Edward Feigenbaum’s Personal Comments on the Collection

Dr. Feigenbaum deposited his material to the collection in two separate accessions, once in 1986 and again in 2005. Both times, he sat with an archivist reviewing each folder in every box, providing context, clarification, or stories relevant to the documents and folders he finds of interest. All of these supplementary notes are listed below, organized by box, folder, and document. The raw audio recordings and transcripts of these sessions are also available at this link.

Expand AllCollapse All

2005 Series
Box 1 Box 2 Box 3 Box 4 Box 5 Box 6 Box 7 Box 8 Box 9 Box 10
Box 11 Box 12 Box 13 Box 14 Box 15 Box 16 Box 17 Box 18 Box 19 Box 20
Box 21 Box 22 Box 23 Box 24 Box 25 Box 26 Box 27 Box 28 Box 29 Box 31
Box 32 Box 33 Box 34 Box 35 Box 36 Box 38 Box 39 Box 40 Box 46 Box 47
Box 51 Box 67 Box 26a
1986-1987 Series
Box 1 Box 4 Box 5 Box 6 Box 7 Box 8 Box 12 Box 14 Box 15 Box 16
Box 17 Box 18 Box 19 Box 20 Box 21 Box 22 Box 23 Box 25 Box 29 Box 30
Box 31 Box 33 Box 34 Box 36 Box 37 Box 39 Box 42 Box 43 Box 44 Box 46
Box 47 Box 49 Box 55 Box 56

Box 1 - 2005

Boxes 1 to 7 came from my office at 1 Welch Road, Building C., Palo Alto. That was the office of the Knowledge Systems Laboratory during the 1980's and until the mid 1990's when the laboratory moved over to the new Computer Science Department building -- the Gates Computer Science Department Building.

Folder 5: American Association for Artificial Inteligence (AAAI) - Feigenbaum Presidency1981

Agenda for AAAI Council Meeting, August 24, 1981
Interesting folder called "AAAI Feigenbaum Presidency". I was the second president of the American Association for Artificial Intelligence. This says there's an AAAI Council meeting August 24,1981. when I was Allen Newell was the first president from '79-'80, and I was the president from '81-'82.


Folder 12: Artificial Intelligence (AI) / Entertainment Project1993

AI Entertainment Project
I think this was some speculations with Barbara Hayes-Roth who was working here of how we could extend artificial intelligence into the arena of computer entertainment. This was 1993. Barbara started off on that work. I went to the Air Force. She continued that work and then she spun off a company to do that commercially.


Folder 13: Aldo Ventures1989

News from Aldo Ventures - 1989
Aldo Ventures is the name of a company run by one of my former students and colleagues, Avron Barr.


Folder 15: Archived Files I1988

draft Recommendations, National Research Council Computer Science and Telecommunications Board
There are some other archived material here in archivedreddy.manuscript. Looks like Raj Reddy's proposal -- Raj is always thinking in terms of massive and massively important computer science projects. So this looks like he was proposing the establishment of a national information network, a computer communications superhighway, a national digital library -- all these massive, national projects.

What Hath Simon Wrought?
This is an article that I wrote in honor of Herbert A. Simon, for a festschrift for Simon at Carnegie Mellon. Simon was my mentor and one of the founders of the field and a Nobel Prize winner (Economics, 1978)


Folder 16: Archived Files II1988

AI Qual Information
Here is a list of the qualifying exams for artificial intelligence with the actual people taking these exams and then who were the examiners for these different people in what topics

Archived files
The first several pages in this file give printouts from the SUMEX-AIM computer system of materials that were archived in the directory of the name of a secretary that I had at the time: Sue Irvine. She was apparently followed by Ellie Englemore (maybe she was later than Ellie.) But in any case, these are printouts of files that were in the directory of SUMEX-AIM. And so we have all the titles of these and the files themselves can be retrieved, I think, from my colleague Tom Rindfleisch who has been doing electronic resurrection of these files from the old SUMEX-AIM system. (As of 2017, Rindfleisch is alive and well).

Archived Files
The first several pages in this file give printouts from the SUMEX-AIM computer system of materials that were archived in the directory of the name of a secretary that I had at the time: Sue Irvine. She was apparently followed by Ellie Englemore (maybe she was later than Ellie.) But in any case, these are printouts of files that were in the directory of SUMEX-AIM. And so we have all the titles of these and the files themselves can be retrieved, I think, from my colleague Tom Rindfleisch who has been doing electronic resurrection of these files from the old SUMEX-AIM system. (As of 2017, Rindfleisch is alive and well).

Expert Systems and their role in society
There's material labeled Dertouzos and that's Mike Dertouzos from MIT. And I don't know what this means. Did Mike do a readings book to which I contributed material or did Mike ask me to do a summary of where things were going in knowledge systems? I'm not sure.


Folder 17: History of Medical Informatics1987 - 1988

American College of Medical Informatics, Fellows - 1988
Medical Informatics is a medical buzzword term that really means computer science in medicine, or computer science applied to medicine. We spun a medical informatics group out of our lab into the medical school. It was called the Section on Medical Informatics of the Department of Medicine. The founding person of that laboratory was Ted Shortliffe M.D. Ph.D. who was one of our students. He was followed by Mark Musen M.D. Ph.D. who was another one of our students. The group of people working in Medical Informatics started an honorary society called The American College of Medical Informatics. This is material from the founding of that.


Folder 25: Babbage Interview1979

Charles Babbage Institute, Oral History interview
This file folder is called "Babbage Interview". That Babbage means Charles Babbage Institute for the History of Information Processing at the University of Minnesota Library. I believe these are the interviews that Pamela McCorduck did for her book "Machines Who Think", reissued in the year 2004 in an updated edition. Quite important book. There's a whole chapter or two devoted to my work in there. And I believe that this is the Pamela McCorduck that donated all the oral interviews to the Babbage Institute. So this is the typed script of the interview plus letters from the archivists having to do with the agreement as to when certain information in there can be released. In addition to that, I believe there is an additional interview with the Babbage Institute that was done as part of a DARPA project (Defense Advanced Research Projects Agency history) .Arthur Norberg , the Director of the Charles Babbage Institute , did a book under DARPA contract about the history of DARPA's support of information processing. I believe that there is another interview at the Charles Babbage Institute which documents more of my career as well.


Folder 31: The Center for Advanced Medical Informatics at Stanford (CAMIS) 1990

CAMIS proposal site visit
There's a file folder called "CAMIS". I don't remember exactly what the acronym stands for but MIS mean medical information systems. This is a proposal about getting more money presumably from the National Institutes of Health. I can't remember why we invented the acronym CAMIS. Was it the successor to the SUMEX-AIM resource in another form and we needed to invent something so we called it CAMIS? The core research has to do with a subproject in my lab called "How Things Work" which we've already mentioned elsewhere in these archives. And so somebody on the site-visiting committee for evaluating our proposal asked a question about CAMIS and we gave some three different substantial answers to that.


Folder 32: Campbell Communications1989

Final version of white paper for Representative Tom Campbell on IT Policy
Tom Campbell is a former Stanford Professor who was a member of the House of Representatives in the late 1980's. This is a set of communications with Tom, including a meeting with him that took place somewhere around the middle of 1989. Some of these are dated in June and September and so on. The substance of the set of communications with Tom is self evident from reading the material. There was a great concern about research funding -- that is, research funding for artificial intelligence and for computer science in general and it was not only because of some things going on at the main funding agency DARPA for AI research but also because there was some national trends that were underway in funding of large science projects. That's about all the detail I can give you right now on this tape except if I would read these memos. But I think they're self explanatory. And Tom was a good friend. He was like I said; a former Stanford colleague and he listened very carefully to what we had to say. And I was personally involved in a lot of the national debate on how much national investment there should be in the computer science and artificial intelligence field. That all started because of that book "The Fifth Generation" which had to do with the Japanese national investment in this area; should the Americans be making a comparable investment?


Folder 34: China--General 1985

This folder contains correspondence that has to do with interactions between myself and Chinese students and scholars. Stanford's interaction with China began at the time of the first U.S. China thaw, in the post-Nixon thaw. But my own linkage with China didn't start until somewhere around 1983 or 1984 when I went to Beijing and Shenyang for a National Science Foundation workshop with the Chinese on artificial intelligence topics. That workshop had Dr. Saul Amarel in it and Dr. Roger Schank and other worthies. So as a consequence of that, I met a bunch of people and through them and to their students, this resulted in a flurry of letters of "Can I be admitted" or "Can we do research" or etc. This folder is interesting because it represents a very early stage in the interaction between American computer scientists and Chinese computer scientists. So historically, it'll be very interesting.


Folder 35: Paul Cohen 1987 -1988

Paul Cohen was a student of mine here. Paul Cohen is an important computer scientist and as of 2004, he worked for the University Of Southern California information Science Institute. He was one of the best students that I had and he coauthored one volume of the Handbook of Artificial Intelligence with me. He was one of the few students that I had who worked in a joint program which was partly computer science and partly psychology. So his Ph.D. was a joint CS Psychology Ph.D. That brand of student has kind of disappeared by now. There's not much of an interdisciplinary communication between us artificial intelligence scientists and the psychology scientists. Paul is one of the best. So this folder is a collection of some of his papers that he sent me. And they appear to be early or mid career papers of Paul's.


Folder 36: Center for International Security and Arms Control 1986

This folder has to do with the Stanford research institute called CISAC which, at that time, in 1986, was the Stanford Center for International Security and Arms Control. And currently it's chaired by Bill Perry and headed by Scott Sagan. I'm affiliated with it as being a member of their executive committee. I've been affiliated with it since returning from the Air Force in 1997 at the invitation of Bill Perry.

Expert systems applied to crisis management
This file says that there was a lunch that involved myself and Joshua Lederberg and it had to do with the possibility of building an expert system to assist humans who are involved in crisis management. There might be a big military crisis or a nuclear crisis or an earthquake crisis or something like that, where a lot of human knowledge would need to be brought to bear in a big hurry and maybe there could be some assistance using expert systems. John Lewis, the Director, proposes that there be some kind of a meeting and he proposes that various of the well-known people in expert systems be brought in to try to make some sense of this. And he includes some other documents in with this and including a paper by Alex George on crisis management. I don't remember a thing about this. I do remember that we talked a lot about crisis management in the 1980's as a possible area of application of expert systems. But as far as I know, nothing really ever happened in that.


