Oral-History:Frances "Fran" Allen

From ETHW

About Frances "Fran" Allen

Frances "Fran" Allen was born on August 4, 1932 and grew up on a farm in upstate New York. There she attended the New York State College for Teachers (now the State University of New York at Albany), graduating with a B.S. degree in mathematics in 1954. Allen taught high school mathematics in Peru, New York for two years before she went on to earn a M.S. degree in mathematics from the University of Michigan in 1957. Later that year she joined IBM's Research Division, where her first assignment was to teach FORTRAN to other researchers. Although she originally planned to return to teaching high school mathematics after paying off her student loans, Allen ultimately remained at IBM for the duration of her career.

As an expert on compilers for high-performance computing, Allen's contributions to the computing field have been seminal. Her publications on optimization, interprocedural analysis and automatic parallelization have been instrumental to the science and practice of computing. In her own work at IBM, Allen transferred this technology to products such as the Stretch/Harvest Compiler, COBOL Compiler and Parallel FORTRAN Product. The Association of Computing Machinery recognized Allen "for pioneering contributions to the theory and practice of optimizing compiler techniques that laid the foundation for modern optimizing compilers and automatic parallel execution" by granting her the nation's top computer science award. As the 2006 A.M. Turing Award winner, Allen was the first female recipient of this prestigious prize. Helping to design and build the Alpha language in the late 1950s and early 1960s, Allen was also active in the work of the National Security Agency on code-breaking during the Cold War.

Since retiring from IBM in 2002, Allen has served as an IBM Fellow Emerita. She continues to be actively involved in the industry and works closely with professional organizations to increase the role of women in computing. Allen is a member of the American Philosophical Society and the National Academy of Engineers and a Fellow of the American Academy of Arts and Sciences, Association of Computing Machinery, IEEE and the Computer History Museum. She has also received honorary Sc.D. degrees from several universities.

In this interview, Allen discusses her 45 year-long career at IBM. She talks about many of the projects on which she worked from 1957 to 2002, including her experiences with early supercomputers on the Stretch/Harvest Program. In addition, she comments on her participation in professional activities outside of IBM during this time span. She served on multiple technology boards and spent a few semesters researching and teaching on college campuses. Allen also shares her opinions on the history and current status of the computing field, as well as of women's changing role within it.


About the Interview

FRANCES "FRAN" ALLEN: An Interview Conducted by Janet Abbate for the IEEE History Center, 2 August 2001

Interview #573 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, 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:

Frances "Fran" Allen, an oral history conducted in 2001 by Janet Abbate, IEEE History Center, Piscataway, NJ, USA.


Interview

INTERVIEW: Frances "Fran" Allen
INTERVIEWER: Janet Abbate
DATE: 2 August 2001
PLACE: Allen's office at IBM's T.J. Watson Research Center, Yorktown Heights, New York

[Notes courtesy of interviewer Janet Abbate]

Background and Education

Abbate:

This is an interview with Fran Allen, August 2nd, 2001. Can you tell me when you were born, and where you grew up?

Allen:

I was born August 4th, 1932, and I grew up on a farm, up in upstate New York. The town is Peru, New York. It’s near Plattsburgh. It’s not too far south of the Canadian border, in the northeast corner of upstate New York.

Abbate:

Did you have brothers and sisters?

Allen:

I’m the oldest of six—six in a fairly short order—and so I have . . . There were four boys and two girls.

Abbate:

Did any of the others go into technical fields?

Allen:

No. No, I’m the only one.

Abbate:

What did your parents do?

Allen:

My parents were—My mother had been a teacher, a grade school teacher I guess, and my father was a farmer. And we grew up—we lived on the family farm, which had been in my father’s family—and it still is in the family; my brother has it, now.

Abbate:

Did they—did your parents encourage you to think about having a career? Or . . .

Allen:

No, they . . . I think they always expected that I—all of us—would accomplish things. They never pressed us. They never really guided us, I would say, except to provide an environment where we were encouraged to read, and where we were encouraged to do and to flourish, I guess! It was a poor—it was a Depression farm—but we did—our entertainment was to do a lot of reading. But we weren’t pressured to do anything, particularly, and I never saw what I did as a career; I saw it as getting a job.

Abbate:

Did you have an early interest in math and science?

Allen:

Yes—though I think my very first interest was probably in English, and in writing. But I had a wonderful math teacher in high school, starting in eighth grade, seventh and eighth grade; and so in the end I became a high school math teacher. I went to New York—what was—what is now SUNY at Albany, but it was the New York State Teachers’ College at Albany at that time, and it was one of—the only teacher training school, I think, in New York state that was paid for by the state—we could go very inexpensively—and which trained people for high school teaching. So it was a possi[ble] . . . It was something that I could do, because I couldn’t do anything that was going to cost tuition.

Abbate:

So you were inspired by that high school math teacher?

Allen:

Yes! Oh, absolutely! Absolutely! Right. And I was perfectly happy to become a high school math teacher. That was just fine.

Abbate:

You went up to—so it was Teachers’ College then . . .

Allen:

Yes; it was Teachers’ College.

Abbate:

. . . and got a four-year degree in math?

Allen:

Yes: a four-year degree. One had to take eighteen credits in education—which was also very good, that it wasn’t any more!—and one then could take the rest of it in more technical subjects, or more—you know, the regular liberal arts training. So I took a major in mathematics, a minor in physics, and studied other things too. So it was very good for me, and I enjoyed it.

Teaching High School Mathematics

Abbate:

And when you got out, you started teaching math for a couple of years.

Allen:

Yes! I started to teach—I taught math for a couple of years—actually in the high school I graduated from! [laughs.]

Abbate:

Ah!

Allen:

Actually it was the school I graduated from—because it was just one big school. And I was the high school math teacher, so I taught everything from elementary algebra to advanced trigonometry, whatever was needed. I also—One of my favorite courses that I put together was for the local kids who were not going to be able to go on to college, but who would become the farmers in the area, because it was a farm community. It was just called “General Math.” So I kind of made up a course, because there weren’t textbooks on that at that time (there are now). I made up a course which was full of practical things, about keeping books, doing taxes, how you—measurements, and making—lots of percentages—everything, all of those things—so they were really handy with that when they had finished that one year.

Abbate:

It sounds like you enjoyed the teaching.

Allen:

Yes, I really did! I enjoyed it a great deal. I found it very rewarding, and working with this community—of course, many of them were relatives and children of friends—and so it was being part of a very good, very nice community, and knowing the kids and being able to do things for them individually, as well as just part of the classroom experience. I just enjoyed it a lot. And I had some great experiences with some of the . . . I remember one young fellow I had in plane geometry—which happens to be my favorite course; geometry happens to be something I’ve always been fond of—but plane geometry is a funny subject, in that one doesn’t understand what this proving theorems is about; I mean, everybody is kind of clueless about it, until suddenly a light bulb goes on, and suddenly one starts to understand what the pattern of the course, of the discipline, is.

And I had a young fellow in the class, who had had a sledding accident, and so he had some problems—cognitive problems—and I think a little bit of speech problems; he wasn’t very good in English and a lot of other things—but somehow, he really loved geometry! And he ended up: it’s impossible to trisect an angle, but he develop this elaborate proof at some point in the classroom, which we then sent to a professor at Cornell; and bless his heart, he looked at it very, very carefully, and wrote back—it must have taken him quite a while—wrote back what the flaw in the proof was. But I hadn’t, couldn’t find it, and it was really very deep! [laughs.] So it was things like that that make a big difference, and I think probably made a difference in this young lad’s life.

