This published version of Oral History #80 with Dr. Jerome R. Cox reflects clarifications and emendations made by Dr. Cox after the transcription was completed.
This is an oral history of Jerome Cox taken on the seventh of June, 2006 at one o’clock p.m. by Simon Igielnik in the King Center of the Becker Medical Library at Washington University, St. Louis, Missouri. Dr. Cox wants to start out making a statement, so we’ll start with that.
Well, I’ve had a wonderful life and I just wanted to say that I’ve had the great support from the lovely and charming wife throughout the entire professional lifetime. I’ve got terrific kids, grandkids, and a great granddaughter. I couldn’t have done most of the things that I’ve accomplished without examples and help from bosses and co-workers and friends. I want to emphasize that as “Roddy” [Henry L.] Roediger has said, “Memory is a sly deceiver,” but I will do my best to remember as accurately as possible. I plan to include some interesting stories and I hope that’s okay.
I guess I want to start at the beginning. Where and when were you born?
I was born in a hospital in Washington, D.C. on the twenty-fourth of May, 1925. My parents were then living with my mother’s grandparents in Hyattsville, Maryland. My maternal grandfather was an Episcopal priest who had his church there in Hyattsville. At the time, my father was working in the U.S. Patent office in downtown Washington, which was before, of course, it moved to Silver Spring [interviewee’s note: Crystal City, Virginia, not Silver Spring]. At the time, I was no more than three – so I don’t remember it very well – but I do remember that I ran away from home down to the train station to see the trains go by, which caused a considerable uproar. But apparently I survived.
Tell me a little bit about your family growing up.
As I said my father was a patent attorney, so he started out in the patent office as an examiner. Then he worked as a patent attorney for Pratt & Whitney in Hartford, Connecticut; [he] later moved to Curtis Wright in Long Island, finally to Bendix in South Bend, Indiana. My mother had a strep throat which developed into rheumatic fever and the valve dysfunction led to her death in 1935. My dad remarried in 1938 and an interesting side note there is that I was the one that introduced my dad to my new mom.
As a youth, what motivated you to study science?
Well the incident that may explain my interest in electronics has to do with Bing Crosby. I liked to listen to the radio at night, but was scheduled to be in bed asleep with the light out at nine o’clock. And the show came on at nine in the Midwest. So I had a little Crosley plastic radio that I took under the covers, so that I could listen without my parents being aware of it, but the pilot light betrayed me because they could see it through the covers. So I took the radio apart in order to disconnect the pilot light and saw all these marvelous components inside the radio and that led me to wonder what they all did. I bought the ARRL Handbook, which is the American Radio Relay League Handbook and [that] was written at a level that I could just barely understand and got my interest up in electronics. Then I guess the next turning point was a geometry teacher, whose name I unfortunately can’t remember [interviewee’s note: Miss Kitson], when I was a sophomore in high school who opened the door to math. I had never been very enthusiastic about arithmetic and maybe just a little more enthusiastic about algebra, but geometry turned me on completely. That teacher enrolled me and two of my good friends in the state math contest. We went down to Bloomington, Indiana to the University of Indiana to that contest and one of my friends came in first in the state and my other friend came in twelfth and I came in thirteenth. So that was an important turning point that convinced me that math was fun. My plan was to – when the war broke out – [was] to sign up for V-12 and go to Purdue, which had this program for officers in the Navy called V-12. [ed. note: The Navy V-12 Program allowed students to complete baccalaureate degrees at civilian universities and earn commissions in the Navy and Marine Corps during World War II.] But in the spring of that year, which was 1943, I was doing the exercises that we were all supposed to do to get fit and crawling across monkey bars out behind the school – [I] fell off and broke my wrist. The Navy decided I was no longer physically fit with a broken wrist and so my admission to V-12 was cancelled. So then it was necessary to decide where I was going to school. And because the friends of my parents were so impressed that I had been admitted to this school that I had never heard of in the East called MIT, I decided maybe that would be a good place to go. But leading up to that, I guess I had a hard time deciding whether I was going to be a comedy writer or a radio announcer or an electrical engineer. So it finally ended up that I chose electrical engineering.
Do you feel that comedy writing was a path that you ever wanted to go down?
Well, I joined the MIT humor magazine staff, called “Voodoo” and enjoyed that. Also [I] was with another guy [and we] wrote the book for a musical in college. And I’d written some plays, or satires I guess, in high school. But electrical engineering took more and more of my time and so most of that has waned except perhaps for the writing of a few edgy limericks now and then.
You received your undergraduate degree in electrical engineering from MIT in 1947?
Yes. I went there in the summer of 1943, but got my draft notice late in that summer and ended up in the infantry as a replacement in first North Africa and then Italy. I didn’t stay in the Army very long because I got an infection in my foot and was in the hospital in Rome for about three months in the summer of ’44 and finally went before a reassignment officer who was a young M.D. who was looking at my record and finally looked up at me and said, “Soldier, if you were discharged what would you do?” And I said, I think unhesitatingly, “I’d go back to MIT.” And he said, “That’s all soldier,” and the next thing I knew I was discharged. I’m not sure that I really deserved to be discharged for medical reasons, but I was.
Then you went from there back to MIT?
Yes, immediately. And I think a semester was starting in November of ’44 because they were on an accelerated schedule and so they started each semester at odd times just as fast as they could go with classes on New Year’s Day and Saturdays and so forth. I was, I think, the first returning combat infantryman to go back to MIT and I didn’t have any trouble adjusting although sometimes the professors, I felt, were watching me carefully to see if I had any strange behaviors. But I enjoyed it very much. I was on something called Public Law 16, which was in effect before the GI bill was enacted. And Public Law 16 was a consequence of World War I and provided support for students who wanted to become teachers. I hadn’t made up my mind that I wanted to be a teacher, but it seemed like a reasonable profession and so I had an open mind toward that.
And immediately after that did you go directly to graduate school?
I was taking a senior course in electronics from “Doc” [Harold Eugene] Edgerton, the inventor of the flash, and he took me aside one day and said, “Would I be interested in graduate school.” I hadn’t really thought much about it before that, but he said he thought I could do it and he fixed me up with a graduate position in the acoustics lab at MIT. “Doc” Edgerton was important in my life in that way and in my wife’s life because after my wife and I were married in 1951, he hired her as his secretary. She enjoyed her fifteen minutes of fame as a result of being shot skipping rope in a darkened room and appearing in a handful of books of his photographs in a picture called [Moving] Skip Rope that was once hung at the St. Louis Art Museum and I don’t know where it is now.
Edgerton was in the Acoustics Department?
No, he was an electrical engineer and his lab was on the second floor of Building 20, which was left over from the war development of radar. The acoustics lab was stuck onto Building 20 and [at] the eastern end of it on the second floor was Edgerton’s lab and the acoustics lab was all the way at the eastern end of the building. So we were just one flight of stairs away from each other in those days, so we used to have lunch together with the folks at the acoustics lab. But he, that is Edgerton, didn’t have any connection with the acoustics lab other than being familiar with Dick [Richard] Bolt, who was the director of the lab and Leo Beranek, who was the associate director.
So when you started graduate school did you work with Bolt and Beranek?
I worked with Beranek immediately. He was my advisor; and [I] knew Dick Bolt, of course, but never was supervised by him. Beranek gave me my project for my Master’s thesis and was my advisor all the way through my degree. One of the things that I got involved with with Leo Beranek was his consulting business. He and Dick Bolt had started a firm called Bolt and Beranek. [Robert] Newman hadn’t shown up yet. I think I was about the fourth employee of BB. There was Sam Labate, Bill [William] Lang and Jordan [J.] Baruch probably preceded me and then probably after me came Bob Newman.
It was an interesting business for a young graduate student. One trip that I remember particularly – being among the first, maybe the first trip that I took as an employee of this new fledgling consulting firm – was to what was called the 8 x 6 Wind Tunnel [interviewee’s note: at NACA, the National Advisory Committee for Aeronautics] in Cleveland, Ohio. They had turned it on once. It looked like a giant horn – [the horn of a] trumpet – with its bell pointed directly at downtown Cleveland. So that when they turned it on at midnight one Friday night – the first time – they woke up all of Cleveland with the tremendous racket that the wind tunnel – it was a super sonic wind tunnel – and so it really rattled the whole community. The judge issued an order that they couldn’t turn it on again until they quieted it down. Leo Beranek got that reversed so that they could turn it on once for measurements and then not again until it had been quieted down. And so, making measurements out there at the 8 x 6 Wind Tunnel was quite an exciting business.