Folder 37: CGL - Computer Graphics Labratory1988 - 1989

Letter from Robert Langridge to Professer Edward Feigenbaum
In 1989, apparently I was asked to be a member of this committee. I was asked to be that member by Bob Langridge who was head of that group. Bob is an old friend that I had met on the National Institutes of Health Peer Review Study Committee for the Mathematics and Computer Science area back in the late 1960s and early 1970s. Bob was one of the early people in chemical computer graphics. I remained friends with Bob for a long time.


Folder 38: Computer Science and Technology Board 1987--1990

The CSTB still exists. As a member of the Academy, I've had several quiet discussions with people including the president of the National Academy of Engineering about what I see as -- basically that it's doing too many studies. It's wasting too many people's time. It's burning up a lot of money doing studies because that's what the CSTB does. It does studies. So CSTB, why don't you do half as many studies, you know, really make a high priority selection. That viewpoint was not well received at the National Academy. And partly because they truly believe that the studies they are doing are excellent; and partly because when there is a big bureaucracy of people employed, they have to continue to be employed. You have to keep getting more money for more studies.

Letter from Edward Feigenbaum to Norman Hackerman
The arm of the National Academy of Science that does studies for the government is called the National Research Council. One of the study boards of the National Research Council that was established in the 1980's is called the Computer Science and Technology Board, otherwise known as CSTB. Its first Chairperson was Professor Joseph Traub and I'm not sure if he was still at Carnegie Mellon at the time or went to Columbia University by that time. But anyway, he was a friend of mine. He happens to be also the husband of my coauthor Pamela McCorduck on several of the books that I've done. Traub asked me to be on this Study Board, the founding Study Board of CSTB which involved quite a bit of work. I did it but I ran into several frustrations with the CSTB, including partly the long trips involved for what seemed like very little gain in the discussions. And secondly, the difficulty of convincing what I will refer to as the bureaucrats at the CSTB and the National Research Council that because we had half of the committee lived out here in California that we should have half of the meetings out here in California, not in Washington. We ended up having one meeting in California but that was always a very sore point. And the third thing was a much more serious one and that was a substantive one that had to do with a steady sort of a national assessment of the field of artificial intelligence which I had proposed and which ran aground with major bureaucratic hassle within the National Research Council. And at that point, I just decided I don't need this. I got a lot of other things I want to do with my life and I got a lot of consulting I want to do where I was earning a lot of money. I didn't need this particular frustrating non-profit work. So I wrote a letter of resigning from this committee which is in here and it's a very milk toasty letter, you know, saying that I'm really busy and really can't do this kind of thing. But actually there were issues and the issues were only slightly covered in that letter.


Folder 40: Congressional Contacts 1983--1993

It appears that this folder "Congressional Contacts" is mid 1980's. It's part of the flurry of sort of high level science politics that I was engaged in during that period as a result of having published that book "The Fifth Generation" and having gotten involved with how big should the U.S. investment in information technology and computer science be relative to the Japanese and the Europeans and various legislators.

Congressional Contacts1983 - 1993
It appears that this folder "Congressional Contacts" is mid 1980's. It's part of the flurry of sort of high level science politics that I was engaged in during that period as a result of having published that book "The Fifth Generation" and having gotten involved with how big should the U.S. investment in information technology and computer science be relative to the Japanese and the Europeans and various legislators. Some state got interested in that question and apparently there was an interaction with an old friend of mine, Wes Clark, who at that time was serving as a House Counsel to Senator Kennedy. Dr. Wes Clark is a famous early computer scientist and the fact that he was associated with Kennedy -- this must be Ted Kennedy because it's the mid eighties -- make sense because Wes Clark was from the Boston area. This is an example of that correspondence including a telegram which supports some position that Wes Clark or Kennedy needed supported at the time.


Folder 41: Consulting--Undersecretary of Defense 1986

This is a folder that has to do with some 1986 consulting that I did for the government apparently under the Office of the Undersecretary of Defense and I was looking at revealing some artificial intelligence stuff to them. So when you do that, you have to fill out a lot of government forms that include financial conflict of interest forms. So there's a rather interesting one here with just form DD1555 in this package which talks about these different companies I was involved with during that period. So this is 1986 when I'm filling out this form and you can see from this that IntelliCorp, TechKnowledge -- those are two of my own companies that I started were on here. What role did I have? Sperry Corporation, I was the Director. Trilogy Corporation, I was an investor, and so on. Heuristic Press is little publishing company that I had with Avron Barr. Feigenbaum, Nii and Associates Incorporated was just my wife's and my small consulting company. Sequent Computer Corporation where I was an advisor to them and had early stock in that corporation. And so on. So that's quite interesting to look at. Then let me see. It says here that I probably had a secret security clearance. I'm also wondering in an employee's withholding allowance certificate whether my social security number should be made public and put on -- if we scan these documents and put on the web. That's an open archival question because that's sort of identity theft kind of worry about, you know, do you really want your social security information public.


Folder 43: Copyrights, Patents 1980--1986

This folder actually contains two different kinds of information which need to be sorted out by the archivists and codified separately. One is just a set of statements by me allowing people to use material of mine or disallowing that they use material of mine and in their publications. And that will help a scholar track who is interested in my work particularly from the point of view of actually quoting it in their published articles or in books that they were writing or reprints or something like that. Then there is a totally different thing in her which has to do with a major flap that was going on at Stanford during the 1980's -- a flap which continues actually to the current day -- but it's seething not boiling -- that has to do with what rights does Stanford as an institution believe it owns in software created by its employees both regular staff employees and and faculty members and students. This became quite a hot issue in the early 1980's when the company bubbles started to boil in the pot. I mean, we had some hardware companies out of this department like Sun Microsystems and we had a whole bunch of software companies because most of us are software people, including a few of my own. Question is what rights did Stanford have? What rights did the developer have? How do we obtain these rights from the government?

Policy on Royalties from Software
There's a letter from me to Pat Devaney in the Office of the Provost for Research where the "New York Times" had quoted something about Stanford's policy is according to Pat Devaney and then Ed Feigenbaum writes to Pat and says 'Pat, we're, you know, we're really interested in this and tell me what Stanford's policy really is because this this isn't the way we understand it'. And then this gets involved later on with the Committee on Research at Stanford, Academic Senate Committee on Research and then policies get formed and then eventually something comes out from Gerry Lieberman who is the Vice Provost for Research. Looks like it gets into something having to do with the Board of Trustees but all of this is an extremely important issue that lasts right into the present day because Stanford is one of the fountains of software innovation in the valley. And one of the ways it gets out is through companies that we license the software to or that faculty members create in order to develop the software.


Folder 46: DARPA Advisory Committee 1984

There's a folder here dated 1984 on the DARPA Advisory Committee, so-called. I have many contracts with DARPA as principal investigator on quite a number of things. And so one gradually acquires a status of some visibility and eminence in the so-called DARPA (Defense Advanced Research Projects Agency) community, the main funder of basic and advanced applied research -- advanced development research for the Defense Department. And it's the main funder of the great developments of computer science and information technology including the internet and computer graphics and artificial intelligence and many other things. Because it's such a big player, as leadership changes within DARPA -- which it does routinely, usually per administration -- and some of the scientists rotate in and out of there every few years, there's always some kind of a little turmoil and hurricane brewing, ranging from a little turmoil to a big hurricane brewing within DARPA. That was especially true around the time of the directorship of Robert Cooper and the so-called Strategic Computing Initiative which is separately documented in a book in the history of computing that was written a few years ago. I believe the book is called "The DARPA Strategic Computing Initiative". Some of the members of this community, particularly a very aggressive leader, Michael Dertouzos of MIT, now deceased, helped DARPA put together a group of eminent people of this DARPA community which would help them search for a new director at the time.

Meeting with Dr. Duncan
It looked like a University of Southern California person, Keith Uncapher, Carnegie Mellon person, Allen Newell -- all of these are very important people -- were participating in this advisory committee. And a subsequent Director of this Office of Information Processing Technology, Saul Amarel, of Rutgers University. Then it looks like by the time you get to 1986, there's another Director named Robert Duncan -- notice that we've passed on from Cooper to an Acting Director, Bulfano to Duncan and then it looks like Duncan discovers that we're not really an official committee because we were never approved by the Office of Management and Budget. And so he sort of says action has been initiated to establish this committee and to fulfill the statutory requirements. And I have no idea whether we ever met again.


Folder 47: DARPA Proposal 1988 -1991

Letter from Edward A. Feigenbaum to Dr. Jacob Schwartz
A DARPA proposal. This is just one of many that is a continuing flow of these proposals from 1988. I think when the first one went in or it was part of the Stanford Artificial Intelligence Laboratory Proposal at the time. Then it became a separate stream of proposals from myself as Principal Investigator of the Heuristic Programming Project, later called the Knowledge Systems Laboratory. All to DARPA, all in the range of something of the order of a million dollars a year up to some being two million dollars a year of research funding to do various things in artificial intelligence mostly, and some computer science things as well. This happens to be one of those addressed to the Director of the Office of Information Processing Techniques, IPTO, had changed its name as part of one of these DARPA hurricanes to ISTO, Information Science and Technology Office, and its Director was one Jacob Schwartz who was a famous professor from NYU. And so we're sending this proposal to Dr. Schwartz for funding. And so it has all the things we want funded and the budgets for the different years and this is especially helpful. I remember Henry Lowood telling me years ago that he would like to see all these proposals as a way of tracking what we were doing at the time and how much money it would take to do that.