Getting a Job at IBM

Abbate:

How did you end up getting a job at IBM?

Allen:

Well, I wasn’t certified; one had to have a Master’s Degree to be fully certified as a teacher. I think that’s still true: you have to have a certain number of course credits, and meet certain requirements. So I went to Columbia University for one summer to get some credits, and then I realized this was going to take forever, so after two years of teaching, I went to the University of Michigan. I got a Master’s degree in mathematics there. IBM was interviewing on campus, and I had taken a couple courses in the very rudimentary computing stuff that was being taught at that time. The University of Michigan was one of the early places where computing was being taught, and so I’d taken a couple of courses in that. I was interviewed by IBM and got the job, and I thought, “Well, I’ll just take this for a year. Longer, no: I’ll just stay for a year, pay off the debts I have from getting my Master’s Degree, and then I’ll go back to teaching, which I know I like!” [laughs.] So it was kind of a throwaway job. I decided that—and it turned out, here I am! [laughs.]

Abbate:

Forty years later! [laughs.]

Allen:

Right! Forty-four years later. [laughs.]

I landed in IBM Research. I didn’t pay much attention to this. This was just a job to earn some money, and it ended up being the beginning of an absolutely fantastic career.

Abbate:

So your first exposure to computers was at Ann Arbor.

Allen:

University of Michigan at Ann Arbor, right.

Abbate:

What was that like? What were they using?

Allen:

Oh, it was a 650 [1], which is a drum machine. We would go in and run our own programs there. A guy named Bernie Galler taught the course. He was a young Associate Professor there at the time, and he was doing some amazing program on traffic control, I guess it was, for the City of Detroit—on what would now be an impossibly primitive machine! [laughs.] But we had real hands-on experience with running the programs and with debugging them, and it was a great course. It was taught in the engineering school, not in the mathematics department, but the mathematics department required that one have two courses outside of their department. That was, I think, a requirement of the university, in Arts and Sciences. So I took it in that, and that was my first exposure.

Abbate:

So you already knew that you might like working with computers, at least temporarily?

Allen:

Yes! Yes. I mean I didn’t view this as [a future career], I viewed this as fun and interesting, though my first love was geometry, and I really liked the abstract mathematics—very abstract mathematics. I took a course from a great blind geometer called [A. C.] Leisenring, who was just fantastic. But that’s not why IBM hired me, that’s for sure! [laughs.]

Abbate:

Did you ever end up using the geometry for anything?

Allen:

No, no. It just delights me, that’s all.

Teaching FORTRAN

Abbate:

What was the first thing you worked on when you came to IBM?

Allen:

I joined July 15th, 1957. The FORTRAN language, which was designed and developed by [John] Backus’s group, was released on April 15th, 1957; so it had been released. I was in the Research Division; the FORTRAN group was someplace else in New York City, if I recall. It was intended the research scientists would all move over to the new FORTRAN language, rather than program the Research computer, which was a 704, in assembly language; and of course there was huge resistance to that, because nobody could possibly make their programs run as well as they could on the 704—especially this high-level language. The scientists knew it would be a disaster! [laughs.] So, my job—because I had been a teacher, I think, and because I’d had a little bit of training—was to teach the scientists FORTRAN! [both laugh.] I had to learn it myself, because I didn’t know it either. So that was my first assignment, was to teach the scientists FORTRAN. That was an interesting experience, because I was learning at the same time they were. And I didn’t have a very happy class, of course [laughs], because they didn’t want to do this. They’d rather be doing their science and writing it in assembly language.

Abbate:

Now that’s so interesting, because, people who do history of computing usually think that higher-level languages were a boon, and that of course scientists would rather have FORTRAN that looks like math than have to write in assembler! Were they eventually happy with FORTRAN?

Allen:

Well, they were absolutely surprised with it—as everyone was. But what had happened is that over the years that it took to develop FORTRAN (and it was a long period of time, from [’53?] until ’57), it went from being quite mathematical formulation—I think I have a copy somewhere of that original formulation, which really looked like mathematical formulas—down to being something that they knew could produce absolutely spectacular code. They dropped a lot of function along the way, because they knew it wouldn’t be accepted if it couldn’t produce great code—and it produced absolutely spectacular code! That was one of the things that turned everybody around in that period. One would write something in the FORTRAN language and then see what the compiler would produce, and sometimes one of them would look and say, “This is wrong. This can’t be right—I don’t understand how it can be this good! What did they do?” [laughs.] They were doing tricks! You know, it was a great project, and there are some great quotes from John Backus about why it came out as it did. It was really knowing what was needed in the marketplace.

So that was my first job. And then—again, I think, hooking back to the fact I had been a teacher—the Director of Research was Art Samuel, who was very busy with his checker-playing program; he had a program on the 704 that was learning to play checkers. It was one of the very early learning programs, and this was in 1957, 1958. And so he was wanting to set up—I don’t know whether it was related to that or not—but he was wanting to set up some relationship with some of the local high schools in Poughkeepsie; Research was in Poughkeepsie at that time, on Boardman Road. And so I was to be the liaison between the Research Group and the high school. Now, that didn’t work out very well; somehow, for whatever reason—whether they weren’t interested, or whether I, you know, didn’t know how to make that kind of connection at that point—whatever.

The Monitored Automatic Debugging (MAD) System

Allen:

And so the next thing I did then was to get involved with a program on the 704 itself. I think it probably became MAD: Monitored Automatic Debugging System. I believe it’s the first operating system, and it had automatic debugging in it, and we added some stuff to the hardware console. It actually changed the mode of operation from being putting stacks of cards—or tape, or whatever, but it was stacks of cards in IBM’s case—into the readers, and then watching what was happening, to actually putting stacks of cards into off-line readers, onto tapes, which would then be read, and jobs could be just automatically sequenced, and automatically debugged—debugged from the console, by single-stepping from the console, from dumping from the console—and then just automatically going on to the next one. There were three of us that worked on that, and so we each had different roles, and it was a nice little group. That ended about 19—I’ve really got to get my dates worked out, but I think it’s some time in 1959. I don’t think I put the dates there, but I would . . .

Abbate:

[Looking at C.V.] No, there are no dates for that.

Allen:

Yes. I dug out some dates earlier, but that was finished up in early ‘59. I wrote a manual on that[2], and I remember giving a course, a lecture on it.

Abbate:

Now, I’m surprised: I didn’t realize anyone was doing on-line debugging, in the sense of being able to do that.

Allen:

Yes, yes. Yes.

Abbate:

Was that widely available?

Allen:

No. I think it was just—it was a research project; we did it on our machine. I didn’t do the actually engineering piece of this; my part was the software part: the debugging, and being able to read input from the tapes, and some other stuff related to the software parts of the subsystem. But you could actually push—we put a big red button on the console, which was called a “Panic Button!” [laughs.] You could hear the machines running, of course, at that time; you could tell when it was in a loop, or if it was hung, or if something was going on. So if you heard this—and every program had its own kind of sound signature, and the operators would get used to that, and the user could get used to that, too—the “Panic Button” would then stop that loop and then dump out memory. Then there was another button—I’ve forgotten what that one was called—which allowed one to single-step under certain conditions, and then you read what was in registers; you could just read the registers off the lights on the console. [laughs.]