I looked at Leo Beranek’s oral history on the I-triple E [IEEE – the Institute of Electrical and Electronics Engineers] web site and noticed that he didn’t mention his disappointments at all. And maybe I’ll follow that lead. (Laughs)
The 8 by 6, what does that refer to?
It had an eight foot by six foot cross section.
[An] eight foot by six foot cross section of air blowing through the neck of this tunnel at super sonic speeds, as you can imagine, is quite noisy.
What was the power source? Electric motor?
An electric motor, yes. The reason they turned it on in the middle of the night is they couldn’t get permission from the power company to operate it during the daytime when the power demand was high.
The next question, I think you’ve covered most of it. Do you have anything you want to add about working for BB [Bolt and Beranek], whenever N [Newman] came in?
Most of the work I did was involved in estimating aircraft noise. I didn’t do any architectural acoustics although the firm did lots of that. It was an era during which they were building bigger and bigger jet engines and so there was considerable concern about the effect on the public. So the test facilities that were being built subsequent to N-A-C-A [ed. note: National Advisory Committee for Aeronautics], which later the name was changed to N-A-S-A [ed. note: National Aeronautics and Space Administration], was building around the country – had to be either way out in the sticks where no one would mind or it had to be properly quieted to avoid problems with public protest. So my job was estimating how much noise yet un-built jet engines and jet test facilities were going to make.
With your electrical engineering background at MIT, had you specifically study acoustics in depth?
Yes, well I had not studied acoustics at all until I got to the acoustics lab, but then took several courses in acoustics. Leo Beranek taught a graduate course in acoustics and then I taught an undergraduate course – course 6.35 called “Acoustics” – that had originally been taught by Leo, but the consulting firm was taking more and more of his time and so, I guess in good MIT fashion, it was passed off to one of his graduate students – me. And I taught that until I graduated with my doctorate in 1955 [interviewee’s note: in 1954], and then it was handed off to another graduate student. I’m not sure whether Amar Bose was the very next one or the next one after the next one, but Bose did teach it and figured out what that course was good for and made a life’s work out of it.
But you never designed speakers or anything for acoustics?
No. I guess I have a tin ear. I couldn’t tell the difference between a good one and a bad one.
During this time you also consulted some time for Liberty Mutual Insurance Company?
Yes, after we’d been married for a little over a year, it became clear that we were going to have a blessed event. And in those days it was pretty normal for the mother to stay home with the baby. And so Bobby’s job with Doc Edgerton as secretary would have to come to an end. And Leo, being sympathetic to my need for a little additional support beyond the graduate stipend, hooked me up with a guy by the name of Chuck [Charles R.] Williams, who was in charge of research at Liberty Mutual in the Loss Prevention Department.
Liberty Mutual was concerned at that time about hearing loss from industrial noise, since their major business was workman’s compensation coverage. So he, that is Chuck Williams, needed someone who knew about acoustics and noise. Leo was probably the world’s authority at that point on industrial noise. And so I started working half-time for Liberty Mutual. I did my thesis research in the morning, but in the afternoon [went] to Liberty Mutual and then for a short time at night went to Bolt, Beranek, and Newman and worked there at night. So between the three jobs I had enough to support us.
While I was at Liberty Mutual I did – with a gentleman from the Industrial Hygiene Section of the State Government in Maine – a study of hearing loss in a cotton textile weaving shed. And that turned out to be the first longitudinal study of industrial hearing loss that we published in about 1952, I guess – something like that [interviewee’s note: publication was in 1953]. Liberty Mutual was scared to death of the potential loss that they could suffer from hearing loss claims and had set aside a reserve of about three billion dollars. They asked me to look at their calculations and they’d made a mistake on the [value of the] threshold [of hearing]. And so it was clear that the reserve was excessive so that was probably the day that I made the biggest transaction in somebody’s balance sheet in my life. They cut the reserve back to about a billion dollars.
How long a period of time did the study take?
It was two years.
That was primarily looking at —
So we did audiograms of people before they came to work. They had not been exposed and then later. And showed that – the office staff that had been hired showed no significant hearing loss and those who were in weaving sheds showed significant hearing loss.
Interesting. You received your doctorate in electrical engineering in ’54.
Also from MIT. And what was your thesis?
My thesis was on the reciprocity calibration of microphones. It was a topic that I’d selected, prompted by my curiosity of why the reciprocity calibration procedure worked. And so I dug into that and did both experimental and theoretical work. It’s a topic that still intrigues me after all of these years. I answered most of the questions but there are still some that are extremely puzzling and I think there’s more to be done there, but somebody else will have to do it.
Do you have a specific question you want answered?
Yes. It’s pretty clear in lossless systems – [Leslie L.] Foldy and [Henry] Primakoff showed why that worked back in the ‘30s, I think [interviewee’s note: the actual publication was in 1945]. But no one has given a satisfactory explanation for why it works in lossy systems. [It] involves – I’m stumbling on the theorem, and can’t remember it. There’s a theorem that is disjoined from reciprocity calibrations that applies [interviewee’s note: the fluctuation-dissipation theorem is related], but how that all ties into reciprocity calibration of lossy systems I don’t think anybody has ever given a satisfactory explanation.
How did you actually decide to focus on acoustics as your area of specialty?
Because that’s where the job was.
Did you have any mentors?
Well, there were folks who had big influences on my life. Particularly Doc Edgerton who got me into acoustics and was a wonderful man. He had a big influence on both my life and my wife’s. Sam [Samuel J.] Mason – who was my best man at my marriage – of Mason’s Law and Feedback Theory. Leo Beranek – who was my advisor was certainly an important mentor. During those graduate days Phil [Philip M.] Morse, who was teaching Methods of Theoretical Physics, a course that was very important to me in my dissertation. Walter Rosenblith – who had [an] office next to me in the acoustics lab and then went on to be Provost of MIT. J.C.R. Licklider who had an office just down the hall a bit and [was] known as “Lick” [was] an important figure in the development of computers and communication. So it was a time of a number of very interesting people around in those days that I think were very important to me.
You mention Philip Morse – do you have any special remembrance of him?
Well, in the course Methods of Theoretical Physics he was writing a book that became this two-volume bible that I’m sure you’re familiar with.
I remember it well.
And so at that time it was mimeographed notes and he would assign some problems which he was making up as he went along. And they were terribly hard problems so that the students in the course broke up into two or three groups and they would work together on the problems. Spending almost the entire week doing the problems for the course and we were fortunate that there wasn’t really other things that we seemed to have to do. It was an intense semester getting the problems that Phil Morse assigned to us done each week. Even if all of us were working together on them it was tough.
It was good to know you were working together. (Laughs) Very hard problems. (Aside) Back to the script here. Shortly thereafter you accepted a position at the Central Institute for the Deaf in 1955 as a research associate and held a dual appointment with the main campus as a professor [interviewee’s note: as an assistant professor] of electric engineering at Washington University. How did you come to accept the position at CID? You were affiliated with CID from 1955 to 1964.
Right. I got my degree in January of fifty-four, and went to work for Liberty Mutual full time. I was asked to start the research lab in Hopkinton, Massachusetts. A concrete block building had been erected in Hopkinton, which is west of Boston – just a marathon distance, since that’s where the Boston marathon starts. So there were three of us in this large concrete block building. I was the director and had just gotten my degree and I was a fairly young director of this research laboratory. There was my secretary [interviewee’s note: Gladys] and Murph – Murphy – who was a machinist. So that was the staff to start. And it gradually grew in size. But one day that year, 1954, Hal [Hallowell] Davis and his wife, Florence, showed up at our doorstep at our residence in Waltham, Massachusetts – and I’d become acquainted with Hal on an American Standards Association committee that he and I were on as members and was extremely impressed by Hal Davis. So when he came to the door and asked me if I would be interested in a position at CID, I hardly thought twice about it and decided to take it. So I said that I was going to resign from Liberty Mutual [and] suggested that Chuck Williams consider a good friend of mine as the replacement director, Al [Allen] Cudworth, and he took the job and made the research lab into something that really was important to Liberty Mutual and made a number of developments including the Boston Arm and other things that were significant for Liberty Mutual. And so in September of 1955 we came out to Central Institute. I learned that there was a position as an assistant professor in the Electrical Engineering Department at Washington U [University] that came along with the appointment at Central Institute. I don’t think I really realized that – although I’d probably been told it – until I got here and was asked to teach in the electrical engineering department.
The first project you worked on at CID was the construction of a mobile trailer designed to be used aboard a Navy aircraft carrier in order to conduct research on jet plane noise. This project you worked with Hallowell Davis. What were the results of the study and what was it like to work with Hal Davis?