Folder 48: DENDRAL Original1965

DENDRAL Original1965
Now we're into the meat of the history of one of the most important projects that we did, the DENDRAL Project. This particular file called "DENDRAL Original" has in it an actual copy of the very first thing which I ever wrote about DENDRAL. I got here at Stanford on January 1st, 1965, initiated my collaboration with Joshua Lederberg. We came up with a kind of a plan -- it's dated April 5th, 1965 -- for what we wanted to do. We -- Joshua Lederberg is not even a coauthor of this document -- were still at the Stanford Artificial Intelligence Project. We hadn't formed the DENDRAL Project yet. There is also AI Project Memo number 30, April 5th, 1965 and it's co-written with a young colleague of mine, a former student at Berkeley who came with me down here to Stanford, Dick Watson, R. W. Watson. And what it does is list what we think we're going to do, and the purpose of writing this was to attract the attention of graduate students. And it was called an initial problem statement for a machine induction research project. What we were looking for was circulate this among the graduate students and find out who wanted to work on this project. This was the very first thing that was written about DENDRAL. I got this copy because Joshua Lederberg on July 4th was looking through his archives and he found that he had copies of this. So he sent me the original and a whole bunch of extra copies and this file-folder contains that. So this is a genuine treasure. It's like finding King Tut's golden crown or something like that.


Folder 49: Historical DENDRAL1977

Letter from Charles L. Coulter, Ph.D. to Dr. Carl Djerassi
The folder called "Historical-DENDRAL" is a letter from Charles Coulter of the National Institutes of Health to the person who was by then the principal investigator of the DENDRAL project. Lederberg and I had convinced Carl Djerassi to become the principal investigator as we transitioned the program out of computer science and much more into chemistry. And this is 1977. Carl and myself and Josh are seeking renewal of this DENDRAL project under the Biotechnology Resources Program of the National Institutes of Health. So the reviewers reviewed our proposal -- the peer reviewers of NIH who are generally very tough reviewers -- and Carl was sent a copy of what the reviewers said about our project at that time. Historically I think this will be very important to see how we were viewed by the rest of the medical and health research community and computer scientists at the time, plus there's some discussion of budget in here. So you can see how much we were proposing to spend at the time.


Folder 50: DENDRAL - Conception1987

Progress Toward History of DENDRAL
The file folder called "DENDRAL - Conception" first of all contains Joshua Lederberg's own attempt to write up the history of DENDRAL and that's why it happens to be dated the 7th of July, '87. So that explains why on July 4th, he happened to come across the original DENDRAL document because he was looking up all that stuff, putting together his own historical view. He writes and says he would be delighted if I could add something to this and my plea is that you give first attention to the accuracy of what's here, especially, Ed, your approval or revision of the long quoted message. So this is Joshua Lederberg's write-up. I assume that a good portion of this made it onto Josh's National Library of Medicine Profiles in Science Autobiographical website. And then there are some notes by myself called "Progress Toward the History of DENDRAL" in which I started to -- I don't remember in what context -- write some notes on the intellectual state of artificial intelligence in the late 1950's and early '60s as a way of laying the ground for why did we DENDRAL. How did that idea come up?


Folder 51: DENDRAL 1985--1986

This is a file called '"DENDRAL". It's not a very indicative title but it includes miscellaneous correspondence concerning a second edition of the main book on DENDRAL.

Letter from Michael B. Morgan to Drs. Bruce G. Buchanan, R.K. Lindsay, Jim Nourse, Edward Feigenbaum, and Joshua Lederberg
There's a letter here to Josh Lederberg, myself, Bruce Buchanan, Bob Lindsay and Jim Nourse from Mike Morgan the President of a publisher called Morgan Kaufmann Publishers -- the publisher of a lot of artificial intelligence books. And it's dated in 1985 and it has to do with some edition of this book. And I don't understand it because the book was published by McGraw-Hill and this letter references the McGraw-Hill book and it's as if Morgan Kaufmann was preparing to come out with another edition of the book. But I don't remember another edition of the book. So it fell through the cracks or maybe it did come out because there's a coming in 1986 ad from Morgan Kaufmann that says that there will be such an edition of the book -- a second edition and it gives an ISBN number. So it says formerly published by McGraw-Hill. So maybe there was such a revamped version of the book on the DENDRAL project.


Folder 55: Encyclopedia Americana 1993

There's a file called "Encyclopedia Americana, 1993". All that refers to is that they asked me to write an article for their 1993 edition of the Encyclopedia Americana on expert systems which I did. And the result of that was that they sent me a copy of the 1993 Encyclopedia American which I still have.

Box 2 - 2005

Folder 1: EPAM 1962--1984

EPAM stands for Elementary Perceiver and Memorizer. That's the name of the computer program which I developed as part of my thesis research at Carnegie Tech which is now Carnegie Mellon University in 1956 through '59 period. And that was a very important piece of research in the psychology side of artificial intelligence, what has been called computer simulation of cognitive processes or information processing psychology. Then I published many papers on EPAM over the years. The first batch of them were joint papers with Professor Herbert Simon, my thesis advisor, and that run of papers ended in around 1965. And that resulted in two versions of EPAM. Versions I and II. Then I worked when I started my research career in 1960 at Berkeley, Herbert Simon was still my collaborator on this work and he and I continued to write papers through 1965 on a new version of EPAM called EPAM III. Simon continued to be interested in the subject and late in the 1980's, he worked with a student called Howard Richman on a new version of EPAM called EPAM IV. And that led to quite a number of papers published in the 1990's on EPAM IV and maybe even another version. There may have even been an EPAM V or VI, I'm not sure. And this folder contains a miscellany of a few EPAM things. Then there are some materials here: Simon made a collection called "Models of Thought". Simon has many collections that he's done of papers of his over the years. And in this collection "Models of Thought", he's included some EPAM papers. And he sent me the copies of what he put in there. So I have several copies of that in this folder also.

A Theory of the Serial Position Effect (1962)
Chapter 3.1 of a book by Herbert Simon (which could probably be found in his extensive bibliography) contains a thing called a "Theory of the Serial Position Effect, 1962" with Edward Feigenbaum. And that was the very first work that Simon and I did together although the date 1962 is very odd because we actually did this work much earlier than 1962.
Theory of the Serial Position Effect probably was in the 1957-1958 period. That was actually what we called EPAM. So this is a remarkable little gathering of papers here that I didn't even know existed.

EPAM-like Models of Recognition and Learning
There's also a paper that Herb Simon and I wrote for the Journal "Cognitive Science" in 1984. So that's one of the latest papers -- Herb Simon and I didn't publish much together in the seventies, eighties and nineties. But this was one of the few papers where we did. It's an extension of the model which we did together called EPAM ELL. Then also in this folder is a very early paper by Simon and and myself, from "Psychological Review", 1962. It's overly titled "Computers and Behavioral Science" but I don't really think it's from "Psychological Review". That might have to be checked, it might also be from the Journal "Behavioral Science". Anyway it's a paper in which we explain how reading can be done by EPAM.

Information Processing and Memory
This a copy of a paper that I wrote called "Information Processing and Memory". There's a note on there which says it's a 1968 paper. Actually I wrote it in 1965 for a conference, the Fifth Berkeley Symposium on Memory, Mathematical Statistics and Probability. It's not only an excellent summary of EPAM at that time but it also is the first place where I introduce the idea of an information processing model of what dreaming activity might be all about. And that model was subsequently reinvented -- not picked up. It was simply reinvented by Crick and other people in the nineties without even knowing that this paper existed.


Folder 2: Expert Systems (ES) on Multi-Processor Arch: Results of Expiriments at the Heuristic Programming Project (HPP) 1990

The multiprocessor architectures was a very big project at our laboratory from, roughly speaking, 1984-5. Around 1990 it was part of DARPA's big vision called the Strategic Computing Project. And that, in itself, was partly a reflection of interest in competing with the Japanese in Advanced Artificial Intelligence Research. And this happened to be our project. And we developed quite a number of parallel computing models for doing the type of software that is called the blackboard model in artificial intelligence. We parallelized that. Now just to comment on the fact that these experiments were done at the Heuristic Programming Project, a historian might look back and say but the laboratory was already called Knowledge Systems Laboratory by that time. And the answer is yes, by that time, we had grouped ourselves into sub-laboratories of which the overall one was called Knowledge Systems Laboratory but there was the one run by myself and Bruce Buchanan, which was the original one, Heuristic Programming Project and then there was the SMI Laboratory which is a section on Medical Informatics which was the medical part of the lab and then there was the SUMEX part which was the computational facilities and new developments in computers lab. So that's why there might be a confusion but it's not really a confusion.


Folder 3: Fifth Generation Book--general 1982--1986

This file is called "Fifth Generation Book General". In 1983,I published a book with Pamela McCorduck called "The Fifth Generation" and it was about Japan's large research investment that they were beginning to make in artificial intelligence called the fifth generation project. First of all, it stimulated my book with Pamela. But it also was part of a large conversation going on in the United States at the time of how do we wrestle with the fact that Japan was becoming either masters of or becoming very good in many different fields and might take over some of those areas as markets, as they were beginning to do in the automobile area. So this file contains a miscellany of materials related to the Fifth Generation book including there was obviously a book lecture tour. So there's information there. There's marketing information from the publisher Addison-Wesley, and a lot of correspondence with Addison-Wesley.

Letter from Ellie Engelmore to Allison Betts
To me, the most interesting thing in here was a letter from my secretary, Ellie Inglemore, to "Time" magazine in connection with something. She said that "Time" magazine had come out with their issue that had special sort of a "We versus them" cover -- U.S. versus Japan -- and that meant that there was a cover article which was essentially stimulated by the Fifth Generation book. We did spend time talking to the "Time" magazine staff. In fact, I think it was Pamela McCorduck and I went to visit them on our book tour in New York hoping that as a result of this interview, they would publish an article about what we wrote and the Fifth Generation project and all that. And yes they did. They did it and they sort of absorbed in all the information, put it into the article and not once, referenced our book. And so Ellie says in that letter, "we are not amused". It was a lousy thing to do on the part of "Time" magazine and we let them know.


Folder 5: Fifth Generation - Misc.1982 - 1983

Research Report on Fifth Generation Computer Systems Project
This is the document that was responsible for kicking off the Fifth Generation project in Japan. And hence this is a document that was circulated widely around the world and that many of us paid a good deal of attention to. This document led to the creation of a research center in Japan called ICOT, whose acronym I've lost now. But it was the Fifth Generation Computing Research Institute. That was dated March 1983. Then in April 1983, they issued another document because April 1st if the fiscal year boundary in Japan. So they actually started the ICOT Institute and this outline of research and development plans for Fifth Generation Computer Systems is sort of their working document saying okay now that we've gotten started, this is what we're really going to do.