Abbate:

Sounds great! [laughing]

Allen:

The great thing about that whole period there is that you could really know everything about what was going on, because it was all very new, and it hadn’t been tightly integrated! You could . . . [laughs].

Abbate:

So you knew the hardware, and you knew the operating system, such as it was.

Allen:

Such as it was! Yes. So. So that was . . .

Abbate:

So that was probably too expensive to actually have people debugging their programs that way?

Allen:

Well, I don’t know why we didn’t take and get it into market; I don’t think we were paying much attention to that then, you know. I don’t know. It certainly was heavily used in Research; it was a very nice thing, and we wrote it up, and stuff like that.

And people may have duplicated it elsewhere, because it was pretty easy to do. And also, it was at the end of this project, I remember, when I had my first experience—or at least, I perceive it this way—with somebody stealing the work. Yes! A guy who was not part of our little team—the three of us—who just had nothing to do with it, as far as I know, wrote a paper on it, on the work!

Abbate:

On the MAD system?

Allen:

Yes! Right. Right. And credited us, as a group, with having designed the cards and some other little stuff. [laughs.] I remember just being wildly angry about this!

Abbate:

He claimed he had done the work?

Allen:

He took it! [laughs.] And I remember being just absolutely furious about it! Because this was a small group, and in fact we were all young, and we were socializing a lot, and he was one of the guys kind of at the edge of the social group, and he just caught all of us totally by surprise! We didn’t know he was doing that. But anyway [laughs], I was angry!

Abbate:

Well, how did that work at IBM? If you came up with something, was there an internal publication? Or external?

Allen:

At that point it was just an internal one. He may have sent—put it out, sent it out; I don’t know.

Abbate:

Was there any sort of usual practice about how people documented things in the company?

Allen:

Well, that varies over the years; I mean, one can document—there is a review process, so somebody usually is assigned to signing off on something. I don’t know whether that existed at that time or not.

Abbate:

It was just more informal at that point?

Allen:

I think so, yes; I think so. And maybe it was just my perception, but I certainly didn’t know what was going on, and I was kind of in the middle! [laughs.] I felt like I was kind of in the middle of stuff, certainly with respect to that project. And I don’t think my—the guy that I was working for, who was one of the three, and he kind of was—you know—I think he got caught by surprise, too. [laughs.] Anyway . . .

On Women in Technology in the 1950s: A Promising Period

Abbate:

Were there a lot of women at the Research Center?

Allen:

Tremendous number! I have talked to people since then, I’ve said to people who were there at that time, “My perception is that there were, percentage-wise, more women there then than there are now.” In technical areas—and the Research Division was mostly focused on the hard sciences and mathematics; because computing was just kind of bubbling up, in some sense. And everybody agrees with me, that that was undoubtedly the case.

And I wanted to give you something—I’ll give it to you now: this is a copy of a recruiting brochure that IBM published in . . .

Abbate:

[laughs, reads:] “My Fair Ladies!”

Allen:

Look at the copyright date on that! I want to make sure that’s got a copyright.

Abbate:

‘57!

Allen:

‘57. IBM was out very actively recruiting technical women in the computing field. And of course there wasn’t—computer science didn’t exist, or any of that [type of training] . . .

Abbate:

For the record, I will say that this is a little brochure with flowers on the cover, and it’s called “My Fair Ladies” [laughs]—I guess a takeoff on the musical, inviting women to come work at IBM. That is remarkable!

Allen:

And it lists four, I think, four areas they were hiring for—we were targeting for: Programmer was one; System Analyst was one?

Abbate:

We have—Let’s see: There’s Systems Service, which is—that’s a customer relations kind of thing?

Allen:

Yes.

Abbate:

Instructing, which I guess is . . .

Allen:

The education program.

Abbate:

Programming; and Applied Science, which is . . . Is that more research kind of stuff?

Allen:

Yes, in mathematics, and lots of . . .

Abbate:

[reading brochure:] “IBMers are happy people!” [laughs.] That’s great; I’ve never seen or even heard of that!

Allen:

Yeah.

Abbate:

So it sounds like they were quite aggressively recruiting women.

Allen:

Yes.

Abbate:

Is that because there was a general shortage of programmers? Or did they specifically want more women?

Allen:

I think it was both. And it was a period when programmers weren’t being trained as such. I mean, the few schools that had little [classes]—like Michigan or Penn State, or Harvard, I guess, had something; there were few places that had some sort of training of this. You couldn’t get a degree in it; you couldn’t take much of anything in that. And yet the field was starting to really take off; at least in IBM’s perspective it was starting to. They were starting to think of it as taking off. So that, in Poughkeepsie—this came out of IBM overall, but in Poughkeepsie, many of the people that joined around the same time—many of the women that joined at the same time that I did—were from Vassar.

Abbate:

Oh, right, because that’s quite close.

Allen:

Yes, it’s right there; it was right there. And so when I joined, actually I needed a place to live, and so four of us just teamed, got together and rented a house out in the country, and the other three were from Vassar.

So there were many women, and I believe that during that whole period, there were many women in the field, as programmers, or in these other areas. I’ve also seen—I just happen to have collected up some lists of conferences that the government was having (I don’t know what agencies there were) earlier in the ‘50s, and it would have the list of attendees, and who was speaking; and there were many women just attending these conferences. And I also went through some very early technical journals in the field, and found that there were many authors who were women—so that they were doing highly technical work in computing; this was all in computing. So they were there! And they were contributing a lot at that point.

My own theory is that that was a kind of a golden era for computing. Because a lot of the processes hadn’t been developed yet; a lot of the requirements hadn’t been developed; or management of this kind of thing hadn’t been developed. And certainly in my own field of languages and compilers, more happened in those early years—starting, say, with real assembly language being developed in about 1951, ’50-’51, which is starting to get into a much more symbolic expression of a problem, out until the next fifteen years, into the mid-sixties—a huge amount was invented: different kinds of languages, different kinds of problems were solved; it was an amazing period. And it was a great period for women, too, even though women were not... You know, it was kind of a funny period for women: we were not part of the work force in the same sense as we see that now—it was not that way at all—but it was really a great period for them!

Working on Early Supercomputers: The Stretch/Harvest Project

Abbate:

Tell me about your first work on compilers.

Allen:

Well, that was this MAD program [laughs]—Monitored Automatic Debugging program. I had to dig into an assembly program to change its format, and so I got very familiar with how that worked. In fact, I completely changed that program. It was a marvelous program to work with; it had been done by a great programmer, [Roy Nutt,] who went on to found a great computing company, Computer Sciences [Corporation], and he and I kept in touch until not too long before he died, because of his interest in languages and compilers and things like that.

But my first one then—okay, having understood, knowing and having taught, FORTRAN—then the next big, major involvement was on a system called Stretch: Stretch/Harvest.

Abbate:

And that was sort of an early supercomputer?

Allen:

Yes. Absolutely. Yes. Stretch was going to be a hundred times faster than anything else. I ended up being responsible for the core center of the compiler work on that. And the Stretch machine, which was a supercomputer, had another machine kind of as a back end for it—attached to it—which was even bigger, which had been built for code-breaking for the National Security Agency.