Well the trailer itself was the project of Art Niemoeller, who was my first graduate student. And my project was designing and making some of the equipment that went into the trailer, including an automated audiometer – a device to measure hearing. That equipment and the trailer were sent to Guantanamo Bay and loaded on board the [USS] Forestall, an aircraft carrier down there. Art Niemoeller and I made a trip down there to make measurements on board the Forestall and we had an adventuresome trip via Cubana Airlines from Havana to Guantanamo City. They confiscated our bags because they couldn’t figure out what the soldering guns that we had in our equipment bags were and thought they were some kind of ray gun. But we eventually got them and got on board the Forestall and went out for a cruise [and] measured the noise that the flight deck crew was exposed to with jets taking off. And I had to run around with a microphone on the flight deck as the jets took off, which was kind of exciting – especially at night. The whole project was supervised by Hal and the objective of the project was to determine the auditory and non-auditory effects of high-intensity noise. And if you work out the first letters of that title it comes out ANEHIN. So that was the ANEHIN Project and the results were that the non-auditory effects were unable to be measured. There had been rumors that it made people a little nutty – the high-intensity noise. It made people sterile, it made people impotent, all kinds of things none of which proved to be the case. However, it was clear that it was making people deaf. And so subsequent to that, ear protection became mandatory and it’s amazing [that] before that ear protection was viewed as sissy stuff.
So that was the result of the project. Hal was a very unusual blend of amazingly serious scientist and a risk taking adventurer in other moments.
Have any anecdotes about the adventurer?
Well, with Hal and Dick [S. Richard] Silverman – led us to a brothel in Guantanamo City, which was during the daylight hours and the ladies in the brothel were cleaning up the place and it all looked very spic-and-span. Dick and Hal led us there essentially on a public health mission – the local public health officer of Cuba was proud of showing how hygienic their brothel was. So Hal and Dick were pleased to have a look at all that.
That’s very interesting. Did they measure the noise levels?
I don’t remember that we made any noise measurements there. (Laughs)
In 1956 you designed the acoustics of the Khorassan Room at the Chase Hotel. What can you tell about the project? Another place you also acted as acoustical consultant was during the construction of the Becker Library that we’re in now. You recommended carpet instead of the proposed granite flooring in the main lobby. Do you remember any other details about any of these projects?
Well, I certainly remember the Koplar/Chase Hotel story. Harold Koplar was a frustrated architect who had a hotel and residence to manage and he wanted to build a ballroom. And the only place that would fit was sort of in the well that was surrounded by hotel rooms. He wanted it to be a couple stories high, but that would — If it were a flat roof it would obstruct the view from the second floor hotel rooms and they’d lose all those hotel rooms. So he decided to make it a hangar shape so that it would only be one story high at its perimeter and preserve the windows on the second story.
They asked me to – I was doing some acoustical consulting on the side in those days – and he asked me to advise him on the construction of this hangar-shaped ballroom and I told him that it was acoustic death to build a hangar-shaped ballroom. He said he wanted to do it anyway and what could I do? I said, “Well, we can make it as dead as possible so that no sound can last in the room.” And he said, “Okay, let’s do that.” I said, “Well, if we do that you have to promise not to have any serious music ever, because it will be inappropriate to have a symphony in there for example.” And he said okay. So we built it with thirty thousand beer cans in the ceiling. Beer cans were opened at the top and the bottom as they were installed in the concrete ceiling to provide a pathway for sound to go up through the neck of the can to a fiberglass layer several inches thick above the concrete ceiling. They acted as acoustic resonators – broadly tuned in the range from about 250 to about 750 hertz. Above 750 it was fairly easy to absorb sound with conventional acoustic tile. So it came out to be deader than a doornail as predicted, and none of the offending sound reflections that would have ordinarily occurred for a hangar-shaped building were present. Unfortunately, the first affair that Koplar had was the St. Louis Symphony on Sunday afternoon. It got panned so unmercifully in the papers he quickly replaced it with “Wrestling at the Chase” on Sunday and that was the way it went from then on.
Then it could be said that you were the origin of the famous “Wrestling at the Chase?”
I guess. (Laughs)
And about the Becker Library, do you remember anything?
Well, I didn’t remember that until I saw that question in your list of questions. And I don’t think I did a serious consulting job there. I was asked perhaps by Mark Frisse [ed. note: director of the library at the time] my opinion of the granite floor and I said, “That’s going to be very noisy and inappropriate for a library. I suggest a rug.” And I think that was about it, but maybe I was more professional than that.
In Helen Lane’s history of CID it is reported that in 1959 you had developed a new interest in data processing and the use of electro-computer techniques for experimental purposes and that you had seven full-time and six half-time graduate students working as your research assistants. Was this the start of your shift from electro acoustics to computing technology? And what were you working on?
Well, yes that began in 1959 with a question from Hal Davis. And the question was, “How would you advise me to build an instrument to measure the hearing of infants by the evoked response method?” Some work in evoked response had been done in England in adults. And Hal saw that it had the potential for measuring the hearing of infants who are obviously not able to speak and therefore can’t say when they hear a sound or not. [The] evoked response method involves the application of electrodes to the scalp, measuring the electroencephalogram – a sound or click is presented repeatedly to the subject and the EEG measured in response to that click. By averaging a number of responses together those portions of the EEG that are unrelated to the click average out and anything that is related to the click will be enhanced. So I tried a number of analog schemes which was what had been done in England and had abysmal failures. Hal was, I think, fairly impatient with me because I wasn’t able to reproduce the analog equipment that had been developed in England. I finally decided that digital computer approach to the problem was the right thing to do and having had no experience with digital computers it took me a couple of years to build a special purpose computer to do the job. And that was assisted by [A.] Maynard Engebretson, who was my graduate student at the time. And he built a computer for his master’s thesis to average evoked responses. Hal used it for many years – at least a couple decades, maybe more. And there were a number of copies made that went to laboratories around the country that also used it for evoked response work. So, that was the beginning of evoked response measurement of hearing of infants and I feel that Hal made an important contribution since that’s an accepted method – the definitive method – of measuring the hearing of infants these days.
Can you talk a little bit about the computer. What type of technology, was it all discrete components or was it —?
There was a small firm in Massachusetts – that Maynard and I had seen ads in engineering journals for – operating in Maynard, Massachusetts. And they had modules – digital modules – so a card about three inches by five inches would be a single flip-flop, for example. So we put those together into a box about the size of a washing machine and it could do averaging and histograms. In fact, the name HAVOC was “histogram averaging ogive calculator.” [interviewee’s note: HAVOC was “histogram average ogive calculator”] Ogive is an integrated histogram.
So that computer, when it was replicated, was it replicated using the same technology?
Yes. Yes. Then subsequently, or maybe concurrently, other people were building digital averagers. In fact, I learned later that Wes [Wesley A.] Clark had built a digital averager at Walter Rosenblith’s lab at MIT, which preceded the HAVOC, but at the time we built the HAVOC, I was not aware of that. Then a number of other people built commercial averagers. CAT was one of them – standing for Computer of Averaged Transients. And so that was one of those things that the time was right and the same idea flourished in a number of different places.
Were there analog to digital converter modules available at that time or did you have to build those up as well?
We had to build them up.
Can you tell us about your work at CID? Did you have any interactions with other divisions of CID – the school, the clinic, professional education?
It was mostly research, but I did some classroom acoustics and I taught a couple courses – a handful of courses for the professional program, what is now known as PACS, Program in Audiology and Communication Sciences.
Among the people you worked with at CID were Hal Davis, Ira [J.] Hirsh, Shirley [K.] Hirsh, Jim [James D.] Miller, Maynard Engebretson, Don [Donald H.] Eldredge, Dick Silverman and others. Do you have any stories to share about these folks?
Well, let’s see. Maynard Engebretson is still a good friend and wonderful guy – [a] doctoral student of mine. And in the days of the Biomedical Computer Lab Maynard worked for me. And I had given him a very dirty job of trying to convert something called a Language Master to use with data and I had renamed it the Data Master. It was very unreliable and not one of my prouder moments at all. It seemed like a good idea at the time, but it never really worked and unfortunately Maynard was saddled with that. And he took to carrying around in his toolbox chicken bones so that whenever the Data Master was acting up he would wave the chicken bones over it to try to make it perform a little bit better. We got LINCs at Central Institute and Don Eldredge got a LINC. A LINC is a computer that Wes Clark designed and I’ll probably say some more about that later and how we got into that. But Don had a LINC in his lab and he complained that the LINC eliminated many, many hours of happy scut work that he used to be able to do – copying numbers from his lab book to get his data results and to making calculations on those results. So he no longer had to do that and this made his life harder because he had to spend more time thinking then doing happy scut work and so he complained that he wasn’t sure the LINC made real progress.