Folder 10: Gannett Center for Media Studies - Columbia University1986

Letter from Frank H.T. Rhodes to Dr. Edward Feigenbaum
This letter just asked me if I would be some sort of an associate for the Gannett Center. The letter describes exactly what they want. Associate doesn't mean very much and I was invited to their events but this is just indicative of the fact that after the book "The Fifth Generation" came out and all of the high visibility press that there was after that, I came to have a certain kind of, I don't know, visibility is one term for it or sort of rock star status is another term for it and get invited to things like this which never had much of an impact on my career.


Folder 15: Shozo Hibiko visit 1989

The title of this folder says "Shozo Hibiko Visit 1989" but actually it's a more general folder. It deals with things that relate to my relationship with people at a university in Japan in the city of Nagoya, Japan, a university called Chukyo University. I was helpful to them in getting a school of artificial intelligence and computer science started in Chukyo University on their new engineering campus. And then I helped them out with a satellite presentation. Apparently this must have been based at Stanford and then it may have been for the opening of their center or some special symposium having to do with their new center. I'm not sure actually. A careful read of these documents will straighten that out. But it has something to do with that. And, of course, I visited that campus many times. And Umemura -- listed here as the Chairman -- is also the Chancellor now of this university. So there are some letters from him in here.


Folder 16: Heuristic Programming Project (HPP) - Presentation, Materials, Achievements1991

AIJ Special Issue List
Barbara Hayes-Roth has a memo in here about how many of our researchers wrote papers that became the most highly cited papers from the AI Journal.

Heuristic Programming Project (HPP) - Presentation, Materials, Achievements1991
Titled "KSE-Overview" dated 1991. So we gave a complete overview of what we were doing and what books we had produced and our achievements and so on.

Heuristic Programming Project (HPP) - Presentation, Materials, Achievements1991
Looks like it was notes for a DARPA site visit -- DARPA being one of our funders. And then looks like there are my notes associated with that.

The Stanford How Things Work Project
There's presentation materials about an early 1990's project called "The Stanford How Things Work Project" headed by myself. The project leader was Richard Fikes who took over the laboratory when I was retired. There are research associates Bob Engelmore who was the Executive Director of the lab for a long time and now is dead; Tom Gruber who left here to form a company in knowledge based systems. And Yumi Iwasaki who left here as a researcher to move up to Seattle with her husband AhuKupta, who is now one of the Microsoft Vice Presidents.


Folder 17: Heuristic Programming Project (HPP) Executive Committee 1984

In a previous note, I've mentioned about the Heuristic Programming Project which was a subpart of the Knowledge Systems Laboratory. The Heuristic Programming Project was the one that I was directly principal investigator of. And it had a little executive committee. I guess that consisted of myself and Bob Engelmore probably, Bruce Buchanan. I'm not sure who else was on it but there's a variety of interesting information in here. There's many lists of students and the projects they did. So I guess each time we met, we must have distributed these notes to each other. Quite extensive review of what was going on at the Heuristic Programming Project. These all seem to be dated 1983,1984. Very, very vigorous time in the project's life.

HPP Executive Committee Meeting
There's a long list of the student research assistants that we had in the 1985 year or something like that. There's a list of the projects that we had ongoing and the Stanford project numbers associated with them. Like a project for Rockwell, project for GT&E, The National Science Foundation MOLGEN project (that is a famous one), a NASA project on Knowledge Representation and so on and so on. Several NASA projects. Quite a number of these projects. So that any scholar later on who wants to study these can track them down by virtue of the names of the students who are working on the projects and what they did. There's another list of that for '84-'85. In fact, it's a more extensive list. It includes the IBM DART project and DARPA's project on multi processor architectures.


Folder 22: Intelligenetics 1989--1990

This folder is called "IntelliGenetics". IntelliGenetics is the name with which the company IntelliCorp was born. The genetics part of it is an outgrowth of the fact that this came from research that we did in the MOLGEN Project. MOLGEN was this pioneering project in computational molecular genetics -- application of computers to molecular genetics.
And that that software that we developed led to a company called IntelliGenetics that was founded by myself, Peter Friedland, Professor Douglas Brutlag, still at Stanford, and Professor Larry Kedes, who is at USC Medical School right now. This project MOLGEN and the company IntelliGenetics has been much studied by Professor Tim Lenoir of the Stanford Department of History. There's extensive work on it and many of the MOLGEN documents have been scanned by Lenoir's people for use in that project. This is just a set of miscellaneous correspondence having to do with IntelliGenetics. IntelliGenetics never succeeded very well as a company, and was sold to Amoco Corporation, and there's a letter from Amoco. And the company itself changed the name from IntelliGenetics to IntelliCorp. And under that name, you'll find lots of other stuff. And I'm going to donate a lot of stuff.


Folder 30: Knowledge Processing: From File Servers to Knowledge Servers1989

Letter from Alison Roberts to Edward Feigenbaum
Here is a letter from Alison Roberts of the Kurzweil Foundation. Ray Kurzweil was putting together a book called "The Age of Intelligent Machines" and this was my contribution to his book. And that book eventually came out.


Folder 31: Knowledge Systems Laboratory (KSL)--Kyoto 1989

This folder is called "KSL - Kyoto". And I had, until I looked through the material, no remembrance whatever of why we were planning to have a Kyoto Branch Laboratory of the Knowledge Systems Laboratory, even though Stanford has a Stanford-Japan Center in Kyoto. I'm not sure -- there's a proposal here for some sheets that say it's a proposal for the Kyoto Branch of Stanford KSE. And I cannot recollect why we wanted to do that. Five-year commitment and what I would do and what Penny Nii would do and what Yumi Iwasaki would commit to this. And there's a letter here to Professor Nagao, Department of Electrical Engineering of Kyoto University and then later on, interestingly enough, Professor Nagao became President of Kyoto University. But I don't understand -- I just cannot recollect why we wanted to do that. The person who will remember pretty well is Yumi Iwasaki who's currently living in Seattle.


Folder 33: Large Knowledge Bases (LKB) 1986--1990

Ever since the expert system era, there was always this view that expert systems would get better and intelligent systems as a whole would expand their capability to be intelligent by getting more knowledge. That's the so-called knowledge principle which you can look up in many ways in what I've written. So this is a folder of miscellaneous items that were gathered in this long running quest for projects that would expand knowledge bases of computers. So, for example, there's a paper dated 1983 here by Doug Lenat and other people on a thing called the Knowsphere. Anyway, this is a very valuable folder. It says a lot about from the point of view of the generation of ideas. Who was involved with it? Who were we getting in? For example, looks like there's a workshop planning list here dated May 29,1986 of who we were thinking of inviting to a workshop, and so on.

KB Project Notes
Then there's some notes from Bob Kahn. Bob Kahn is one of the co-inventors of the internet. But Bob was running the Information Processing Techniques Office at DARPA. And these are notes on a knowledge bank project. So he too was clued into thinking about how to expand the base of knowledge that computers AI programs had.

Pursuing the Knowledge Principle
And there's an article here by two students of mine, Mark Stefik, Doug Lenat and myself called "Pursuing the Knowledge Principle". That is how the knowledge principle says that we just have to get onto the subject of expanding the amount of knowledge in knowledge bases and there are dates associated with these things, acknowledgements. It says many of these ideas were drawn from some lively discussions in July 1986, and then it mentions a whole bunch of people who were contributing to this discussion at the time.


Folder 34: Large Knowledge Bases (LKB) / How Things Work (HTW)--Technical Background Materials 1991

There's another folder here called "LKB" which stands for Large Knowledge Bases and "HTW" which stands for How Things Work. That was the name of another project that we were working on. Most of this material here is related to How Things Work. All of these materials are dated in the 1990 and '91 timeframe so it's very localized in time and I hate to disturb this folder from that point of view because it gives us some really nice sort of framework of time that goes along with this. But most of the pages in here have to do with How Things Work.

KB interchange standards
There's one memo here by two people, Doug Lenat -- I've already mentioned him, former student and researcher on Large Knowledge Bases and another former student named Guha, on knowledge sharing. And Guha, by that time I think, had finished here and he was working with Lenat at the Micro Electronics and Computer Consortium, MCC. And these are some ideas about large knowledge bases and knowledge sharing. It linked in actually with the How Things Work Project in a way because we were thinking of different projects around the country, codified knowledge of how various things work. Then how would they all share that knowledge?


Folder 38: MACSYMA Materials undated

The project Macsyma is an MIT program for doing symbolic algebra. In McCorduck's book, "Machines Who Think" Macsyma is accorded sort of equal status as the project that co-invented expert system ideas. Here at Stanford, we normally think that it was the DENDRAL project that Lederberg and I and Djerassi did that invented the expert system ideas. But McCorduck makes the case that there is an equally good case that Macsyma was there too.


Folder 41: Microelectronics and Computer Technology Corporation (MCC) 1985--1988

The folder is labeled "MCC" which stands for Micro Electronics and Computer Technology Corporation in Austin, Texas. It was an industry sponsored research institute that was intended to blunt the thrust of the Japanese Fifth Generation Project and get us going in the right direction. Many companies joined. So it was not government sponsored. It was industry sponsored. And it was in this laboratory that my former student, Doug Lenat, was asked to set up his psych project which was this large knowledge base project. I, myself, personally had no relation whatever to MCC. I believe I never was even a consultant for them but I was known to those people. I was known to the MCC Chairman, Bobby Inmati and to Doug Lenat and his boss and some other researchers there.

Report on Expert Systems Building Tools
But what's particularly interesting about this folder is that at the back of it, there's an MCC report from a person called Joseph Scullion which evaluates tools for building expert systems. And that's particularly interesting now because, looking back, a historian or scholar would want to find out what tools were available around the 1984 time-frame when Scullion was writing this report. And what did he say about them? What tools were best? Remember IntelliCorp was one of the companies that had such a tool, KEE. Another company that I formed, TechKnowledge, had a tool like that. And there were many others tools that were out there. So that would be interesting for scholars to know. And that'd probably be interesting to scan, that report on expert system building tools.


