working out, we can discuss it, and discuss whether it’s a time issue, or whether it’s some other set of issues. But I’m not a time-clock person. I don’t expect people to be punching the time clock.

Abbate:

Do you find they’re actually more productive if they are able to have a more balanced life?

Berman:

I don’t think that’s the key to it. I think productivity has to do with the approach you bring to the job and the skills you bring to the job and the expectations you bring to the job. A lot of the time, if you’re good at your job, you’re going to be good at making a balanced life for yourself, too! I think it has to do with the person, more than the number of hours they spend doing one thing or another. It’s an attitude thing, I think, and an approach thing.

Abbate:

Well, great! I think you’ve answered all my questions.

Berman:

Okay! So: Thank you so much for talking to me!

Berman:

Thank you!