Abbate:

Ah! Was that the Harvest?

Allen:

The Harvest machine. So I ended up—starting in the ‘59, ‘60, ‘63 period—ended up being the liaison with the National Security Agency on the language: the Alpha language that was being designed to express the problems of code-breaking. Just as FORTRAN is the [language for] science, they were designing, with us, a language called Alpha (it doesn’t stand for anything) for doing code-breaking on the Harvest computer. So I ended up having the optimizing part of the compiler and being the liaison with the National Security Agency on this Alpha language; and that liaison mostly entailed making suggestions—working with them on what we could handle in the compiler, what we couldn’t, and, in some cases, additional things that would be useful because we knew we knew how to do that, we could figure out how to do that. Then I ended up spending a year down at the National Security Agency, when we installed the system.

So I ended up, out of all of that, learning a lot about the optimization. We borrowed a lot of things from the FORTRAN people, who had done such a magnificent job of optimization. But we also failed to really make the machine—the Stretch machine—run as fast as it should on the big problems. I mean, we had overreached in some real way! One of the customers for Stretch was—I guess it was the National Weather Service; anyway, they were buying the machine in order to do weather prediction, short-term weather prediction. That was being written in FORTRAN, and the first version of that took 18 hours to do a 24-hour prediction! [laughter.] And so it wasn’t . . . [laughs.]

Abbate:

Only six hours of—

Allen:

—six hours of value! [laughs.]

Abbate:

Well, weather is certainly very computationally intensive.

Allen:

Oh, yes! And it’s been a wonderful, wonderful problem to work on over the years, because the models, and the computationally intensive, . . . You know, things have just gotten better and more in that field.

So that was my very first deep involvement with compilers and program optimization.

On the Status of Compilers Today

Abbate:

Just for the record: When you have a compiler, there are many ways you could actually implement the instructions: so [optimizing means] you want to do it the fastest way, right?

Allen:

Right, right. And it requires, in order to do that . . . I think of a compiler as mapping what the application person—like the person who wants to do a weather prediction—the expression of that problem (usually in a high-level language, like FORTRAN or whatever) to the machine in such a way that you get the right answer—but also that you utilize the resources of the machine to the greatest extent possible. And that’s what optimization’s about. You can just map things down, and not take advantage of registers, and lots of other computational units, [but that wouldn’t be as efficient]. So optimization’s about taking [advantage of the machine’s resources]—knowing, also, the machine very, very well. So it’s kind of bridging a gap—so the user won’t have to know all about it!

Abbate:

Exactly. And isn’t that crucial in the sense that part of the skepticism people had about high-level languages is that they wouldn’t produce very good code?

Allen:

Yes! Right. They wouldn’t . . . . And FORTRAN put that piece to rest! [laughs.] But it didn’t stay to rest, because . . . But it still is a lot of discussion about “How important is optimization on these machines, [now] that they themselves have become very fast?” I disagree with that point of view altogether; but I think we’re working on many of the wrong problems these days; we’re still focusing on solving some of the same problems we were trying to solve at that time. And the problem bottlenecks have shifted to memory and moving data around so that the data’s ready at the point of computation. But we aren’t doing a good job in compilers with that at all, so . . .

Abbate:

Hmm.

Allen:

Compilers have fallen into—I think this is a very low point for compilers. All fields kind of have their ups and downs. Usually the ups are when one has a nice breakthrough of some sort; or something shifts the problems that get worked on, or the view that one brings to the problem. And I think there’s no doubt that we’re not . . . We need a big shift right now! [laughs]

Working with the National Security Agency

Allen:

I gave a talk on this, on that project, out at the Computer History Museum recently—or in December [2000], I guess it was—not on the whole Stretch part of it, but on the Harvest piece, because that was a [secret]: the existence of NSA wasn’t even known at the time! [laughs.] We thought of ourselves as working for the Bureau of Ships—because that was code-name for NSA in the budget!

Abbate:

Ah!

Allen:

But we didn’t know! You know, we were . . .

Abbate:

Well, you must have known at some point . . .

Allen:

At some point, yes. [laughs]

Abbate:

When you went there, and there were no ships! [laughter.]

Well, I had never heard of Alpha, so I guess that was not . . .

Allen:

Yes. The Computer History Museum now has a—NSA now has a History Museum, which is on-line, it’s “outside the wall”; and I was really delighted to see that. We’re starting to see—because then I was able to get some retrospectives, a little, on the work that we had done: not a great deal, but some.

But that worked out—the Harvest part—worked out very well. And they went on—the NSA people—went on to develop a follow-on language, called “Beta,” but I wasn’t involved in that. I mean, that was quite a few years later.

Actually, Harvest lasted . . . I think we turned it over to them in ‘63—‘62 or ‘63; ‘63 maybe—and the machine was decommissioned in ‘76, I think? I’d have to look those numbers up. And it was decommissioned [only] because a part could no longer be found or made. I have a tape that—down there on the floor is a tape, and it was a part for one of those tapes that . . .

Abbate:

And that was a one-of-a-kind computer?

Allen:

Yes. But it was quite an amazing computer, in that the I/O rate . . . An instruction could last for hours and hours, and the I/O rate and the instruction rate . . . It just kind of streamed instructions through, and the instructions—there were about seven instructions, though many thousands of variants of these—but one instruction, for example, would be the sort, and it could do sorts, and do statistical analysis of the data that was streaming by it—manage those in various ways. So it was a different machine! [laughs.]

On Early Computing Culture

Abbate:

Did you feel like you were on the cutting edge?

Allen:

I think everything we were doing at that time was on the cutting edge; there was no question about it. Everything certainly that . . . Yes: It all was; we were inventing a lot of stuff!

Abbate:

It must have been exciting!

Allen:

It was! I can remember, when I was managing the Stretch Optimizing Group (part of the compiler): of course, nearly all of us were young; not all of us, but most of us were pretty young people, and I had these two guys who came in one weekend. The system had some basic data structures in it, of a fairly standard sort, and they decided they were going to turn them all into list.[3] LISP had been invented, you know, a little earlier— John McCarthy’s LISP—and LISP programs were coming out; in fact, a guy in IBM, [David] Gelernter, had a LISP program, so people were getting quite familiar with this. And so over the weekend they switched the whole basis of our optimizer over onto list. [laughs.] And another weekend they came in and introduced hashing, which was a brand-new idea at the time! This is a way of doing some lookup in a table, which is . . . And I would come in on Monday morning kind of aghast! [laughter.]

But that was the kind of free-for-all that we were working in! You know, people felt that that was okay, to give it a try. Brand new ideas would just go in and hit product systems—systems that were running against a schedule! [laughs.] It was fun!

Abbate:

Did you keep it in LISP?

Allen:

Yes! We did . . . Not in LISP—but we introduced list underneath it; we didn’t write it in LISP itself. The whole idea, notion: We kept a big piece of it in list; it was in mixing tables and list. It’s common now.

Abbate:

Now what is “list”?

Allen:

A list is a data structure.

Abbate: Oh, just a list.

Allen:

Yes. Yes. So we would—we mixed tables and list. But it was really a time when there weren’t as many restraints on processes and things.