Ira is a good friend – Ira Hirsh – who I’ve known over these many years, and aspire to retire as many times as Ira has retired from various jobs. He was head of research at CID. He was head of psychology at Washington U. He was dean of the School of Arts and Sciences. He was head of psychology again. And he was director of Central Institute [for the Deaf] – all of which he retired from – so that’s a pretty good record.
Dick Silverman was a wonderful director of CID. I don’t think we’ve seen his like since then. He built it up to a world-renowned organization. [He] had the habit of having lunch in the lunchroom in the research and clinics building and all us research folks would sit around the table and Dick would sit at the head and preside. His wife made him lunch every day and was convinced that he never really knew what he ate and so to test that theory she made him a bread sandwich – two slices of bread with a slice of bread in between and he never commented on it.
Let’s see. Who else did you ask about?
Well, I mentioned Don and the LINC. And Hal [Hallowell Davis]. Hal was inspiring in a couple of ways. He was inspiring as a mentor and he was awe-inspiring in his ways of thinking about physiology problems. Awe-inspiring in his wide range of interests and his ability to concentrate and persevere.
What about Jim Miller?
The thing I remember about Jim Miller was that he decided to give up smoking and the only way he could do it was to go to movies continuously because you couldn’t smoke in theaters and so he would see one movie after another in order to break the habit.
Was he successful?
And Shirley Hirsh?
Shirley was a lovely lady. She has unfortunately passed away a few years ago, but she was Hal’s assistant on developing the field of infant evoked response. I think [she] was, and Hal would agree, was an important contributor to that development.
As of 1993 you’ve been a member of the Board of Managers for CID. What does that involve?
Well, CID has a very large board and so it gets its work done by committees. I’m on the Development Committee and the Program Committee as of now. At the time of CID’s difficulties, I felt my job was to supplement communication between the factions that were at war during the late ‘90s. Fortunately, Bob [Robert G.] Clark, the head of the board, stepped in and resolved the problems for CID in what seemed to me at the time to be a miraculous way. And Bob gets an awful lot of credit for seeing CID through a very troubled time.
In 1964 you founded the Biomedical Computer Lab (BCL) and served as director until 1975. Can you tell us how this came about and what was its significance over some of the early challenges of establishing the new programs? What was your biggest challenge as an administrator?
Well, I was on the NIH computer study section essentially when it was formed. One of the members of that study section, Lee Lusted (spells) L-U-S-T-E-D —
Is he related to Pat Lusted?
I don’t know. He was from Rochester, I think.
There was a fellow over in Psychiatry, a professor of Psychiatry, of that last name.
Maybe. In any case, Lee was on the computer study section and one day he said to me [that] I ought to consider setting up a grant for a laboratory devoted to biomedical computing here at Washington University. And I hadn’t really thought about that, but after giving it careful consideration and talking to many friends on the committee and folks who I felt would give me good advice, including Mary – I can’t remember the name. Maybe it’ll come to me. [interviewee’s note: Mary A. B. Brazier] [I] talked to a number of people. I wrote a proposal and to my very great surprise it was awarded and I left CID to start the lab across the street here at 700 S. Euclid. A couple of good stories there – Ed [Edward W.] Dempsey was dean of the medical school and so I had to get his support for such a lab in order to be able to submit the proposal through Washington U. So I put together a presentation and went over to Ed’s office. I found that the date for the presentation to Ed was delayed quite a bit and I didn’t understand why it was taking so long but eventually it happened and I went over there. And there was a man sitting in the corner who I didn’t know and who Dempsey didn’t introduce, who just sat there as I made my presentation. And after I was all done Dempsey said, “Okay, I’ll let you know” and then a little while later he said, “Yes, it sounds fine.” And it was only perhaps a year later that I found out that the man who was sitting in the corner was Fred [Frederick L.] Stone who was the head of the newly-formed Division of Research Resources [ed. note: part of the National Institutes of Health, now known as National Center for Research Resources], which funded the grant.
Let’s see, what else did you ask me there?
When you came, the lab was just an idea. What were the early challenges, really starting from nothing? There was no model or anything.
Well, space was a problem. Where were we going to put the lab? And when I talked to Dempsey, he said there was no space on campus – [the] medical campus. So I tried the streets around here and found a garage at the corner of Euclid and – [I] can’t remember the name of the street – Laclede, Euclid and Laclede, that was vacant and could be renovated to do the lab. So plans were moving forward to do that when Shriners [Shriners Hospital] allowed Washington U. to buy the Shriners’ nurses’ quarters in Shriners Hospital. And it was obvious to me that having something as close to the medical center as that would be much superior. So we got out of our obligations to the garage down on Euclid and decided to move into 700 S. Euclid. It had been left unheated over the winter of 1963 and all the paint fell off the walls in sheets as a result of that oversight. But we completely renovated the first floor and left the basement and the second floor in the shambles that they were. I guess my personal difficulties had to do with the transition from being one of the guys to the boss and that was a tough transition for me, but apparently it worked okay.
How did you gather staff?
I was teaching graduate courses and so I taught a course in digital logic and developed that course. It was originally 560, I think, as a graduate course and it gradually worked its way down and it’s now 260. So it’s gone all the way from a graduate course to a sophomore course, but it’s still pretty much the same course as I started it in about 1960. So I would get students in that graduate course and invite them to come work at the biomedical computer lab. And of course since we had funding that was a fairly easy sell.
About the same time the computer research laboratory was established – how does this tie in with BCL and who were the staff members involved in CRL [Computer Research Laboratory]? CRL was involved with a LINC laboratory. How did George Pake persuade [William N.] Papian, [Wesley] Clark and [Charles E.] Molnar to relocate here from MIT? So just that whole general topic.
At CID I’d received a grant for a building a computer for Hal Davis and Don Eldredge for their physiology lab. Again to my surprise, it was granted and I needed to figure out how to do it. We’d had success with the DEC [interviewee’s note: Digital Equipment Corporation] modules, so I decided to see what they had available for building a computer for the physiology lab. I took a trip to Maynard, Massachusetts and walked in unannounced and asked if I could get some technical questions answered and they said our president is available. So I spent pretty much of the morning with Ken [Kenneth] Olsen, at a time when Digital Equipment [Corporation] was still a small company. I told him what I wanted to do and he said, “Well, you really ought to meet Wes Clark at Lincoln Laboratory.” So that afternoon I went over to Wes Clark’s operation at Lincoln Lab and saw the prototype LINC and realized that there was absolutely no point in trying to build something that was different from the LINC. Wes was clearly well ahead of my thoughts and had a successful prototype already that did everything that I thought about doing and more. So I joined forces with Wes Clark and his folks. Actually Bill Papian was his boss, but Wes was the computer designer. Bill had built the first ferrite-core memory back when he worked at Project Whirlwind. And so Papian, Clark, and Molnar were the drivers there of that project. And a good friendship developed.
Wes was, he claims, fired for insubordination from Lincoln Labs. I think the real story is that he insisted that applications of computers to medical research was an important topic and that Lincoln Lab should work in that area and the management of Lincoln Lab declined to do that. So he and Papian started a project at MIT with promises from MIT of faculty positions for a number of the folks. They got a thirty-one million dollar grant in the early ’60s when thirty-one million was quite a bit of money for a New England Center for Computer Technology and Medicine, but MIT reneged on some of their promises and the staff became very unhappy. Because I was going back and forth to Boston and Cambridge a lot, I was aware of the unhappiness and suggested to George Pake, who was provost at the time, that he might have a profitable trip if he stopped off at MIT. And he did and Papian and Clark and others were impressed by George Pake.
I think one of the things that impressed them was that he told them that they could kick his tires. He didn’t seem to want to test them at all, but rather wanted them to test him. So he was able to convince a large fraction of the folks that were then in what was called the Center Development Office at MIT to come to St. Louis. They arrived here in, I think it was, July 1964.
Molnar stayed behind because he hadn’t finished his doctorate yet and he came roughly about a year later – after he finished that and went directly went from his ScD to Associate Professor in the Physiology Department. And Wes Clark and Bill Papian and Tom [Thomas T.] Sandel all got departmental appointments. Tom Sandel eventually rose to be head of the department of Psychology. So it was a big and productive move for Washington U.