Folder 44: MOLGEN 1980--1985

This folder is on MOLGEN and I've spoken about MOLGEN before. That was one of our most famous projects and I've mentioned that it stands for Molecular Genetics and it was a National Science Foundation sponsored project that had many collaborators over the years. It was essentially started by myself and Joshua Lederberg back in the mid 1970's when I finally convinced Lederberg whose Nobel Prize is in genetics to let us do expert system work in the field that he was a world famous expert. And there's a long story behind that which I won't mention here. But we finally succeeded in launching this project called MOLGEN and made very early and quick substantial contributions. This particular folder contains some of the MOLGEN proposals to the National Science Foundation plus a lot of the routine correspondence about administering them the MOLGEN Grant. The name Betty Scott appears from time to time. Betty Scott was the Computer Science Department lead administrator at the time. There's also quite a lot of stuff in here about a three-way discussion between the founders of IntelliGenetics, namely myself, Friedland, Brutlag and Kedes as founders of IntelliGenetics. The only other thing here that's a little bit of a mystery is some discussion about getting one of our collaborators into the budget as a co-investigator. And I have lost track of that in this folder here -- can't find it. Anyway Professor Yanofsky is a collaborator of ours in the later stages of the MOLGEN project and there was some effort here to be get him to be a co-investigator on this project. I don't know if that ever happened.

Letter from Laurence H. Kedes, M.D. to Professor Bruce G. Buchanan
There's a letter here from collaborator Larry Kedes to my co-principal investigator, Bruce Buchanan, or at that time he wasn't co-principal investigator, but at least he was -- well he may have been co- principal investigator by that time, 1985. And it had to do with a minor discussion of did one of the students -- former student Mark Stefik -- get enough credit in an article that Peter Friedland and Larry Kedes wrote for a journal "The Association for Computing Machinery Journal".

Letter from Niels J. Reimers to Mr. Charles H. Herz
The Stanford Office of Technology Licensing which was asked to approve a licensing of the MOLGEN software to IntelliGenetics for commercialization and the National Science Foundation whose rules about intellectual property, ownership of software, at that time were not totally clear. And it wasn't clear whether Stanford owned the rights to the software or did the developers own the rights to the software or was it in the public domain or what actually was the situation with intellectual property rights. So the Office of Technology Licensing person, Niels Reimers, got involved with some correspondence here to the National Science Foundation on that issue.

Stanford Docket S80-63 MOLGEN Project: Molecular Biological Software
This memo is from Bill Osbourne dated January 14th, 1981, and it's called NSF Petition. "Attached is a copy of the NSF waiver which allows Stanford to copyright MOLGEN, and hence to license the software for commercial activities." The background behind this memo is that the company IntelliGenetics was formed by four of us in the summer and early fall of 1980. The first step was to license from Stanford's Office of Technology Licensing the software that we ourselves had developed. That software happened to have been done under NSF sponsorship. And the NSF grants manual says that therefore it's in the public domain unless a waiver is specifically granted. Well, there was no way that IntelliGenetics as a company was going to license this stuff and pay Stanford if it was in the public domain. So what we asked Stanford to do was to ask for a waiver to get it out of the public domain, and then Stanford would own it, then IntelliGenetics could license it from Stanford. And therefore, Stanford could receive money from IntelliGenetics, and that money could be passed on to my research group, one-third of it anyway. Bill Osbourne was working in that office at the time. He was the software guy working in that office. Here's a letter from NSF to Bill Osborne making reference to his letter, and it grants him the waiver. And then there's a copy the relevant part of the grants manual with red lines that indicates what section is being discussed in the waiver. This document is a copy of whatever it is gave Stanford the rights to license.that software. So for us it's either under Commercialization or it's under IntelliGenetics or it's under MOGEN.


Folder 47: Toward the Library of the Future1989

Box 3 - 2005

Folder 1: National Academy of Engineering (NAE) 1986 -1987

This is a folder with a lot of miscellaneous stuff in it about the National Academy, most of which is not important to me -- I mean, it's okay National Academy history but it's not Feigenbaum history.

Edward Feigenbaum selected as Phi Beta Kappa visiting scholar
There's also, oddly enough, at the same time, something which I didn't remember at all. I was selected as Phi Beta Kappa visiting scholar and that means I probably had to go around and give a lot of lectures. So I'm supposed to travel to universities and colleges meeting with students and delivering lectures. I didn't remember doing that but here it is in black and white.

Letter from Alexander Flax to Edward Feigenbaum
There is a letter here dated February 26,1986 signed by Alexander Flax the Home Secretary which informs me that I have been elected as a member of the Academy effective immediately. And that is, of course, one of the most important professional letters of my whole career. So it invites me to the inauguration ceremonies and then it gives me the citation that they're going to use or that was listed for me in the application material that was approved. It says what I am and what I did.


Folder 6: National Productivity Board1986

Letter from Edward Feigenbaum to Freddy Soon
This letter is from the National Productivity Board of the Government of Singapore. This is a letter which says they're moving into a new building and they're establishing a Productivity Hall of Fame and they would like to feature individuals of international repute in the productivity arena who've had something to do with Singapore. And I did. So they asked me to agree to be one of the personalities in the Hall of Fame and I sent them a big photograph and I said thank you very much. I'm glad to be in the Hall of Fame.


Folder 12: National Library of Medicine (1 of 3)1989

Implementing Meta-1: The First Version of the UMLS Metathesaurus
National Library of Medicine, one of three. There's an article here about the Metathesaurus, NLN's Metathesaurus. That was one of the very earliest applications of semantic information to intelligent indexing and retrieval. A very fine piece of work. It had nothing to do with me. It's just that I have an article about it because I was a member of either the Research Advisory Board of the Lister Hill Center for the National Library of Medicine Research or, maybe by that time, on the Board of Trustees of the National Library and so I got this sort of thing.


Folder 13: National Library of Medicine (2 of 3) 1989--1990

Then the next one is the Board of Regents. I was a member of the Board of Regents so I went to Board of Regents meetings. I was a member of the Board for I think a four year term. It's a quite prestigious position. You have to go to Bethesda, Maryland for these board meetings from time to time and very interesting topics get brought up.


Folder 15: National Institute of health Site Visit 1985

This is a folder called "NIH Site Visit, December 2 through 4,1985. It's a renewal of the long lasting SUMEX Project that was started by Joshua Lederberg and myself back in the early seventies. And now at this time which is 1985, I notice that the director of the project, Tom Rindfleisch, mentions that as he's planning the site visit that its principal investigator is Edward Shortliffe. Ted Shortliffe, one of our students who headed the spin-off laboratory that we had which was mentioned earlier, the section on Medical Informatics of the Department of Medicine.
So, by that time, obviously Ted Shortliffe had taken over as SUMEX Computer Project Principal Investigator. And so this was a site visit.

SUMEX-AIM Site Visit Agenda
This site visit agenda. Because we were asking for more money and they routinely come and review us, there were many different talks that were given in this site visit and so there was a kind of a plenary session in which several of us gave orientations and historical overviews.


Folder 16: Allen Newell Award 1993--1994

This folder is the Allen Newell Award. When Allen Newell died at a relatively early age in his early sixties, I think, he died of prostate cancer -- a really tragic death -- Gordon Bell and I decided to use personal funds to establish an Allen Newell Award at the Association for Computing Machinery which is the main computing society. It's the society that also gives out the Turing Award. But they also give out several other awards which have the names of eminent people associated with them. So there is an Allen Newell Award of the Association for Computing Machinery. The funds of Bell and myself were more than matched by the Association Machinery themselves. And a very nice little endowment was set up for the Allen Newell Award which is awarded annually. And I served on the initial Allen Newell Award Committee in which we chose people for this honor. And this folder is a folder with what the award is about and who was nominated at what time and who was considered and so on and who was on the Allen Newell Award Committee.


Folder 17: Allen Newell memorial1993

Allen Newell, 1927-1992 by Edward A. Feigenbaum
Apparently I wrote a tribute to Allen Newell after his death. I wrote it for the National Academy of Engineering. I don't know where the academy published this tribute to Allen Newell. It becomes part of the archival record of the history of the engineering profession I guess in the academy offices. So obviously Allen Newell died in the early '93,1 guess, or maybe in '92. Allen Newell is one of history's greatest computer scientists. He was what you might call an advanced graduate student at the time that I was an early graduate student at Carnegie Institute of Technology, now Carnegie Mellon University Graduate School of Industrial Administration. But when I say advanced, I don't mean he was just simply advanced in being there a few years more than me. He was already a well known person in the field and now is regarded as one of the co-founders of artificial intelligence and winner of just about every award you can win including the National Medal of Science from the President which he received actually on his deathbed and every other award. And he was just a stellar, superb scientist in every dimension. And I guess my tribute would say that. Anyway he's one of the greats of computer science.


Folder 20: NTT Proposal 1988

There's a folder here called "NTT Proposal 1988". The subject of all of this material had completely escaped my mind. Apparently we were interested -- we at the Knowledge Systems Lab, Heuristic Programming Project -- were interested in developing a long term relationship with NTT that started out with having Dr. Okuno with us. I mentioned him before in one of my notes. And then it looks like this idea took off in different directions. I'm not exactly sure how that developed or even what came of it. But this is just a tangle of a relationship with NTT that never quite matured.

complex and confused NTT affairs
Apparently we were interested -- we at the Knowledge Systems Lab, Heuristic Programming Project -- were interested in developing a long term relationship with NTT that started out with having Dr. Okuno with us. I mentioned him before in one of my notes. And then it looks like this idea took off in different directions. First of all, it looks like it got tangled up in some idea that the Asia Pacific Research Center -- at that time called the Northeast Asia Forum under Dan Okimoto -- would kind of handle the Stanford relationship with NTT in some kind of centralized way for Stanford. And there's a letter in here in which I'm asking for clarification of what that means for the future relationship between my project and NTT. Why are w e suddenly adopting the Moscow model for Stanford on this.

Proposal for a Seminar on the Problems of Complex Software-Intensive Systems
There are several copies of a proposal for a seminar that we were going to give related to some seminar with NTT people on either real time software or this particular seminar talks about complex software intensive systems.