Within IBM, what happened after that—kind of in parallel with some of that piece of work—the 370 was starting up, you know, and the base technology in the machines was changing a little bit. When we started on Stretch there was the 709, and that moved to the 7090, with transistors and things in it; so the base underlying technology was shifting, also. And then by the mid . . . When I finished Harvest, I wrote a final report on my section of the Harvest system—which I couldn’t keep, of course: it was confidential! [laughs.] I really regret, because I did spend—you know, I was one of the [last?] people on it, and I spent the good part of a summer on that, several months, writing the report—and it just disappeared into Fort Meade somewhere! [laughs.]

Returning to Research: The Advanced Computing Systems (ACS) Project

Allen:

But then the next job I was offered was on the 360, which was coming out, and I was in Poughkeepsie at that particular point. But the job they offered me was to coordinate the work on the languages, between . . . By that time that project had parts of it all over the world: the PL1 language was being done in Hursley, and there was a piece [of the] implementation at Burblingen, and, of course, the center of all the 360 was in Poughkeepsie. And I had spent a lot of time traveling back and forth to Washington, D.C., and I said I just didn’t want to spend the rest of my life traveling. I really didn’t want a job that involved a lot of traveling. So, I declined, and came back to Research.

Abbate:

So you weren’t going to do the 360.

Allen:

The 360: no.

Abbate:

What was the second town? You mentioned two other places besides Poughkeepsie, where they were doing work.

Allen:

Oh! That was in Burblingen, in Germany; and in Hursley, in England. I’m sorry; those are IBM cities that people know about. Yes.

Abbate:

Oh, so that would have been a lot of traveling!

Allen:

It would have been a lot of traveling, and I just didn’t want to do that. So I returned to Research, which was now located down here, at the Watson Laboratory. This laboratory where we are today opened in 1961.

Abbate:

Here in Yorktown Heights.

Allen:

Right, here in Yorktown Heights, physically. Though I had worked in IBM Research for several years—not in this building, but in Poughkeepsie, when it was still there. And then it moved; some groups moved down early, into a place called the Lamb Estate; and Herman Goldstine was the manager of that group, so we got to know him some. Anyway, this building started to be populated in 1961. So I returned here, I think in probably ‘63; and I specifically came to work on a project called “Project Y”—the letter “Y.” And John Cocke, whom I had worked with on Stretch some—he wasn’t as involved with the software as he was with the hardware, but I’d worked with him on the Stretch machine—he was involved with this other machine. And I came here and we built—we actually had a compiler and a simulator running for that machine, for this Project Y machine, long before the architecture froze; which was something one should do all the time, and we don’t! [laughs.]

Abbate:

So that they could make architectural changes to make the compiler run better?

Allen:

To make the system run better. It could find out how fast—you know, could run real programs through the compiler, through the simulator, and then look at how much time it was taking, how much it could overlap, how much it could use . . . how many registers one needed to have; you know, how deep the pipelines in the machine should be—that kind of thing. That project then ended up being called ACS—Advanced Computing Systems—and this was, again, a project to be faster than anything else!

And that moved, then, to [the west coast] . . .

There was a big meeting the summer of—I think it was in August of 1965; I have a lot of the papers on that stuff—which gathered a lot of people over at a place called Arden House, which was the Harrimans’ estate across the river, that he had given to Columbia University. And the Arden House conference, which was very small, was gathering people together to launch this as a product.

A funny incident: When I went to that, I was the only woman invited to that conference; it was a small number. But at this time, there were not that many women; this was now getting into the mid-’60s, and [laughs] . . . I walked into the [conference building] to register, to get my room; and the guy at the desk, who was from the West Coast, which was where we were going to move to—he said something. Without looking up, he said, “We’re having to double up; everybody has to double up, because we don’t have enough rooms.” And then he looked up, and he said, “Aah! [gasps] You’re—a woman!” [laughs.]

Abbate:

Horrified! [laughter.]

Allen:

He became a great friend, afterward. But it had never occurred to him that a woman would be coming to this conference. So the doubling-up was going to be Allen and Amdahl; Gene Amdahl! [laughs.]

Abbate:

So you got your own room?

Allen:

My own room. [laughs.] So did Gene!

So, that launched the project from here, from a Research project here to a West Coast project. So we moved out there, into what is now Silicon Valley; but we were in a tomato canning factory on Lawrence Expressway—I think it was Lawrence Expressway—which was just tomato fields! [laughs.] And apricot fields, and peach fields.

Abbate:

So IBM had just started this western research outpost?

Allen:

No, they had the Research Division out there before, earlier. I don’t know just when that started, but that started very early; and they had a plant site out there. But this was not going out related to that; it was going out [laughs] to get away from the East Coast—and its management!

Abbate:

And this was the ACS Group?

Allen:

Yes, ACS; right.

Abbate:

How long were you on the West Coast?

Allen:

The project was canceled in—oh, I guess it was . . . I really should dig out these dates. Well, it was in 1968 or ‘69. It was canceled then, and people kind of went different ways: some coming back here, some staying out there, some leaving the company. I stayed on and formed a compiler group, because the compiler was one of the very advanced pieces of work that we had. We had taken—I had taken—the lessons from the Stretch, and then really, really focused on optimization for this machine. And it was a group of [us doing this]—you know, I won’t take credit for [everything]. I’ll take credit for trying to do a pretty rigorous piece of work on it, in terms of turning it into a set of algorithms in mathematics, for a lot of that; but not for [everything]. There were two or three others; John Cocke, this person that I mentioned before, was very, very involved with that with me, and a couple of other people. We had focused totally on the methods to translate the language to various open-ended machine constructs, because the machine itself was not well-defined, and we wanted to experiment with it. And [it was a ] very complicated machine—very, very complicated.

So I stayed on because I wanted to; I didn’t want to lose that compiler work. And I convinced the Software Division, which was in New York at that time—not the whole Software Division, but the Compilers and Languages part was down at Time-Life [Laboratory]I convinced them that they should continue to support this work. So I then stayed out there with a small group and worked directly for Gene Amdahl, who also stayed on before he left to form his company. But I also wrote up a lot of what we had done in this whole period. So those papers laid in a nice foundation for compiler optimization.

Abbate:

Were you getting well known for this?

Allen:

Yes.

Involvement in Professional Societies

Abbate:

When did you start becoming active in professional societies?

Allen:

Oh, I think I became active—it was during that period, certainly; during that period.

Abbate:

[looking at Fran Allen’s C.V.:] It seems like you’ve done—I mean you were, I guess, fairly recently the Chair of ACM, or the President?

Allen:

No, not of ACM. I was head of the IBM Academy—but I was on the board of the ACM.

Abbate:

I’m not totally sure.

Allen:

But it was during that period. And was a member of ACM from very early on.

Abbate:

Right, I mean, I see various: “ACM National Lecturer,” and . . .

Allen:

Yes.

Abbate:

. . . editor of the [ACM Transactions on Programming] Languages and Systems . . .

Allen:

Yes.

Abbate:

That was just a natural part of the work, to be involved in these? Or was that something you particularly sought out?

Allen:

I think I sought it out, some. It was a very free-flowing period.

[break: recording pauses, then resumes]

Abbate:

So: I guess I was asking about professional societies, and . . .

Allen:

Oh! Yes . . .

Abbate:

So you would go to conferences and publish papers, and so forth?