I note that we’re an hour and a half in, and we’re on page two.
[ed note: Dr. Cox and Interviewer discuss length of interview and whether to take a break]
What were some of the BCL’s early projects? Didn’t some of them include computer imaging systems for cardiac purposes, nuclear medicine [and] radiation treatment planning?
Well, radiation treatment planning was the first big project. And it originated from a course that Wes Clark and I taught on computer design. Wes asserted the strong position that the only way to teach a computer design course was to design a computer. I’d never thought of that. It seemed a little absurd to me, but he said, “Well, we could do it” and we did. We developed something called the FW2 [interviewee’s note: the project was actually called “4W2”], which stood for the Four Week Wonder. It was a time when TW3, “That Was The Week That Was,” was a popular television program so that’s where the FW2 [4W2] came from.
Actually one team finished in four weeks and the other team finished in five weeks. They both began with Fs so FW2 works out alright. The computers worked and one of the students in the class, Bill [V. William] Gerth, wrote a radiation treatment planning program, which was just a very trivial version of radiation treatment planning. But when I showed it to Bruce [D.] Waxman who was the program officer at the Division of Research and Resources he was excited about it and said, “Let’s make a special program.” He managed to push some funding on me to build a number of copies and develop the software. Bill [William E.] Powers, who was the director of radiation therapy at the time in the basement of Mallinckrodt, led the development of specifications. And Bill [William F.] Holmes wrote a good bit of the software. There was a team that did a lot of the work, but Bill was the important guy in that.
So we got radiation treatment planning software system working, then our company that was building the computer was acquired by Becton, Dickinson [and Company]. The company’s name was Spear, (spells) S-P-E-A-R. And Becton, Dickinson said they didn’t want to have anything to do with this so I was running a big project that depended on having a number of copies of the computer that we designed and then nobody to build it. So a small startup company in St. Louis called Artronix volunteered to build it after some negotiation; they won the contract to build what we called the program console – which had the initial “PC” – later appropriated by IBM. We built about a dozen copies of it and distributed them to six or eight universities around the country. That development by Artronix eventually transitioned into another company called CMS, which is still in business building radiation treatment planning systems and has over the years had a relationship with Washington U.
The Spear PCs, were they ever used here or did any of those actually get built?
We had a prototype, I believe, that Spear built, but the Artronix one was the one that was built in decent numbers.
Did you ever have any discussions with IBM about compensation for stealing the name?
No. (Laughs) No.
Do you have any anecdotes about Arne Rostel [ed. note: the president of Artronix] or any of those people who were involved in the early days of Artronix?
Well, they were fearless and they went on from the success with — No, I’m wrong. There were a number of Spear PCs built. And it was after they were distributed that they were bought by Becton, Dickinson. Then they didn’t want to go on and that was when Artronix came in so my memory initially was in error there.
Did anybody at Artronix – had Arne ever worked for Washington U. or —?
Oh, wait. Yes, they had built a radiation dose measuring system that Bill Powers was using and, I guess, [was used in] a number of radiation oncology departments around the country. So they were familiar with the radiation therapy business and so jumped into radiation treatment planning fairly readily. They’d never built a computer before and it’s much to their credit that they were able to make it happen. I lent them Bill Gerth for several months to help them get started building computers. Bill Gerth was the assistant director of the lab. So it was kind of amazing that they were able to pull it off and get into the computer business. I indicated they were kind of fearless because the next thing they went into was [CT]. They saw computed tomography as a growing business and built a CT scanner with the help of Dick Zacher from the Physics Department, but financial woes and I don’t know what all beset them and they never really made a success out of that. You may know more about that then I do.
Somewhat, but they definitely didn’t have success. Did you have any dealings with Michel Ter-Pogossian?
Yes. Well, let’s see – let’s talk about PET.
Mike [Michael E.] Phelps and I had gotten together at Mike Ter-Pogossian’s urging in connection with making measurements of radioactive isotopes. So, we enjoyed each other and got into the habit of meeting Friday afternoons to just brainstorm about things that we were doing, [things] going on in radiation physics and computing. And one Friday afternoon Mike Phelps said he thought it was possible to build a tomographic system based on positrons and he wondered if the reconstruction method that we used for x-ray tomography – transmission tomography – would work for emission tomography. And Mike and I went to the blackboard and fiddled around for a bit and decided that “Yeah, that really ought to work.” We saw how to apply the reconstruction algorithm to emission tomography.
Mike then went to work to do an experiment to verify that what we’d calculated ought to work. And working at night he was able to do that. He showed the results to Ter-Pogossian who thought it was nonsense and didn’t want Mike to spend any time working on it. Mike Phelps – not to be deterred – he continued to work on it at night and on weekends – around the clock. [He] always was a person with enormous energy. And [he] finally got results which were incontrovertible. Again [he] showed them to Ter-Pogossian and Ter-Pogossian, that time, recognized that this was an important development and to his credit turned the lab around and focused all its energy on developing a positron emission tomograph. [The] two Mikes – Ter-Pogossian and Phelps – were both ambitious characters and so I think it was probably inevitable that trouble about priority and who’s responsible for the development of the PET would develop and it did and Mike [Phelps] eventually left.
Do you know where he went?
He went originally to [the University of] Pennsylvania and then that lasted I think only about a year and then he went to UCLA where he still is and has thrived and been extremely successful.
In terms of reconstruction techniques, was that before Fourier reconstruction or was that already at Fourier reconstruction?
Yes. [Godfrey Newbold] Hounsfield, when he originally developed CT scanning used arithmetic reconstruction methods. Don [Donald L.] Snyder and I were fascinated by the original transmission tomograph and said, “Gee, that’s a clunky way to do it.” There ought to be a Fourier way to do it and we, I think, contemporaneously, with a number of other people, developed the Fourier algorithms for reconstruction. We also identified some ways to sample the data – that is Snyder and I – and applied for a patent on the data sampling method. We could hardly apply for the patent on the algorithm and wouldn’t have been successful anyway because others were doing it concurrently. But we did get a patent on the data sampling method and that led to a contract with Picker, a radiology equipment company, to do research on computed tomography and led to a number of other patents – both with members of the lab, Bill Gerth, and with members of the staff at Picker. Picker never used the patent and we got no royalty – even though Picker cross-licensed it to GE who did use the patent. But since it was a cross-license and no money was exchanged, Picker didn’t feel obliged to pay us anything which was, of course, disappointing. But universities at that time and maybe still are not in the habit of suing anybody who infringes their patents.
The patent revolved around coincidence measurements?
No. The patent was on basic reconstruction method for computed tomography – focusing on how to sample the data. Near as I can tell, GE used it in their CT scanner.
Interesting. Let me come back to that – it’s an interesting time. Let me move forward. In the middle of all this there was the development of the MUMPS computer language.
I was impressed by MUMPS. I had encountered it at [G.] Octo Barnett’s laboratory at MGH in Boston. I had gotten to be friends with Neil Pappalardo, who was the developer of MUMPS. MUMPS stands for MGH Utility and Multi Programming System, which is a second order acronym because it contains an acronym. That led Joan Zimmerman, who was then working at the lab, to start the MUMPS users’ group [interviewee’s note: the users’ group was called MUG] which I think is one of the early third order acronyms since it contains an acronym of an acronym. I think getting the MUMPS users; group started and publicizing MUMPS was one of the successful things for the Biomedical Computer Laboratory. But the effort was transferred to MCS when Joan left St. Louis and you probably know more about that than I do.
That’s one small area here. You talked a little bit about the old Shriners Hospital facilities and so forth – the paint falling off and so forth. As I recall it was many years before the basement was actually renovated, and some of the floors still are probably being renovated today. Do you have any stories?
We gradually grew in the number of employees and the number of labs. BCL occupied the first floor. When CSL came… We started BCL on April 15th, 1964. Then the group from MIT came in July of ’64 – so we renovated the second floor to house them – the second floor of the nurses’ quarters. BCL grew with the radiation treatment planning project. CSL grew with – what was then CRL – grew with the macromodular project, which I’ll probably say some more about in a bit. So we needed more space and the Shriners Hospital space was available and so it was renovated. CSL moved out of the second floor. BCL moved into the second floor and gradually more and more of the Shriners Hospital space was renovated to contain projects that were being developed in both BCL and what later became CSL.