Folder 38: Knowledge Systems Laboratory (KSL) Retreat Slides1987

Knowledge Systems Laboratory Retreat Slides
Each year the Knowledge Systems Laboratory would hold a retreat in which we'd take our staff off to some place like Asilomar or some other place like that and be away for a couple days and present material to each other and try to generate new ideas, generally recharge the batteries. This was slides prepared in connection with the Knowledge System Lab retreat in September of'87. But it would appear from these slides that I had already knew a lot about the industrial applications of expert systems and was well on the way to doing the book "The Rise of the Expert Company" because there is a slide in here which sort of announces that it's coming out and it says forthcoming September 1988. So I was really presenting to my own group a lot of what we had found out in doing the research for the book.


Folder 39: Expert Systems (ES): Payoff and Pointers Cambridge1988

Expert Systems: Payoff & Pointers
Expert Systems Payoff and Pointers. There are slides in here for a talk given in Cambridge on January 27 ,1988 and I have no idea who in Cambridge it was for. Who was I doing this for MIT or a company or something. But the slides are almost the same as the ones that I gave at the KSL retreat. So it is an example of my just using slides again and again in different ways to construct new talks on the same subject.


Folder 41: Rise of Expert Company - Slide Set 1990

Rise of Expert Company - Slide Set
This is is a folder here "Rise of the Expert Company Slide Set, 1990". And it represents the state of my thinking in the late 1980's and into the early nineties about the industrialization, commercialization application of expert systems in the world of industrial/commercial work. Many slides are repeated from other slide sets that we've already gone through. Big discussions of the payoff of expert systems, how you can tell -- lots of notes associated with it where some of these notes elaborate the bullets on the slides with some specific case studies that I found in "The Rise of the Expert Company", naming particular companies, naming the dimensions of the benefit. So getting quite specific about the economic gain from using expert systems and where the future lies with reference to some of the earlier work on DENDRAL.


Folder 49: Sumex 1984

Apparently I wrote a tribute to Allen Newell after his death. I wrote it for the National Academy of Engineering. I don't know where the academy published this tribute to Allen Newell. It becomes part of the archival record of the history of the engineering profession I guess in the academy offices. So obviously Allen Newell died in early '93, I guess, or maybe in '92. Allen Newell is one of history's greatest computer scientists. He was what you might call an advanced graduate student at the time that I was an early graduate student at Carnegie Institute of Technology, now Carnegie Mellon University Graduate School of Industrial Administration. But when I say advanced, I don't mean he was just simply advanced in being there a few years more than me. He was already a well known person in the field and now is regarded as one of the co-founders of artificial intelligence and winner of just about every award you can win including the National Medal of Science from the President which he received actually on his deathbed and every other award. And he was just a stellar, superb scientist in every dimension. And I guess my tribute would say that. Anyway he's one of the greats of computer science.


Folder 53: Talks1989

Astem address
ASTEM is, or was (I don't know if it still exists), an institute that was started in Kyoto, Japan for the Mechano Electric or Mechano Electronics Sector. That's what the EM is (actually should be ME), and it sort of includes robotics and other things where mechanical systems and electrical systems are combined. And I was asked to give a speech at ASTEM. So I did and these are the notes. But I used a lot of the same lecture material that I've used before and have already dictated notes about for the archives.

EXPERT SYSTEMS AS POWER TOOLS FOR THE KNOWLEDGE WORKERS' SOCIETY: A QUIET REVOLUTION
This appears to be a sort of generic talk on the Expert System Case Studies that I described in the book "The Rise of the Expert Company". And therefore, it accompanies those generic slides (the transparency stuff) that we put into the archives a while back in the early part of this archiving. So this really does represent a kind of a state-of-the-art in the use of expert systems in industrial work as described in that book round about the late 1980's.

Talk at the University of West Indies
The next one has to do with talks that I gave in India and Trinidad. The dates are actually listed here. This talk originated by my giving it at Aston University in June of 1989. A friend of mine, Fred Crawford, who was a Professor of Electrical Engineering here at Stanford, became Vice Chancellor which means the head officer of Aston University in Britain and asked me to give a speech on my view of computer science and expert systems,knowledge systems of the future which I did. Later on, I was awarded an honorary Doctor of Science Degree from Aston University in the early nineties. Actually, it was by coincidence in the same year and probably in the same week that my daughter was receiving her DPhil at Oxford. So both happened very close to each other. Anyway, this India, Trinidad talk looks like it was a variation on the Aston University talk. And the Trinidad one was University of the West Indies. And all the dates are given. They're all 1989.


Folder 58: Texas Instruments Expert Systems (ES) Tools1984

Letter from Ralph A. Olivia, PH.D. to Mr. Edward A. Feigenbaum
TI is Texas Instruments. This is marketing material and what's particularly interesting here is that they're offering a software package called "TI Personal Consultant Software" that's modeled after one of the original expert system development tools software that we did here at Stanford in my lab. The Emyson system here in the letter is referred to as Stanford University's pioneering work in expert systems development. What I don't know is -- I notice that there's some PC like machines shown in the marketing material and I don't know if those machines are -- well one of them looks like a real PC. The other one could conceivably be a Texas Instruments Lisp machine which was a peculiar kind of personal computer that was based on lisp on a chip. So the computer language lisp was actually put on a chip and TI was selling a computer based on it. I'm not sure that this was the lisp machine but you might think about that.


Folder 68: White House Committee1985

Dick Martin's White House Committee 1985
Actually the document inside says Dick Martin's White House Committee 1985. Who is Dick Martin? Dick Martin is -- or was -- a technologically savvy Navy Captain in military jargon that's an O6 which means the captain is the Captain of the ship and it's the equivalent of a colonel in the Air Force or the Army. So Dick was a pretty high ranking officer and he was very technologically savvy and he introduced early workstations and high speed computer networks onto the aircraft carrier that he was helping design and then ultimately he was the captain of. Eventually Dick left the military service. I don't remember if this 1985 date was before or after he left the military. When he left the military, he went to Carnegie Mellon University at the Software Engineering Institute. This committee envisions that they're going to bring some of this new and fancy technology to the White House Crisis Management Center. So I guess that means sort of the war room where the National Security Council and the President decide what to do in difficult situations. But apparently he was trying to get me to serve on this committee and I really don't know if the committee got set up or whether it didn't. I have very little recollection of this.

Box 4 - 2005

Folder 10: Rise of the Expert Company Fan Mail 1989

This is a collection of complimentary letters that came in about "Rise of the Expert Company". Just in case someone might think that I threw away the non- complimentary letters, I didn't. The general reception for this book was of this kind of "we love it", "thank you very much for making this kind of a presentation", etc. And what surprised me in this file which I didn't remember was there is a letter there from Walter Riston, the CEO of Citibank and there's also a letter from the CEO of American Express among the other people who wrote about the book.


Folder 17: Expert System Book--EAF's Book, with Pam McCorduck undated

The file labeled "Expert System Book, Ed's book with Pam McCorduck" just contains a set of budding stories. I don't know if these were stories that arose after the book came out or before the book came out but there's one story about a General Motors expert system that I'm sure TechKnowledge for them. There's one on Arthur Young's accounting company and there's another one on new expert system for a newspaper pagination. So I was obviously just collecting stories at that point.

Box 5 - 2005

Folder 45: Anne Feibelman--Movie (HPP)--Selling, Advertising 1986

During the peak period of productivity and esprit de corps of the Heuristic Programming Project and Knowledge System Laboratory, the entire group collaborated with the Department of Communication which is the Media Department in Stanford in the making of a movie called Knowledge Engineering. It really turned out to be an excellent movie. The movie director was a student named Annie Feibleman. And we should definitely scan that movie. I'm sure we have it on my shelf of videotapes. And the movie was one of the most important pieces of exposition about expert systems and knowledge engineering because it was easy to understand and it was very visual and it was shown a lot of places. So it was kind of a key PR piece, not only for us here at the laboratory but also for the field. And then Annie submitted it to a film competition for documentary films. It may have been documentary films in the medical computing area, I'm not sure. But anyway it won one of these awards. The organization was CINE, and it won the CINE Eagle Award in 1986 although the award letter was dated May 26,1987. So we were all very proud of this movie. Oh and I should say we worked very, very hard on the movie script and we also worked very hard to get good shoots for Annie so that she had a lot of good material.


Folder 58: Knowledge Systems Laboratory (KSL) Info Packet1985

Addison-Wesley/brochures for books
File named KSL Info Packet 1985 contains a small brochure which, if you read the notes, is referred to as a mini brochure. It was something which would be inexpensive and we could get out large numbers of them fast. When it says mini brochure, that compares with the full brochures that we did as sort of serious documents to get out the message about what our laboratory was doing. We were one of the world's great artificial intelligence laboratories, maybe for a time there the premier world's artificial intelligence laboratory. And there's a cause and effect issue here with regard to our brochure. Maybe because we did such a good job on our brochures, maybe that's why we were so highly visible and important or maybe because we were so important, we were driven to get out these wonderful brochures to make everyone know how great the work was.


Folder 60: Visitors Day--Knowledge Systems Laboratory (KSL) 1985

The letters in here and the notes and all that are dated, roughly speaking, around the end of August, early September of 1985. It seems that the reason for having this open house at the Knowledge Systems Laboratory was the fact that the International Joint Conference on Artificial Intelligence must have held a meeting here on the Stanford campus. Their biannual international meeting must have been at Stanford that year. And since we were going to have a collection of people from around the world, probably we decided to set up a visitor's day in which they could all come over to our place and take a look at what we were doing because they were already here at Stanford. I can even see the visitor's day was on August 17th. And in this file here is an agenda of who's going to give talks. Tom Rindfleisch gave one, Larry Fagan in the medical area, Matt Ginsberg in the logic group and other people are listed here. I see that I am not listed here. So apparently I didn't attend this visitor's day. But anyway, this is the whole record of it.