Allen:

Yes.

On Women in Technology in the 1960s and 1970s: A Tough Period

Abbate:

How many women did you see at these conferences?

Allen:

A number; quite a number of women. I think that women in my field of languages and compilers—which is a very, very old field compared to some of the newer ones—that’s been a good field for women. It started early, and there were many women involved with it, for a long time. And in leadership positions, too.

Abbate:

Writing the first compilers!

Allen:

[laughs.] Right! Yes. And doing the first languages—and the programming, which is a piece of all of that, too. And certainly within IBM that was the case. We had a laboratory in Boston—in Cambridge, actually—which had a lot of focus on that, and that was run by one woman, and then another; and then our Time-Life Laboratory in New York, which was focused on that area, had loads of women involved with it. So it was a field that women were involved with. But . . . That, I think, has also changed. The percentages [of women] are just—it’s because, of course, the field itself has grown some—but its percentages have just continued to fall. I don’t have my hands on the numbers the way I wish I did, but I was following the NSF numbers for a while, in terms of undergraduates (B.A.s), and our computer science field was certainly not a stellar field—for women it was just falling precipitously!

Abbate:

Well, how do you think the field has changed over the years?

Allen:

I think it may have—and I’m just guessing; I hadn’t thought through that question—I think it changed in the ‘60s and ‘70s—late ‘60s and ‘70s, and probably into the ‘80s—in terms that it became a field: computer science got itself established as a field of study. (And it was a question for a while.) It grew out of, in some cases, the mathematics departments, but in many cases it grew out of the engineering departments. And in those periods—I believe that in the ‘60s, late ‘60s and ‘70s, there were not that many women in the engineering departments; not at all. And the other thing that happened is that as it became a profession, and as processes and requirements emerged, there were structures that were put in place, many of which were management structures and process structures; and as a profession then, it became an avenue that women were pretty much shut out of—in general. That’s my belief of what happened, from this very early, really wonderful period, with a lot of potential for women, to what one saw in the ‘60s and ‘70s and ‘80s.

Abbate:

So when you say it was more structured: Is that creating barriers to entry? Or was it that there was more of a promotional path upward . . .

Allen:

A path, right.

Abbate:

. . . that was favoring men?

Allen:

Yes, yes: the latter. Right. And I think we were not the—companies were not, and institutions as a whole were not focusing on diversity the way we do now. So there was a great—there was an old boys’ network [that] really thrived in that whole area. And for good reason: I mean, [laughs] there weren’t many women, and they weren’t reaching out to include them.

Abbate:

Did you feel that as a woman you had less access to pay or promotion?

Allen:

I think my situation changed dramatically in the ‘70s, for sure!

Abbate:

In what way?

Allen:

You know, it was, again: There were fewer women, and everywhere I looked, within where I was—not in the professional societies so much; or not within my own subfield, compilers and languages; or not in some parts of IBM—but in lots of places, there was a huge glass ceiling.

Abbate:

Did you personally feel affected by that?

Allen:

Yes! Yes, I certainly did. Absolutely.

Abbate:

So initially you thought you were being promoted, but then [later] you weren’t? Or [was it a problem] all along?

Allen:

Well, it was just . . . I didn’t think so much about promotion. I was being promoted fine¬, you know; not a problem. But I think then I started to realize I was working for men that I felt I should have their job! Over and over again! [laughs.] I’d think, “What am I doing working for this person?” [laughs.] “Why not me?” But it wasn’t about me; it was about the fact that there were no women being promoted. I mean, one would look around and see: the ceiling had really dropped, or at least it became very well perceived as having dropped.

Abbate:

So did you eventually find some way around this?

Allen:

Well, I think that that’s why I went and did a lot of things outside; I got very involved outside. And that’s a very common thing for women to do. I certainly had an opportunity, I was supported inside [IBM], in terms of having big groups to do a lot of nice experimental work; but I knew it wasn’t going to take me anywhere other than to be able to publish papers and get some very nice scientific results. It wasn’t going to take me anywhere [in terms of promotion]; that’s where it was going to be!

Abbate:

You weren’t going to be Vice President.

Allen:

I probably had my own [laughs]—there were probably other reasons why I wasn’t going to be Vice President! But it certainly was evident that those kinds of promotions—promotions where there were control points—[were] not going to be held by women, almost entirely.

Managing the PTRAN Team

Allen:

But I was well-supported in terms of having groups, and [intellectually?]. One thing I do remember: Because I had been doing a lot of work with compilers, one of the men in the department—who is now one of the IBM Vice Presidents that is in the paper all the time; he’s absolutely wonderful, he’s Irving Wladawsky-Berger—Irving came to me and said—it was a drop-by conversation; he came into my office, and he said, “Why don’t you start something on parallelism? Parallelism is going to be important, I think.” And IBM wasn’t really in it in a big way, and he just said: “Get something going on parallelism!”—in compiling in parallelism. And that was another fifteen years of nice work! [laughs.]

Abbate:

Well, it was a very hot topic.

Allen:

But it wasn’t a hot topic in IBM! But Irving has always been on the leading edge of stuff. And he finds the hot topics before they’re hot—and he recognized that one. And I think I’d been dancing around it a little bit, and so that was really great. And so I had several years of . . . That was good—terrific!

Abbate:

And this eventually became PTRAN?

Allen:

PTRAN. Yes, right.

Abbate:

Parallel TRANslation. And that was a product . . . ?

Allen:

It became—it actually found its way into products, the core of it did, yes. And loads and loads of papers. I mean, it became a very, very, very well-known system and group of people—just stacks of papers; we just poured ideas out all over the place! [laughs.] It was great! And I was able to hire whole lots of young people. That was another very exciting time from an idea point of view. And I wasn’t so much involved that time as I was, you know, providing the environment for these people to do it.

Abbate:

So you were a senior manager at that point.

Allen:

Yes, right. And I was able to provide a good environment for them.

Abbate:

What is a good environment?

Allen:

In a field like that, I’ve felt that we had to do two things: One was that it had to be not just papers; it had to be both experimental and theoretical. I felt we had to build a system where we could validate our ideas, as a group; and I made sure that every person had their own space in it. We had a base system, and they could take it in whatever [direction they wanted]. You know, they could do the experimentation and get themselves established professionally—because that was a bunch of young people there—with their own name on pieces of this. Yes. You know, that’s the kind of thing that a good university professor will sometimes do: keep a core system going for a while and get a lot of theses out of it, and stuff like this. But this was better—this was the same sort of idea, but there wasn’t going to be an end to their work, in the sense of, you know, “Get a thesis; get a job!” [laughs.]

Abbate:

But it sounds like you were really a good mentor to them, in the sense of having an idea what would be good for them to grow professionally . . .

Allen:

Well, yes. They’re just out of . . .

Abbate:

. . . in terms of having a project, being able to publish, things like that.

Allen:

Yes, right. Right. Rather than just work very close to a product, and just work-work-work on code, because that wasn’t going to be very challenging or influential either, in the long run—or writing a lot of papers. I mean, it had to have something that had both components in it.

Abbate:

You must been a great person to work for!

Allen:

[laughs.] Well, it was a great team. It was absolutely fantastic; it was a great team. It was described—I saw a description recently of that project, and it was one of the most successful groups in Research at the time. I thought, “That was written down, in this paper I saw!” [laughs.] PTRAN was one . . .