Maybe I should mention a few words about the transition from the Computer Research Laboratory to the Computer Systems Laboratory. Bill Papian was director of the group at MIT – both at Lincoln lab and at the MIT itself. And so it was natural that he became director of the Computer Research Laboratory. Wes Clark was the genius behind computer system design. And Wes and Bill fell into conflict because Wes felt that Bill shouldn’t be making commitments for him when he was the one who was developing the ideas. So the resolution of that problem was for Bill to start a laboratory on integrated circuit design that would be in the Engineering School and Wes would run a new laboratory – the Computer Systems Laboratory – here in the Shriners Hospital. The Computer Systems Laboratory had NIH support for continuing to augment the LINC and support the users, but most of its money came from ARPA, then later DARPA, for the macromodular project which was started by Wes and funded by Ivan [E.] Sutherland. And that was the first asynchronous or clock-less computing system of any size that was ever built and still is a landmark in that regard.
Ivan Sutherland is still working in that area for Sun [Microsystems], and Charlie Molnar who made the project happen after Wes left Washington U, is a leader in the field and has left an important legacy. Charlie passed away a half dozen years ago I guess.
I guess a couple of specific things in that realm if I could ask you – the big development there was the development of the Glitch Theory.
And I guess I know a little bit of the story – could you tell us a little bit of the story of how that came about and also what involvement you had in that.
Well, I had very little involvement in the original Glitch business, but it’s an interesting story. Severo [M.] Ornstein and Michelle [J.] Stucki were developing the macromodular specifications and I think it was probably Michelle Stucki [interviewee’s note: “In retrospect, I think it was Severo Ornstein and not Stucki.”] who said, “What happens if two asynchronous events happen to occur at just the same time, particularly if they are leading to two different outcomes?” And none of the staff had ever thought much about that before, but Charlie Molnar and Don [Donald F.] Wann, Fred [Frederick U.] Rosenberger, and Tom Chaney all thought hard about that. It was not a topic that was in the textbooks. Fortunately Tom Chaney did experiments which indicated that the phenomenon of what came to be called meta-stability in flip-flops was real and could be demonstrated. Charlie and others wrote a paper which they submitted for publication, but was rejected because two out of three reviews said, “Don’t publish it.” One review said, “Don’t publish it because it’s old stuff.” And the other review said, “Don’t publish it because it’s impossible.” But they eventually did get it published and I think Tom Chaney’s measurements established the field of computer glitches – what became sometimes known as the St. Louis Glitch – and meta-stability is now in almost all textbooks and is really pretty well understood.
Ivan Sutherland said that the whole macromodular project was worthwhile because of the development of the understanding of the glitch phenomenon.
Another important development of that was the electric baseball bat that sat in the CSL trophy case for a long time.
Oh, I don’t know about that.
There was a baseball bat with an electric cord attached to it that apparently somehow grew out of the macromodular project. Another thing – that group was a part of ARPA. Did they have any connection or did BCL have any connection with the Internet early on? Did Al Gore ever come to visit us?
There’s a book, Where Wizards Stay Up Late [ed. note: Where Wizards Stay Up Late: The Origins of the Internet by Katie Hafner and Matthew Lyon], that recounts the history of the Internet. And that and another one on Licklider – I’m trying to remember the name of the book, but I don’t remember it [interviewee’s note: The Dream Machine: J.C.R. Licklider and the Revolution That Made Computing Personal by M. Mitchell Waldrop] – recount an episode in which Licklider decided that what became the ARPANET should be built and they were trying to understand how to do it. The ARPA contractors had periodic meetings and at one of them they were discussing the architecture for this network that they thought they ought to build – and were stumped. And the story has it that I think on the way to the airport Wes suggested that using a small computer as an intermediary between the network and the big computers that they wanted to connect together would be a wise idea – and essentially was the first router – and so Wes is credited with a breakthrough idea for building the ARPANET. The ironic part is that Wes didn’t want to have anything to do with having the ARPANET here in St. Louis because he thought it would be a distraction.
Interesting. Your work involved collaboration with other departments and programs. How did you foster cooperation? Were some more difficult than others? Who among them was your favorite collaborator? What makes for a good collaboration? Is competition better than collaboration? There’s a lot of stuff there.
Well I’m not going to answer all those questions, but I think collaboration is better than competition and the collaborators were Bill [William E.] Powers in Radiology, Harry Fozzard in Cardiology, Gerry [Gerald] Wolff in Cardiology, Charlie [G. Charles] Oliver in Cardiology, Mike [Michael E.] Phelps [in] Radiology, Burt [Burton E.] Sobel in Cardiology, Gil [R. Gilbert] Jost [in] Radiology, Marc [Marcus E.] Raichle [in] Radiology, David [C.] Van Essen [in] Anatomy, Mark [E.] Frisse [in] Medicine and [the] Library, Jim [E. James] Potchen in Radiology, and Mike [Michael E.] DeBakey [interviewee’s note: at Memorial Hospital in Houston] consulted. [I] consulted with Mike DeBakey on monitoring in his intensive care unit. Maynard [V.] Olson on algorithms and genomics, Sid [Sidney] Goldring was one of the original LINC participants. He used the LINC in neurosurgery. So those were all medical collaborators and they were all great. I think there were probably ones who the chemistry didn’t work right and so nothing developed, but in each of those cases that I named there was a good problem and we worked well together and got some good results.
There are lots of other collaborators who I view as very good friends. In CSL – Charlie Molnar, Wes Clark, Mich Stucki, Severo Ornstein, Tom Chaney, and Fred Rosenberger. And in BCL Don Snyder, Jim [G. James] Blaine, Lew [Lewis J.] Thomas, Ken [Kenneth B.] Larson, Bill Gerth and there are many other BCL stalwarts. All of those – never had a fight with any of them.
That’s a good record. Some of your projects included commercial partners and interests as well. Commercial interests were intertwined with scientific interests during your career. Are there any lessons that you learned? What about intellectual property issues? Will you talk a little bit about that?
Well, I started out in industry at Liberty Mutual. And I guess I was discouraged. Although starting the research lab was great fun – getting into the higher realms of politics at Liberty Mutual was not something I wanted to do. I felt that, as I saw it, what the politicians at Liberty Mutual did was send each other memos and there was very little solving of problems. And so that disappointed me. BBN gave me a quite different view and I found that exciting, but didn’t want to spend my life at BBN.
Artronix was the first exposure here and that was a pretty small company. Mennen-Greatbatch [Electronics] was a collaborator which took the algorithms that we developed for monitoring electrocardiographic rhythms and put them into their product.
Each one of our attempts to do collaboration with industry I felt was flawed and maybe that’s just the nature of things. Each time we tried it we tried a different way and we tried to collaborate with DEC, with Picker, with Positron Corporation. Teramet was a networking scheme that was developed contemporaneously with ethernet and was licensed to CSK – a Japanese corporation – but was unsuccessful. Later on Jon [Jonathan S.] Turner and Guru [M.] Parulkar and I licensed networking equipment to Synoptics and finally when that hadn’t gone well we looked at each other and said we’ve tried all the options. The only thing we haven’t tried is to start a company ourselves and so we started Growth Networks. And that worked out amazingly successfully and it was eventually sold to Cisco. So I think we can’t say that that was the right way to do it. We were just amazingly lucky at that point in time. Other lessons to be learned – intellectual property is very sticky. (Laughs)
Any thoughts you have about directions the university should go in terms of intellectual property?
I think that being fairly generous with intellectual property is the right way to go. That means that you don’t anticipate you make as much money from it directly in license revenues, but you may do more good for the university and for the community as a whole by making the transition of ideas to the public good easier. I think, assuming that it’s a source of revenue – as a profit center for the university – is a bad idea.
I think we may have time for this one more question, then I going to have to change the card [in the digital recorder]. In 1967 the Wash. U. Computer Laboratories was organized and included BCL’s new unit. You were appointed as chairman of the Computer Laboratories, a position you held until 1983. Can you tell us how this came about and what were some of the projects that were involved?
We had two grants from NIH in 1964; one for BCL and one for CSL. The Division of Research and Resources in the person of Bruce Waxman, said, “I want to manage just one grant so you figure out how there will be one principal investigator.” So George Pake volunteered to be principal investigator for the two laboratory grants, but he obviously had his hands full later on in the ‘60s and couldn’t do that any longer and asked me to take over as principal investigator. So that’s how it happened. I took George’s place and BCL ended up doing medical applications of computers, whereas CSL did fundamental computer development. An interesting side note is that at one time or another I had been the boss of two-thirds of Wes Clark’s wives. (Laughs)
Very cute. On that note, I’m going to stop this right now and I’m going to change the cartridge.
[Recording cartridge is changed]
We’re still in the King Center of Washington University on the seventh of June, of 2006. We have just finished talking about the funding. Do you have anything more that you want to talk about the funding of the BCL?