Folder 61: Display Case - Knowledge Systems Laboratory (KSL) 1982

Display Case - Knowledge Systems Laboratory (KSL) 1982
There's a file folder here labeled "Display Case". The display case that got built and hung up on the wall, was meant for Margaret Jacks Hall. The Computer Science Department was housed in Margaret Jacks Hall around 1982. I was Chairman at the time and remember I said that Harold Cohen, the artist, had been one of my great discoveries and he had done this wonderful work. We decided to install in Margaret Jacks Hall a work of art on the patio. Harold Cohen and his computer program and his wife and his son and several other assistants put together a piece of artwork.
Actually Harold and his program did the piece of artwork and the other people assisted in the making of the artwork which was a gigantic ceramic mural of one of the drawings from an early version of Harold's program Aaron. That was then fired onto ceramics. It was then brought up to Stanford in a vehicle by Harold and Becky and on some nasty, rainy, windy, cold December days in I guess late 1981 because the dedication was a little later -- so the mural went up. And then there were pictures of the mural being constructed and that's what the display case was supposed to handle: the story of the mural and pictures of its construction. So that's what this whole file folder is about. The department is no longer there. It's now something else. It's the Department of English, I think, or something like that in there now. But the mural is still there and it's a great mystery to the people who occupy the building. And I've vowed that someday I'm going to over and tell them the whole story of this, maybe bring Harold with me but, as of yet, I've not done that.

Box 6 - 2005

Folder 6: Heap #7--Stanford in Japan 1986

This folder is called "Stanford in Japan". It represents the correspondence that was done in memos among a number of people who were involved in establishing the Stanford Japan Center. I was involved with that. I was actually in Kyoto as they were driving around searching for appropriate sites. The name of the person who was leading that was a law school professor named Tom Heller. Tom may have been Head of the Overseas Studies Program at the time, I'm not sure. The President, Don Kennedy, was involved in this process. And so I was on the mailing list for this and there are a variety of documents in here which are, I think, important from Stanford's point of view for the history of the Stanford Japan campus. And they're not particularly important from my point of view except that I participated in this process.


Folder 11: Heap #14 - Articles from French Literary Journal1986

MONDiAUX DE L'INTELLIGENCE ARTIFICIELLE
This file is related to a publication in a French literary journal in 1986. The context, of course, was the big discussion going on in the world that we've already talked about in these notes concerning artificial intelligence and its place in the scheme of things and the big boom in AI and the Japanese Fifth Generation Project and the articles in magazines like "Time" magazine and so on, a lot of which are in my archives here. The article was the result of an interview or two interviews done by a relatively distinguished French journalist, and interviewer named Mademoiselle Pasternak and her card is included in this folder. There's a French literary magazine called "La Nouvelle Litteraires". There's a pro and a con about artificial intelligence. The pro is an interview with me and the con is an interview with one of artificial intelligence's most vocal detractors; a professor of philosophy at the University of California Berkeley named Hubert Dreyfus. There is a letter that was sent by the article in my file dated March 31,1986. There's a letter dated April 2 from Carnegie Mellon University written by Herbert Simon, of course, one of the greats in artificial intelligence and Nobel Prize winner and - and also my PhD mentor, in which he discusses the Dreyfuss article. So this is one of the most interesting short discussions of the views of Dreyfuss that I've ever seen and therefore, really deserves to be scanned and put in the archives.


Folder 20: Heap #24--Blackboard 1986

This folder entitled "Blackboard" is a collection of two papers that were written by Penny Nii who, along with Barbara Hayes-Roth, were the two major researchers working on the blackboard expert systems software during the 1980's. So these manuscripts were published in "AI" magazine then the Journal of the American Association of Artificial Intelligence in the summer of 1986 and they were really good papers that became sort of big hits in the "AI Journal" right away, instant successes. And there's a note to myself about maybe writing some sort of a commentary and remembrances. This is work that - that Penny and I started back in 1973 when we were working on the HASP/SIAP Project.
Actually it was the HASP part of that, the Heuristic Adaptive Surveillance Program. And then it became generalized into a widely distributed software package called AGE, which meant Attempt to Generalize. And this document says it's part of a retrospective monograph on the AGE Project currently in preparation. So these two documents are really important summaries of what we were doing.


Folder 32: Heap #51 - article rdf:about edison1986

Labors at Menlo Park
This is a file called "Article about Edison" from May 1986 "Science" magazine. It's a book review of a book called "Edison's Electric Light: Biography of an Invention". I often pull out things out of magazines I read with the dream that I'm going to write a comment or I'm going to write a small note about it or a working paper or something. In this case, there's a note that says that I'm going to do this, compose a memo about Edison and experimental science. What it really means is that I was thinking about my long term view and methodology, strategy, of approaching the science of artificial intelligence which is an experimental strategy of try something out and see if it works. And if it doesn't work, improve it and if that works, improve that a little more and so on and so on and so on. Not a big theoretical thing but rather what's called the cyclical development method or the spiral development method in computer science. And this is a book in which this method of working by Edison is described. So I was going to try to couple what I do in computer science with what Edison did as described in this book.


Folder 44: Computer Industry Project (CIP) Software: MIT and JIPDEC Doc.s 1989--1993

In 1993, as part of my work on the Stanford software industry study, I decided to start with a study of the software industry of Japan, partly because I was scheduled to go to the Stanford Japan Center to teach in the spring of 1993. So I thought, you know, what better way to do this research than to involve the students at the Stanford Japan Center as my research assistants on this project and do a seminar course on it. So I interviewed many people in Japan and this is a folder of material gathered at that time for that study. The study was eventually published in a book called "The Future of Software" MIT Press January 1996. And it's called something like "Japanese Software: Where's the Walkman"

Urgent Proposal: The New Age of Software (Draft)
This particular folder of material contains a particularly interesting draft proposal from one of the big shot committees in Japan -- that is a committee full of big wheels called the Information Industry Committee of the Industrial Structure Council -- called "Urgent Proposal: The New Age of Software". And so some people in Japan knew that the software industry was going to be like the major industry of the world or close to it and were trying to tell the Japanese government that Japan was way behind in this industry. And actually nothing happened on the basis of this very cogent proposal called "Urgent Proposal".


Folder 46: Computer Industry Project (CIP) Japan Interviews undated

During this time period, and in particular, 1992, I launched a software industry study within the confines of the Stanford Computer Industry Project, a joint project of the Business School and the Engineering School sponsored by the Sloan Foundation. It lasted for several years. My first target of study was the Japanese software industry. As with any study, one accumulates lots of different sources of information, lots of articles, scholarly articles, industry magazine articles and economists and so on. This material is that collection so someone who wants to study the software industry and particularly that of Japan at that point in time. I've already described how at the Stanford Japan Center in spring of '93,I conducted a seminar with a number of Stanford students, mostly graduate students but some mix of undergraduates as well. Very successful, wonderful seminar with these students who were helping me do a study of the Japanese software industry. This folder called "CIP Japan Interviews" is exactly the raw material for that. It's probably just a collection of what we did in class and some notes that the students got together. I see the name of Totten in here. That's Bill Totten who's an American who started a Japanese software company called Asisto and gave an interview to the class and was very helpful in our study. And there's another one here called "Key points from the Lotus Interview". We interviewed Lotus Japan and so on and so on. This is the raw material from which that study was built. That's the study that's reported in that book I mentioned, "The Future of Software" MIT Press January 1996.

Box 7 - 2005

Folder 1: Computer Industry Project (CIP) Company Information1994

10 YEARS OF QUOTABLE QUOTES
Looks like PR material from Sun.

Handbook of Artificial Intelligence
At a particular point in time The Handbook of Artificial Intelligence, which was not actually to go to the publisher until about 1980 was in a sufficiently complete condition that the optical printer outfits around here could be organized and bound. We could see what we had as a volume. And for that a few copies were custom bound. And this one was mine. It's written up here "To Ed" and signed by various people - not all the people - who had put the major amounts of energy into doing it, all signed it and dated it August 1979. And this was the state of The Handbook of AI, Volumes I and II at that point. All done in output here at - Stanford. And we sent a bound copy of this to DARPA because they were sponsoring it as kind of a product. And Avron Barr got one of these, I got one of these, DARPA got one of these, and I'm not sure if there were any others.

SGI software fiasco system 5.1
A very detailed email about a software project at Silicon Graphics that failed badly. It was called System 5.1. And this was an internal memo that someone in University of New Mexico or New Mexico Tech wrote about the fiasco that went on in this software project. Now that's an important thing to have kept. It doesn't belong in with the Sun folder but it does relate to the topic of the software industry project or subproject of the computer industry project because one of the main dimensions of corporate life in software is that almost everything is a software fiasco and nobody quite knows why. It's just called the continuing software crisis. And it's as alive today. If you read the Economist of two months ago, maybe one month ago, talks about in an article the management of software, the software crisis. Well that's 2004, December. And there was a software crisis all the way back to the mid 1960's. So to study that is very important. What is the software crisis? And I wrote a lot about it for the Air Force. That was a critical thing for the Air Force because the Air Force was in a continual multi-billion dollar a year software crisis and some of it affected equipment that could kill pilots and reduce the availability of the airplanes and so on. So big topic.


Folder 3: Research Seminar Materials 1993

There's a folder here called "Research Seminar Materials". This was a research seminar that I taught when I was a faculty member at the Stanford Japan Center in the spring quarter of 1993. The students who took this course were Stanford students -- mostly graduate students but some undergraduate students -- and they were trying to understand the Japanese software industry as I was trying to understand it. We were doing essentially research for a paper together. And there was something like fourteen of these students. They did a wonderful job. I took them around as joint interviewers. I would conduct the interview. They would be with me. They'd help to prepare the materials. They'd help to work up the materials after the fact. And this folder contains stuff that was distributed to these students during the course of the quarter.

Endgame
There's a memo here from me to my students called "End Game" which is May 25. And it just says here's what we have left to do because we're all going back to Stanford on June 10. That's the end of the quarter and we don't have much time. So here's, you know, here's what we have to do and here's the layout of it. Here's what we have so far.

Red Paper (1of3)
So then there's a particularly interesting memo in here called "Red Paper 1 of 3" by an American who has been living in Japan for decades. He started a software company in Japan which has been doing very well. He's one of the few Western entrepreneurs who did well in Japan. His name is Bill Totten. And Bill wrote a thing called "Red Paper: Are we Abandoning our Computer Industry to Japan?" And his point of view is that we as Americans or at least at that time in 1993, we simply didn't do enough homework, enough background work to succeed in Japan. And since he succeeded, he knew what the background work that we needed to do was. And that's what the paper was about.


Folder 9: Japan S/W--interviews 1993

Folder is labeled "Japan S/W" which means software "Interview Spring 1993". The study that we did was an interview-based study. And these are interviews with many of the big shots of the software world, of the relatively small-company software world in Japan except here's the first interview was with the General Manager of Nintendo which is not a small company at all.