One doesn’t get an opportunity like that [very often]—and I did get that opportunity! So, it wasn’t that I didn’t have opportunities; it was that there was clearly a set that were off-limits.

Becoming the First Female IBM Fellow

Abbate:

But you became an IBM Fellow at that point? Or at some point?

Allen:

At some point in there, in 1989.

Abbate:

And what does that mean?

Allen:

Oh, it’s the highest honor a technical person in IBM can have. It means that one has the expectation and the freedom to work on what you want to. Within limits, of course; you have to convince somebody it’s going to have the potential of being relevant. But it does give one a huge amount of freedom—and also a lot of responsibility, because one is called on to make lots of recommendations on lots of things.

Abbate:

I saw the photographs of IBM Fellows in the cafeteria. It looked like they were maybe fewer than a hundred? Not a lot . . .

Allen:

Well, I think there are probably—the total is about 140 now, I think. There are around 50 active people.

Abbate:

And I saw maybe three or four women there.

Allen:

Yes.

Abbate:

I don’t know if you were the first woman . . .

Allen:

I was the first. Yes, I was the first. [laughs.] But, it was—you know, it was . . . Well, I think there’s a story. Listen, here’s something I’d love to have written up some time, is a story about the programmers. Because I was hired as a programmer, and so were many of these—the ENIAC women were programmers. And programming was not a very rew[arding]—[didn’t] have the kind of cachet that it does now, so much—I don’t know what we’d call it now, “software engineer” or something—but it was something that women were believed to be good at! [laughs] But I think there’s a whole story in there about the programmers of that time. Maybe that’s what you’re going to be able to bring out, some. Programmers were not recognized—the value of programmers was not recognized for a long time. My first award was on ACS, and it was only engineering awards: there were no awards for programmers, there were awards for engineers.

My first award, actually, was in the ‘60s. I think it was for—it didn’t say “programming,” but for something like that; for software or something—but it was a cuff-links and tie-clip! [laughs.] And, in fact—I mean, look at this! [Points to two nearly identical certificates hanging on the wall.] These are my IBM Fellow documents. The top one says, “In recognition and appreciation of his outstanding technical contributions, and to enhance his opportunity to add this to the record of achievement in the future.” And then [they redid it]. They wanted this one back; I didn’t give it back! [laughs.]

Abbate:

[Looking at the second document:] They did it over with “her”! They just hadn’t even noticed when they were writing that up?

Allen:

No! [laughter.] I accused them of going to the stock room! [laughs.] And they didn’t stock female versions of this.

Abbate:

Apparently!

Allen:

But, you know, that was a . . . That whole era [of the late 1960s-1980s, judging from her other comments] was not—you know, there were lots of good things about it, but it was not a great time for women to flourish. And that’s one of the reasons, I think, that at this time we’re struggling; because there aren’t the number of women leaders—technical leaders, I mean, the people who are very senior in the field. There’s a gap in here.

Abbate:

In here?

Allen:

In between [Q: women who came of age in the “golden era” of the 1950s and are now retiring and the younger generation of the 1990s onward who aren’t senior yet?]. You know, there’s not that . . . There are going to be some younger ones coming up; but even if one looks in the universities, the number of professors in departments that are women [is small]. Over and over again: you can to go to Berkeley, you might find [one]—well, I haven’t looked recently: five years ago there’d be one senior professor; you go to Harvard, you’d find another—

Abbate:

If you’re lucky.

Allen:

. . . MIT, you’d find a third, in computer science. [laughs.] So it’s a . . .

Outside Work at Universities

Abbate:

Now, some of the outside work you did, I notice, was at universities: Stanford, NYU. You were teaching in graduate programs at both places?

Allen:

Well, at NYU I took a sabbatical. IBM Research has a sabbatical program, and so when I finished on the ACS on the West Coast, before I returned back to Research, I took a sabbatical at NYU, at the Courant Institute.

Abbate:

This was 1970–73?

Allen:

It was actually 1970 and ‘71. I then taught as an Adjunct Professor for a couple more years, so that’s why those dates. But the sabbatical was for a year. And I got married in ‘72, to Jack Schwartz, who was the head of the department at Courant. [laughs.]

Abbate:

Ah! [laughter.]

Now how did you end up there? Did you just want to get back to teaching for a while?

Allen:

Well, John Cocke had taken the year before, he’d taken a sabbatical, and it had become kind of a center for compilers and languages. It really was a very, very strong department for that and so that’s why I went there. So I was wanting to... And actually there was a book in progress, which I was going to get involved with, and then I didn’t. On compilers . . .

Abbate:

Were you able to shape the program at all, based on your experience actually writing compilers?

Allen:

I taught compiler graduates two or three years there. And I think apparently I did—I’ve heard people who were in my class say that it was good! [laughs.] But you know, this was a big [lecture], once a week, sixty or seventy people in the class.

Have you heard of Anita Borg?

Abbate:

Oh yes!

Allen:

She was one of my students! In one of those classes. [laughs.]

Abbate:

Ah! I interviewed her in January.

Allen:

Yes, yes: a wonderful lady. And there have been others . . .

Abbate:

And then I guess in 1977 you were . . .

Allen:

At Stanford.

Abbate:

Teaching, or . . .?

Allen:

No, that was just a sabbatical, at Stanford. I came back here and was working and then Jack and I went to Stanford for a year. I’ve forgotten . . . I think I had some interest in going back to California. Yes. I mean, I enjoyed living in California! [laughs.] At that time it was a pretty great place! It’s still a great place.

Abbate:

But a bit different, I guess.

Allen:

It was less crowded.

Abbate:

Did you find there was a lot of interaction between the industry and academic people in computing? Or were they really two worlds?

Allen:

Well, everyone wants it to be stronger. I don’t find it two worlds so much as hard to really get a lot of strong—well, we work very hard to get a lot of collaboration going. I mean, I worked . . .

[CHANGE DISC - BREAK IN INTERVIEW]

On Balancing Work and Family

Abbate:

We were talking about family. I don’t remember . . . Let’s see: you take big vacations . . .

Allen:

Right. I don’t know what else I was saying. I do like to, yes.

Abbate:

So that kind of gets you away from [work] . . .

Allen:

Right—very much away from work.

Abbate:

What kind of hours do you work? Is it a 40-hour week? A 60-hour week?

Allen:

No, it’s much more [than 40 hours]. I tend to—my work style, or just my personal style, is to get very involved with something. It’s kind of the way I work. I like that; I get very engrossed in something.

Abbate:

Was it hard [for you and your husband] to balance two careers?

Allen:

Well, fortunately, Jack was at NYU and I was here, and we kept two places—one in the City and one up here. [laughs.] So that got confusing! And he had two daughters, so we spent, of course, a lot of time with them. But he is very involved with his work, too. So that was good. But we’re divorced now. But I don’t think it was about balancing careers, or anything like that.

Working on Technology Policy

Abbate:

So, we left off: You’d worked on the PTRAN project. What are some of the other major things you’ve done since then?