Major sources of funding were primarily NIH. Later on at the Hilltop campus – the Danforth Campus – [we] got both NSF, DARPA funding and various other sources. I got quite a bit of NIH funding during the days of BCL. In fact, it was sufficient to put me in the top five percent of PI’s. I was kind of amazed at that, and as they were doing some project they contacted me and said that I was in the top five per cent and they wanted to know about some history and so forth. So I filled out their questionnaire and sent them a CV. There was no funding from the medical school, except for renovations – taking off the peeling paint and fitting out parts of the nurses’ quarters and the hospital. So it was all soft money.
Do you feel that the independent funding gave you [freedom] in terms of pursuing your research interests?
Well, having as much money as we did was an advantage. I could pick projects to work on and some of them of course did not work out, but many of them did work out reasonably well. Later on the funding got tighter and I guess I perhaps ran out of obvious ideas to follow up on and got the itch to go do something else. I was then following the life plan that said ten years at a place was about the limit and then you should go do something else. So I was at MIT from ’44 to ’54, then there was a year transition and then I was at CID from ’55 to ’64. And then at BCL from ’64 to ’74, and then at the Hilltop at the Computer Science Department from ’75 to (oops!) ’91 – so I overstayed my term there. Well, I’m still there. So how did I get into that? What did you ask me? (Laughs)
Well, how did this come about was the question. In 1975 you were appointed chairman of the Department of Computer Science —
Oh, yes. How did that happen? Well, I’d found myself getting more and more involved in database issues at BCL and thought that the future of Biomedical Computing was going to involve more in the way of databases than I knew about. Also at the same time, the Applied Math and Computer Science Department at Washington University Engineering School was getting a divorce. The applied mathematicians and the computer scientists found they could not co-reside. And Dean [James M.] McKelvey developed the idea of separating it into two departments; the department of Systems Science and Mathematics and the Computer Science Department. John Zaborszky was clearly the candidate to be the head of the Systems Science and Mathematics department, but Computer Science did not have a head. So a Search Committee was formed and I was asked to participate in the Search Committee. Someone, and I can’t remember who, suggested that I should apply for the job. And because I felt that either I was going to have to build a database group at BCL or learn more about databases myself, I decided to consider the possibility of becoming Chair of the Computer Science Department. So I resigned from the Search Committee and was asked to serve as the Chairman and I accepted. I was already a professor of Electrical Engineering and so I was made a professor of Computer Science and that’s how I got to the Computer Science Department.
An interesting story is how I got to be a tenured professor. Back in 1961 I had been asked by Jim McKelvey to be the chair of the search committee which located Bill [William S. C.] Chang as Chairman of the Electrical Engineering Department. And after he had accepted the job, it seemed natural to appoint me as chair of the Tenure Granting Committee for Bill Chang. And someplace along the line, someone must have noticed that I didn’t have tenure. And it was not the right thing to do, to have the chairman of the Tenure Committee not have tenure. So mysteriously, I think with George Pake’s assistance, suddenly some documents appeared and I was granted tenure. I don’t know how it really happened.
Well, things were different in those days.
Do you ever have concern that you didn’t take any courses in computer science, but yet you’re the chairman of the department?
Well no one in the department had taken courses in computer science in 1975. So the department consisted, when I took it over, of Rich [Richard A.] Dammkoehler, who was an industrial engineer, Bill [William E.] Ball who was a chemical engineer, and Sy [Seymour V.] Pollack, who was a chemical engineer, and me who was an electrical engineer. We all picked it up as we went along.
In 1977 you were appointed as a Fellow of the I-triple-E [IEEE] for your contributions to the applications of computers to clinical medicine.
Yes. That was probably a result of an I-triple-E [IEEE] review paper on the electrocardiogram, the electroencephalogram, and the blood pressure wave [and the] monitoring thereof. I wrote that paper with Floyd [M.] Nolle and [Randolph] Martin Arthur and it, I think, was well received. Also, I got elected to the Institute of Medicine about that time – in the early years of the Institute. And I suspect, although I don’t know it’s true, that it was probably a result of a recommendation from Walter Rosenblith who was then Provost at MIT and a friend from the Acoustics Lab. So those probably contributed to the Fellow position at the I-triple-E [IEEE].
Between 1983 and 1984 the Institute of Biomedical Computing was established at the merger of BCL and CSL. Two groups were added – a medical informatics group headed by Mark Frisse and a Center for Molecular Design headed by Garland Marshall. This new IBC was a joint effort between the Medical Campus and the Danforth Campus and the School of Engineering and Applied Science; [it] allowed for faculty appointments at IBC. Dr. Molnar was head of CSL. Dr. [Lewis J.] Thomas was head of BCL. What can you tell us about this organization? During this time you were appointed professor of Biomedicine at the Institute for Biomedical Computing, a position you held until 1998. Do you have any administrative duties with IBC? That’s a long question.
I sat on the Governing Committee of IBC. Charlie Molnar was the Director of the IBC and actually it was his grand vision for the cooperation between the medical school and the School of Engineering. It eventually failed. I think that was because it neither had a charter from the University for teaching, nor did it have sufficient soft money to guarantee its survival. Because it didn’t have a charter to teach, its ability to grant tenure was fragile and eventually failed. CCB [the Center for Computational Biology] now occupies some of the space that the Institute occupied. But previously CCB was only a small part of IBC and then known as CMD [the Center for Molecular Design]. Is there anything else you want to know about that that I might know?
I guess as a structure spanning campuses it once was kind of a fulfillment of a goal that both campuses have had for many years to try to cooperate and so forth. But on the other hand there was that tension between the campuses. I wonder if you could speak to what lessons in terms of future cooperation might —
Well, there are different cultures on the two campuses. And I don’t think IBC was strong enough to bridge those cultures, particularly as a result of the fact that there was not a large source of funding for IBC. And there was not a defined teaching goal. So those are the two, I think, root causes.
But then there are a bunch of other difficulties. How do you manage sharing overhead when faculty is from both campuses? That’s a knotty problem. How do you manage tuition for graduate students – another knotty problem. And I don’t think that either one of those have been solved. I think there’s even more argument for good collaboration across the park [ed. note: Forest Park] between the Engineering School and the Medical School. There already is collaboration between the medical school and Arts & Sciences through the Division [ed. note: the Division of Biology and Biomedical Sciences]. But a similar arrangement has not been accomplished – and I hope will be accomplished – with engineering.
I have a question here about Virginia Bixon. Do you have any remembrance?
Yes. Virginia was the Business Manager of the Biomedical Computer Laboratory and was the mother hen of the laboratory. She really defended the interests of the laboratory heroically and occasionally excessively.
In 1986 you were appointed professor at Mallinckrodt Institute of Radiology, a position you held until 1998. Can you tell us about what research you did there?
Well, that goes back to a conversation that [C.] Frank Starmer, who was then at Duke, now at South Carolina – the Medical College at South Carolina. The conversation that Frank and I had [in] about 1980 when we recognized the potential for digital images in radiology. I think Frank may be still peeved with me that I took that idea and ran with it and he was not able to do that at Duke. I became a staff member of the Electronic Radiology Lab that was headed by Jim Blaine and I think ERL [Electronic Radiology Lab] was a major force in the development of digital images for radiology. Gil Jost supported that lab effectively and enthusiastically and we now see through arrangements that ERL had with Kodak and with others the successful use of digital radiography throughout BJH [Barnes Jewish Hospital] and BJC [BJC Healthcare]. I was in on the beginning before ERL was formed, but Blaine and Jost and many others really carried through that work and so I’m pleased that that worked out as well as it did.
From 1991 to ’95 you served as Director of the Applied Research Laboratory at the Department of Computer Science. Can you tell us a little about that?
ARL was originally conceived as a vehicle for transitioning developments that Jon Turner had made in the design of electronic switches for digital communication – transitioning that technology to industry. We had in mind a full time staff plus graduate research assistants and that lab was successful in transitioning technology that Jon had developed to Synoptics and then to other companies, perhaps not as successfully. But the lab is still going and still being a major contributor to technology in digital networking. It was under the direction of, in addition to my position as director, Guru Parulkar and Jon Turner. Jon is still the director.
In 1997 you received the Washington University Elliot Society Search Award and in 2001 you were awarded the Lifetime Achievement Award from the St. Louis Academy of Science.
