Oral-History:Saburo Muroga

About Saburo Muroga

Saburo Muroga, who received Ph.D. degrees from Tokyo University, was for many years a Professor of Computer Science at the University of Illinois. Interested in creating powerful algorithms for logic-design automation, he developed minimum-gate design and the Transduction Method. He was an honorary member of the Information Processing Society of Japan, and IEEE Fellow, and a Distinguished Visitor, IEEE (1988-90). He was also honored with a Contribution Award from the Information Processing Society of Japan in 1991. In the interview, he discusses the course of his research and the funding he received from the National Science Foundation. Though he cites instances when NSF decisions were influenced by politics, he concludes that it was nonetheless less political and more transparent than many other government agencies, both in the U.S. and in Japan.

About the Interview

SABURO MUROGA: An Interview Conducted by Andrew Goldstein, Center for the History of Electrical Engineering, July 23, 1991

Interview #111 for the Center for the History of Electrical Engineering, The Institute of Electrical and Electronic Engineering, 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:

Saburo Muroga, an oral history conducted in 1991 by Andrew Goldstein, IEEE History Center, Piscataway, NJ, USA.

Interview

INTERVIEW: Saburo Muroga

INTERVIEWER: Andrew Goldstein

DATE: 23 July 1991

PLACE: Telephone Interview

Education and Career

Goldstein:

Can you tell me about your education and your academic career?

Muroga:

I received a garcushidegree from Tokyo University in 1947. A garcushi degree is the equivalent of something that is between a bachelors and masters degree. Then in 1958 I got a Ph.D. from Tokyo University in EE Department. I have two different Ph.D.s from Japan. One is just like a United States Ph.D., and the other was done the traditional Japanese way, where the Ph.D. was earned without taking courses. In the Japanese method, I submit a collection of papers, and then based on that the Ph.D. is granted.

After graduation, I went to work in the national railroad laboratory in Japan. There my responsibilities partly involved in the bullet train program. After two years I went to work for the Japanese government in the [inaudible] Commission of the Japanese government, which is their version of the FCC. In 1959, I joined the NTT, which is the Nippon Telephone and Telegraph Corporation. I became a staff administrator for their research laboratories.

In 1960, I joined the IBM Research Center in Yorkshire Heights, New York. Thereafter, in 1964, I went to the University of Illinois. They brought me in as a Professor of Computer Science and Electrical Engineering. I had a joint appointment.

University of Illinois and Parametrons

Goldstein:

I’d like to ask you about your history with the National Science Foundation. The earliest grant that I have on record of you receiving was in 1970. Had you gotten any money from them before that?

Muroga:

No, I don’t think so. When I came to the University, I didn’t know about the National Science Foundation. I had some money from the ILLIAC-IV project, which had the big money.

Goldstein:

Why did you go to Illinois?

Muroga:

I heard from everybody it was quite a nice place. I had the job in language , but we had communication problems. In other words, by coming to U of I, I can be my own boss; I had more freedom on research.

Goldstein:

What work were you working on between 1964 and 1970?

Muroga:

After NTT, from 1959 to 1960, I designed a data computer, with parametrons. Parametron is quite a unique logic element. Parametrons don’t use transistors or vacuum tubes because they are made of completely active devices. It has a core, many cores in fact, and then capacitors and resistors, nothing else. Using parametrons, I designed computers and was in charge of operations of this computer, until 1960.

Goldstein:

Were those components handy for your later work, or did you stop working with them after this period?

Muroga:

Parametron is basically a sersar logic gate. In other words, by working on parametrons, I tried to develop a theory that was suitable for the design of parametron computers. I guess I can say I am the pioneer of sersar logic. Around the same time in the U.S. there was a Professor, from the University of Pennsylvania, who came up with the sersar logic at the same time that I did, but he did so completely independent of me.

Sersar logic is quite expensive to implement. It’s difficult to handle the many elements in sersar logic. Many people gave up on sersar logic because it is so complicated. I, however, applied it to the [inaudible] design. This idea came from integer programming in the logic design method.

Goldstein:

You describe this in your book, correct?

Muroga:

Yes, and this is [inaudible] design of the sersar logic. In fact, these days there are some people that are partial to this specialty. Therefore, in this case the method turned out to be quite important.

Minimum Gate Design

Muroga:

Next, we developed the design procedure for the power adder using a minimum number of NOR gates. NOR gates are a special case of sheshagi. NOR gates are the gates on many of the simple gates. We asked people to try and design them so that they used a minimum number of NOR gates.

In a typical example of the first design [inaudible], it would use a minimum number of NOR gates. Then, later they would be cascaded for others because that makes it more sophisticated. We found that this was not so. Consequently, we concluded we needed special modules for this.

We needed special modules that could carry many inputs and outputs. In the case of this other module, it offered only one input and output carry line. In ours, we had numerous input and output carrying lines. Therefore, we designed a strange kind of module that could be used for a small network using the minimum number of NOR gates, and one that used the integer programming approach.

Goldstein:

When doing this, you must have worked it out on paper as opposed to doing it with computer hardware, correct?

Muroga:

Yes, this work was done on paper first. By doing this, by cascading certain modules, the entire network gets a minimum number of NOR gates, no matter how many modules you are connecting.