Allen:

Well, there was a whole series of projects—including PTRAN and some stuff after that—and I also ended up for a while having an office out in California, at one of our software labs, so I’d spend part of the time out there. I also then, after that, got very involved with some of the boards in Washington, D.C. I actually had too many at the same time! [laughs.] And it was a time when IBM Research wasn’t too interested in our being involved with outside things, or as far as I could tell; they didn’t value it very much. But I was on a board for five or six years at the National Science Foundation, and these would meet, but there would be work in between. And I was on another board at the National Research Council, called the Computer Science and Telecommunications Board, which did involve quite a lot of work in between, working on different reports. The board, like the other National Research Council boards, kind of independently develops reports on various subjects. So I got involved with some of that, which I enjoyed, but it was a different kind of thing than getting involved into the meat of some technical kinds of stuff. But it was a very, very, yes, edu[cational] . . . I found it intriguing about how decisions get made, and things happen, and ended up with a tremendous respect for the people who do that sort of thing full-time.

Abbate:

Sort of technology policy things?

Allen:

Yes, technology policy. Really good people. So, it was the same thing as when I was with NSA: fantastic people, whose names will never be known in the technical community, but who were doing incredible work! It’s good: a lot of goodness there. And one of the other things is that, at the national level, there’s a huge focus on diversity, and it’s at every point. “We’re going to have a study”: the question of who’s on this study is—a piece of the question is the diversity of the people. And they find women all over the place! I don’t know why we can’t! [laughs.]

Abbate:

Good for them!

Allen:

Good for them, yes! And it’s not just about diversity with respect to women; they do a fantastic job on that. And the National Science Foundation is just having an amazing impact, as you’ve . . . you’re closer to that.[4] [laughs.]

Abbate:

They do try hard. I hope they’re having an impact.

Allen:

Yes, I think they are.

On Mentoring Women in Technology

Abbate:

I noticed there’s a mentoring award named after you.

Allen:

Oh, yes! It’s that big one in the corner. [points to a poster on the wall.]

Abbate:

[reading:] “The Frances E. Allen IBM Women in Technology Mentoring Award.”

Allen:

The one with the torch. [laughs.]

Abbate:

Is that an annual thing?

Allen:

I think it will be annual. I got it last year—I received it last year.

Abbate:

Can you tell me about that?

Allen:

Well, I was—it was . . . IBM women, the technical women, have been having a conference every [two years]. That was, I think, the second conference that we had; and they decided to give the award at the conference. I had become a big advocate for women—that’s not too surprising!—but I also was particularly focusing on women who were just joining the organization, in part because [when] people come—and I’m sure it’s true everywhere—come to an organization, particularly if they’ve just come out of school, they suddenly find themselves in a very alien environment! [laughs.] And the culture is different; and how to succeed is very puzzling; and there’s a management system, and a performance evaluation system—you know, every company has all this stuff. So I had started to try and make myself a focal point for women that were entering our organization, so that they could come [and] I could try and help them build networks. I think that’s one of the big things that mentoring can do, is to help expand the networks. Because people get very isolated when they first start.

Abbate:

Did you have mentors yourself?

Allen:

No. And in fact I don’t have one now, and there are plenty of times I wish I did! [laughs.]

Abbate:

That’s an interesting question!

Allen:

Yes, yes, yes. And I don’t know why. I’ve gone to somebody, to people; I think of something and I say, “Okay, I’ll go and talk to this person about this problem.” But I’ve never had a mentor.

And in fact, general mentoring is relatively new in most organizations.

Abbate:

Well, formally, yes.

Allen:

Yes, formally. And also, in technical fields, it’s still trying to find its way, I think. [But] for promotions, new job opportunities, there’s a pretty clear—there is a mentoring structure in there.

Abbate:

In the more managerial track, you mean?

Allen:

Yes. Right. So I think we still have to figure out what to do for technical people, technical women: what guidance and help do they need that’s different? How does that work? So a number of us have been trying to toss that around. That is, within the technical community, how do we think about mentoring differently than people on the executive track?

On the Rewards of a Career in Computing

Abbate:

Do you have any advice for young women who are thinking about entering computing?

Allen:

Oh, I think it’s a great field; there’s no question about that. I’ve given some talks at univ[ersities]—not so much anymore, but for a while I was giving some talks about “This is the real beginning of the interesting part of computing!” [laughs.] Which it absolutely is! I said, “The first fifty years were all about laying the basis for it, the groundwork; but things are getting integrated now, and it’s the integration, and the ability for it to reach—touch—everybody; that is where computing is really—where the real excitement is starting.” And affect the lives of women in, I think, that way.

Abbate:

What have you found to be the most satisfying aspect of being in computing?

Allen:

Oh, I think it’s because [laughs]—because it’s new!

Abbate:

It’s always new!

Allen:

It’s always new—and changing so fast; it’s so dynamic! And the number of challenges: I just love that—that it’s intellectually and technically challenging in every way. So that’s what I’ve really, really enjoyed about it.

My hobby—or I shouldn’t say “hobby”; I don’t think of myself having a hobby so much—but something I do a lot of, is hiking and mountain climbing. And, you know, it’s somewhat of the same sort of thing: it’s kind of challenging, and interesting; and how does one involve oneself in it? What capabilities does one bring to it that will make a difference?

Abbate:

I never thought of that.

Allen:

It’s exploring! [laughs.] And I do like to explore.

Abbate:

It’s an adventure.

Allen:

It’s an adventure. [laughs.] And when you can make everyday work an adventure, it’s not boring!

Abbate:

Before I wrap this up, have I missed any major milestones?

Allen:

I don’t think so. No.

The Blue Gene Project

Abbate:

What are you doing right now?

Allen:

Well, I’m on a project called Blue Gene—G-E-N-E [laughs]. And that’s going to be—it’s a set of projects; there’s actually a bunch of things—but it’s very much in the same pattern as the earlier projects I was on, in that it’s going to be some very powerful, fast computing capabilities, and it has solving some important problems—some problems that need that kind of capability—and then there’s a big software hole in the middle of it. Because [of] the problems that need to be solved, and the power of the computing, there needs to be some very clever, good software put in there so that the biologists can do the experiments that they want to do. Model the [processes]—protein-folding is very specifically one of them that we’re doing, but there are others that will be related to it—on a machine whose scale will be very, very large, in order to provide the computational power that’s needed. So, it’s kind of a revisit of the old—of the kinds of project I had early on worked with—or have worked with pretty much all along. So that’s what I’m working on.

Abbate:

[It’s similar to the old projects] in the sense of being a really huge computational [problem]?

Allen:

Yes, right.

Abbate:

And the “Blue” is for “Big Blue?” [the nickname for IBM]

Allen:

Yes! [laughs.] Actually it came from the chess-playing [machine]: We used the words “Deep Blue.”

Abbate:

Right. But wasn’t that also [the IBM nickname]?

Allen:

Yes, right. And this one is Blue Gene.

Abbate:

That’s great!

Allen:

Blue Gene, yes.

Abbate:

Well, that sounds like that will keep you busy for a while!

Allen:

Yes! [laughs.] Right.

Abbate:

Well, I want to thank you very much for taking the time to talk to me. It’s been really fun.

Allen:

Well, great! I’m glad you came up.

Notes

1. All computers mentioned in this interview are IBM models, unless otherwise noted.

2. “Monitored Automatic Debugging System,” a user’s manual for PK MAD, February, 1959.

3. “List” is the name of a data type.

4. The oral history project of which this interview is part was sponsored by the National Science Foundation.