That’s right. The award from the Elliot Society was a complete surprise – that’s the tradition of the Search Award. I had no idea that I was going to get the award. When they started reading the citation I said, “Gee, who else at Washington U went to MIT and came here in 1955?” I had no clue it was me, but then gradually it dawned on me. My knee had swollen up as a result of too much running, and so I had been walking on a cane and when I came to the dinner I came in with a cane. But when I finally realized that they were calling on me, I strode up to the dais and forgot my cane and didn’t have a moment’s pain from the knee that had been bothering me all the time. Really quite remarkable. So it was a great surprise and a great joy.
We’ll lead that into some work with psychological effects of medicine. You also were given the Lifetime Achievement Award by the St. Louis Academy of Science.
Yes. I knew about that ahead of time and that was a great honor also, which I was very pleased to get. I served on the Academy of Science Board for a bit, but they tended to have their meetings at four o’clock on Tuesdays, which was the same time as the CID Board meetings and so I felt my greater allegiance was to CID. So I resigned from the Academy of Science Board, but I have a strong feeling for the Academy of Science and think they’re doing a great job.
You’re now a professor of Computer Science in the Department of Computer Science and Engineering. You’ve been with this department since 1975. What do you think of the students of today compared with those of twenty or thirty years ago? Do you recall any memorable students and any names we would recognize?
Well Jon Turner was a student of mine and he certainly was someone whose exceptional gifts were recognizable as an undergraduate. There are a couple [others] who have given me sayings which I use a lot. One of them is Eric Kiebler, to whom I attribute “Kiebler’s Law.” And Kiebler’s Law is the “software isn’t done until the last user is dead.” And Engebretson’s Law – Maynard enunciated this one time. Engebretson’s Law is that “In computer projects ninety per cent of the fun is over when ninety per cent of the work is yet to be done.” So there are lots of memorable students, but those I remember because they have given me great joy in their work and in their sayings. I’ve been with the Department of Computer Science essentially since it started and I’m delighted to see its growth.
For over fifty years you’ve been a leader and innovator in the development of computer technology applications including biomedical computing. What do you regard as your best accomplishments?
Well, I first think of facilitating, under Hal Davis’ leadership, the testing of hearing in infants which is now mandated in most states in the U.S. The interactive radiation treatment planning with Bill Powers, funded by Bruce Waxman, was an important achievement which lots of folks contributed the important stuff to. But I still feel good about it and felt particularly good about it when I got a course of radiation therapy for prostate cancer recently and thought about all those bits banging around in the therapy machine. I think the development of the AZTEC Algorithm by Floyd Nolle, which is now a component of many ECG rhythm monitoring systems, was a good application. I’m proud of the fact that I was a contributor – although Mike Phelps was the real spark plug in the development of PET and Mike was helped by Ed [Edward J.] Hoffman, Don Snyder, and Nizar [A.] Millani in getting that going and of course it was in Ter-Pogossian’s lab. And finally the growth of a productive and happy CS department – one that’s flowered under the leadership of Jon Turner and Catalin [Gruia-Catalin] Roman is something that is very pleasing to me.
Did you foresee any advancements in these fields you’ve worked in when you first started? Could you see or have a vision of where this all might head?
I knew that computers were going to be very important, but I had no inkling of how sweeping and rapid the changes would turn out to be. I foresaw the importance of small computers in the days when most of the computers were large monsters in glass-enclosed rooms. There was a conventional wisdom that bigger computers were more powerful at that time, but I could see in the early 1960s that small computers were going to be increasingly dominating the business. I wrote a letter to Licklider [interviewee’s note: Licklider was then director of DARPA] – who was then championing from DARPA – was championing time sharing and arguing that individual computers like those that eventually got called PCs were going to be the future, not the big monsters. Unfortunately, I can’t locate that letter.
What do you deem as your greatest achievement?
Well that’s easy – having a fantastically supportive wife and three wonderful children, who are now successful adults, eight healthy grandchildren and one beautiful great-granddaughter.
Are any of them involved in computing?
My youngest son worked for a while for IBM managing and selling software to the automotive industry. And he now works for Trilogy, a company that’s also selling software. He’s managing a group that’s developing and selling software to auto industries.
The only one that I can think of is that I have not expressed my gratitude adequately to the many, many colleagues and co-workers who have been essential in whatever successes there were.
What was your biggest disappointment? Was there a project or goal that you did not attain?
Many, many, many. In fact, when I saw this question I was trying to figure out what the ratio of successes to failures was and I figured it was five failures to one success, which reminded me of I think it was Churchill that once said, “Move on from each disaster with undiminished enthusiasm.” And I guess I’ve done that and also have a little, or great deal of, perseverance. But the hard part is knowing when to quit.
How do you know that?
I don’t. (Laughs)
What future goals do you have for the medical school?
I think that applications of computing and computers should be as integral to the medical school education, research and practice as [it] is [in] genomics.
This is not on the list – do you feel that there ought to be specific courses taught in the medical school on computing? Should there be a computing curriculum?
I don’t know how to organize it, but there should be more understanding of computers for those, well for both those that are going into research and those that are going into practice. But those that go into practice need a different set of facts about computers than those that are going into research. And I don’t pretend to know how the curriculum should be set up, but I do feel strongly that it’s as important to the future of medicine as is genomics. The government has, under [David] Brailer, who recently resigned but got a good start, recognized that the electronic health record is an important thing, but the docs haven’t really adopted it yet. And I think it’s a generational matter [and] that now that more and more docs have the ability to type, it may become easier and easier.
Actually related to that is a question we didn’t ask earlier was about the Medical Care Group, MCG, and at some point there was a collaboration between BCL and MCG regarding medical records.
Did anything ever happen with that?
Well, medical information systems have been around the corner since 1960. And I wrote a paper once entitled “The Physicians Workstation, or the Nearest MIS yet,” where MIS was (spells) M-I-S for medical information systems. And the paper just observed that medicine is an information-rich activity and an obvious candidate for computers [now] that the technology is getting mature, but the payoff for the individual physician in clinical practice is still not quite strong enough to make the physician want to use computers daily in his interaction with patients. It seems like it’s just around the corner. Many attempts have come quite close, but it’s still not there and I’m not smart enough to know when it’s going to be there.
That was the last question. Do you have anything more to add?
We haven’t talked about computers in cardiology. In 1974, Paul [G.] Hugenholtz and I talked about starting an organization, an annual meeting, which we decided to call Computers in Cardiology. The kick-off meeting was at NIH in 1974, but then I had meetings alternately in Europe and the United States. Paul Hugenholtz was from the Thoraxcenter in Rotterdam, [and a] cardiologist. So he was the cardiologist and I was the computer scientist. Computers in Cardiology is still active, so that’s pretty good. Its thirty years, more – thirty-two years I guess – and we’ve had some wonderful times. Paul Hugenholtz had the saying, quoting others, [but] I’ve forgotten who, “Make no small plans, they do not have the power to stir men’s souls.” And so he had ambitious plans for Computers in Cardiology, including the “social.” It was an annual meeting and everyone looked forward with great anticipation to the “social” and each new host tried to outdo the last host in terms of the “social.” The idea was that you would ride on buses, boats, whatever transportation – trains – that would throw people together in random ways that would allow them to talk about their families or their work and get to know each other. And that has worked out very well over the years. There were very exciting socials. One of them was [when] Paul rented a train – an old steam train – that went up the path of “A Bridge Too Far” [ed. note: a 1974 book by Cornelius Ryan and a 1977 movie], and back and we had champagne and lunch on board the train and Paul and I got to drive the steam train and that was great fun. There have been others that have been equally exciting. So it’s nice to look at the bound copies of Computers in Cardiology that line my shelves at home and to see that it’s still going on.
One other story. I was on the [interviewee’s note: council of the] National Center for Human Genome Research for five years from 1990 to 1995. And that was a very exciting time. I was the token computer scientist – everybody else was a geneticist – on the committee. And we got to review the grants that had been approved by study section [and] were coming through council to be approved. And toward the end it was clear that things were coming along, but it was not so clear that the goal was going to be achieved. There was still a lot to do and things were going slowly and there were disputes here and there. And Bob [Robert H.] Waterston had been invited from here to give a talk on his proposal on how to make a factory to turn out base pairs and so Bob came and gave a talk to the council. And after it was over everyone on the council applauded. And I have been on a lot of councils and I’ve never seen people applaud before. So it was a turning point.
Well, I think that’s a nice interview and I appreciate your talking to us. I think everybody who listens to this, I’m sure, will learn a lot and I just want to thank you.
Oh, sure. It was fun.
I have [enjoyed it] myself.
I’m glad you were able to do it. It was fun to have an old friend.