Goldstein:

This work was being done for the ILLIAC-IV, correct?

Muroga:

No. I believe I received money from the military around that time.

Seeking Funding from the NSF

Goldstein:

When you were doing this work with the adders, were you working on a particular computer system, or were you doing general development? I am trying to ascertain what brought you to the National Science Foundation to seek funding?

Muroga:

When I came here, the ILLIAC-IV project had just started, so the Department had plenty of money. Meanwhile, I tried to get research funds from outside, because when the ILLIAC-IV project went to committee, it was decided I needed to find the research funds for my projects. As a result, I applied for funds from the military.

Goldstein:

Did you also go to the Department of Defense for funding?

Muroga:

No. I focused on the NSF. I had heard the NSF was quite a good program. I found this to be very true, because in dealing with them they have been very organized and demand high efficiency.

NSF Funding in 1970s vs. 1980s

Goldstein:

I see that you received funding from them fairly regularly through the ’70s. Did you also get money from them during the ’80s?

Muroga:

The support was fairly regular in the 1970s, but that wasn’t the case in the ‘80s.

Goldstein:

Why do you think that was so? Were you working on something outside their area, or was there just less money to allocate?

Muroga:

I think the NSF probably began supporting more people. Also, it seemed that the funding process became very political. Some of the people would team up with university professors and make big noise.

Influence of Work

Goldstein:

Can you name for me any computer systems that implemented your designs, your minimum gate designs?

Muroga:

I don’t know. My design is a logic design mural, so any computer designer can use it as part of their design.

Goldstein:

Was it your intention to have your work, say for example a research paper, read by some other researcher who might then decide to implement your work into their projects? Did this happen?

Muroga:

That is difficult to know. But the data works were used for the transduction missiles. Transduction missiles were made for ArCom  or the study of special logic in the integral programming approach. These products are used extensively by major chemical companies.

Goldstein:

How did they come to adopt the method? Did they read about it?

Muroga:

Yes. We presented many papers on the subject, and they realized the importance of our approach. [Inaudible phrase] graphics, philosophies, and many other companies, began using our approach.

NSF Politics

Goldstein:

Did you have much close contact with the program officers at the National Science Foundation? Did you discuss your research with them?

Muroga:

No, not very much. I don’t travel very much, so I never go to Washington. Occasionally, I would meet the people from NSF at the conferences. But initially, I had quite a bit of difficulty in promoting our approach. Oddly enough, [inaudible] paper became really political, so we submitted the paper to many developed nations. Our papers were always rejected, for many reasons. The review committees would say there was nothing new, or something was wrong with the theory, or they wanted proof, and so on.

All of these comments were nonsense. As I said, oddly enough, the project became very political. So when I submitted the papers to NSF, they made their comments by [inaudible]. They already had my paper about the textbook on missile design. I made the correct association at this point, and due to this received their funding for a while.

Goldstein:

When was that?

Muroga:

Around 1987, ’86, or maybe ’89. For some reason or other, I couldn’t get the funds, and it was only after it was found that our approach did work that they found the money. Our paper was finally accepted by Roxsholt in 1989. The Roxsholt paper provided comparisons that tested the benchmarks on the different approaches. We then used the benchmarks to show the accuracy of our approach. The results were very surprising to other people. So for the first time, our paper on [inaudible] method was accepted by a conference. Then, Barbara Dietwood changed their attitude. At one time they even had a special class for studying our approach.

1981-1989 Period

Goldstein:

Can you point to any other key stages of your career? I see the transduction method was a major contribution. Are there anything other moments or contributions you would like to talk about?

Muroga:

I studied the transduction method in 1981. The paper on it finally came out eight years later.

Goldstein:

What were you doing during those seven years between the study and acceptance of transduction?

Muroga:

Between that time I had to argue with the [inaudible]. It was controlled by IBM [inaudible]. At one time, [inaudible name] was IBM guy. He wanted to promote IBM’s [inaudible] missile. There emerged a moment of unexpected opportunity for me.

I had another quibble, there were complaints about the Adder, and they [inaudible]. One time I even gave up, and tried to publish my study as a part of a book about [inaudible] information, but the book publication was cancelled. So, I went back to active reviewing. Then, finally, the paper came out in 1989, eight years later. This turned out to be quite a [inaudible] missile.

NSF vs. Research Funding in Japan

Goldstein:

Is there’s anything you would like to call attention to in your research career, or in your dealings with the National Science Foundation. Are there important issues or moments?

Muroga:

I’m really happy with the National Science Foundation. When the research office became political they had the tendency to intrude, but even when they did it was still less intrusive than other research funding agencies. Moreover, whenever I complain, they are willing to listen, and if they agree with my position I can continue receiving funding. I believe they are the best research-funding agency throughout the world. I’m really happy with the National Science Foundation, and I probably couldn’t have achieved what I did without them.

Goldstein:

In Japan, then, how is basic research funded, if you say it’s not done through a federal organization like the NSF?

Muroga:

There is a single agency there and it is much more political, and less informative. In the case of NSF, they show why the research proposal was rejected. There is no such thing in Japan. There’s no way to argue or complain, it’s just a one-way communication.