Just System Interview
Just System is one of the most important small companies in software in Japan. At that time it was a credible competitor for Microsoft. But then Microsoft just clobbered it. And there are many other interviews here. I think with Just Systems, we actually interviewed the president. He and his wife were the cofounders of this company. These are the original interviews.


Folder 11: Newell Speech1982

On Some Future Directions of Artificial Intelligence
This document is by Allen Newell, Carnegie Mellon University dated 20 May, 1982. Newell, of course, is one of the geniuses of the field. He's counted as one of the fore-founders of artificial intelligence and first President of the American Association of AI. A deep friend of mine, he and I studied for our PhD exams together at Carnegie Mellon and a true genius period. Winner of the National Medal of Science and every other thing you could win in computer science and in psychology. He's also the winner of basically what amounts to the Nobel Prize in Psychology, the Distinguished Scientific Contribution Award. Now this is the very detailed outline of a lecture that Newell gave at a special conference held to celebrate the tenth anniversary of the birth of a University of Southern California Research Institute called Information Sciences Institute. We all had a hand in that in that. Newell had been an employee at the Rand Corporation in Southern California. I had been an early consultant there. Herbert Simon had been a consultant there. We all had deep connections into the Rand Corporation's computer science program which they called at the time "Numerical Analysis Program Department". But because of kind of a drifting apart of Rand's interests from computer science, many of the Rand people spun off from Rand into a separate institute and that's Information Science Institute at USC led by Keith Uncapher. Keith was a friend of all of us and he got us to come together to celebrate their tenth anniversary. This is Newell's collection of notes. I remember his speeches being spectacular. It was so insightful, so wonderful. So I must have asked him for his notes because he not only sends me the outline which is printed out but he also sends me the extra handwritten notes that he's made at the back of this.


Folder 14: Science Chicago Talk, DENDRAL1982

AAAS Talk
This represents a little story here. The story is that by 1982, the journal "Science" which is the leading American journal in all of science in terms of readership -- compare with "Nature" in Britain -- had gotten to the point where it was going to publish an article on expert systems which is the field that I co-invented. I was invited to give a talk at a symposium of the American Association for the Advancement of Science (AAAS), the publisher's of "Science" magazine. And I gave a paper at a symposium in honor of one of the famous people of the society, a woman named Minna Rees. And there was a transcription of the talks. So what is present here on paper is a somewhat lightly edited version of what I said in the talk. But about the same time, two local researchers, one of them who was one of our students here, Ted Shortliffe, who graduated - got his MD as well as his PhD, started his own sub-laboratory of our laboratory called"The Section on Medical Informatics of the Department of Medicine", at that time called "The Division of General Internal Medicine". Ted Shortliffe and another friend of all of us from SRI, Dick Duda, put together a similar article dated Sep - 30th of September 1982 and they say that in their letter to Dr. Abelson, he says we have been in touch with Professor Edward Feigenbaum who is preparing at your invitation an article on symbolic computation based on his invited speech at the AAAS meeting in Washington last January. Because of the potential overlap, blah, blah, blah. So they want to coordinate this. Well it ends up that we don't do a coordination of it. It ends up that what we do is just that their paper got published. That's because I just decided that I didn't want to put the time in on redoing this paper. So it's an unpublished paper. Big deal. Their paper was good enough and so that's the one that the big chunk of the scientific world, through the magazine "Science" found out about expert systems.


Folder 35: Misc. Papers, Journals, Articles ("Heap" ?) 1986-1987

A hanging file here which contains one file folder called "International Electronic Mail Service" plus a lot of other things. I just want to comment on what these red dots mean. Every red dot has a number on it. It means that this was, at that time, my filing system which relied on basically accession numbers, each of these is an accession number like the number 22 or the number 24. And then the title of the thing would go into a Word document and then Word's find function would be basically the search engine of the day. That was my way of retrieving information rapidly to find out what number I would go to. So, you know, suppose I wanted to look up something on Fifth Generation computers. Well, one of the keywords for Fifth Generation was 5G and here's a paper by Professor Isd at Keo University called "Application Fields and Social Impacts of the 5G Computers". And I would just find it as number 21. I'd go to the thing called the pile or -- we had different names for it but the pile was one of the names for it. And the pile was just a linear flow of these accession numbers.


Folder 36: Misc. Papers, Journals, Articles ("Heap" ?)1986 - 1987

Letter from Ronald Reagan to Dr. Wun C. Chiou, Sr.
In accession #79 on an envelope, Penny and I participated in some conference called Advances in Intelligent Robotic Systems. And there's a section on space station automation. Anyway somehow President Reagan was told about this and he welcomed the organization to Washington. I am pleased to extend warm greetings to the members of this organization that was holding this conference. And so there's a copy of the letter from Reagan along with the program for this meeting. Didn't even know that existed.

Box 8 - 2005

Box 8 contains file folders of many, many talks that I've given over this particular period which apparently starts in 1987 and goes through the early 2000's. There are not only talks and talk notes in here but there's also view graphs, that is, transparency material which is the way we used to do slides before Powerpoint. There's some 35mm slides that is another way that we used to project slide material. In addition to that, there are other notes in here: like a file that says "here's a possible talk you can give based on the talk that you gave in Singapore". In fact, that one I think says "Singapore 1992" -- so I must have given a talk there and I thought well maybe there's another article or talk which can be mined out of that material. And I tend to do that a lot. I tend to make up a lot of talk notes. Looks like there's an article there. I put it in a file says looks like it'll be a bang-up article. Here's some material for it. Then I never get around to it and I never write the article. So there's probably dozens of those things throughout my archives. But here you'll find a full range of talks from - for example, right here at the front is a talk that I gave in the Netherlands on expert system applications and a talk in 1988 that I gave to the Computer Forum. There's a - a file here on - so when I put together these lectures, I sometimes need additional material like quotes that will be interesting or some other stuff that is useful

Folder 2: Slides (EAF's Talks)1987 and n.d.

Japanese National Fifth Generation Computer Project
There's a whole bunch of things in there about a theme of the early 1980's that I was dealing with which was the Japanese fifth generation project and the American and European responses to it. That was a book that I wrote, 1983 the book came out. It was a big best seller in U.S. and many foreign translations. It happened to hit the crest of a wave of concern in the U.S. about losing out on key industries to the Japanese. You know, we lost the steel industry. We lost the auto industry, blah, blah, blah. Are we going to lose the computer industry to the Japanese? And the fifth generation project in Japan was a very large artificial intelligence project that the Japanese had launched to try to advance state-of-the-art. And I wrote a book about it. And because the book became so popular, there were many lectures to be delivered on the subject. These 35mm slides have lecture material of that kind. They could be composed together in various ways to give various kinds of talks.

Knowledge Engines: The Payoff of Expert Systems
The speech is a copy of view graphs called "Knowledge Engines: The Payoff of Expert Systems". And it's listed as being by me with the initials AF/ST which is the initials of the Air Force Chief Scientist. So I know that I gave this talk somewhere as Air Force Chief Scientist and there are other view graphs in the same folder. They all look like the kind of view graphs that I would be publishing as expert systems began to pay off and subsequent to the publishing of the book "The Rise of the Expert Company" which was in 1988. These were, you know, I was the collector of the great stories about expert system applications and then I would give talks on that. And that's what these viewgraphs were about.


Folder 5: Various Well-Tried Materials and Quotes for Knowledge Systems Lectures (EAF's Talks)1983 - 1987

Rise of the Expert Company Talk
Then there's a [file] based on the book that I did with Pamela McCorduck and Penny Nii called "The Rise of the Expert Company". Once you write a book then there's a whole series of talks that you give based on that book. So here is a sort of a canonical talk called "The Rise of the Expert Company Talk" which is the talk that I would have given many times after that book was published. SW: And when was that book published? EF: Book was published in 1988. 1988, assembled in 9 - '86 and '87.


Folder 6: For General - Expert System Talk (EAF's Talks)1991

From What to How: Automating Knowledge and Its Use
There's another similar folder here called "For Generic Expert System Talk" so if I'm giving a generic expert system talk, I could probably whip one together in a relatively short time by looking in this folder for material.


Folder 11: Japanese Language Slides (Sasari)1988

Japanese Language Slides (SASARI)
There's some Japanese language slides and it says "Sasaki". It means Sasaki Is it - who's a friend of mine in Japan, undoubtedly he did the translation of some of my materials into Japanese to give one of my normal talks on expert systems and the expert system business to Japanese people. LMKB is an acronym which stands for: L is Large, the KB is Knowledge Base I'm blocking on what the M is. So LMKB is a project that we were heading into which was attempting to create very large knowledge bases. I just don't remember what M stands for. But maybe it will show up here. SW: And the time period on that was in the late 1980's, early 1990's. And so you see a lot of things here like LMKB slides and one talked about an IBM talk that I gave on LMKB. M might have been Multipurpose Knowledge Base.


Folder 16: Slides / Knowledge / Toward the Future (EAF's Talks)1988 - 1990

Slides / Knowledge / Toward the Future (EAF's Talks)1988 - 1990
This folder labeled "Slides/Knowledge/Toward the Future" apparently was a talk Igave. These are notes and slides from a talk I gave at a thing called AI Day on June 14, 1990 but presumably AI Day was for the industrial affiliates program. And apparently I was giving a talk sort of being visionary. And the talk was called "Toward the Future", maybe the future of AI.


Folder 20: Feigenbaum Medal - World Congress on Expert Systems - Speech (EAF's Talks)1991

Notes for Feigenbaum Medal Speech
Just especially noteworthy file here is called "Feigenbaum Medal Speech" which was the speech that I gave in accepting the [Feigenbaum medal]. It was December 17th, 1991. It was the first World Congress on Expert Systems held in Orlando, Florida. And that's when the Feigenbaum medal was awarded - the first awarding of it was to me. And that was the speech I gave.

Toward the Library of the Future
Inside the Knowledge Talk folder was a set of notes but the notes include a heavily annotated copy of one of my more important articles called"Toward the Library of the Future", published 1989 in a publication called "Long Range Planning". It says here this article is based on a lecture given at Aston University in November 1986 to mark the opening of the university's new computing suite. So anyway, this article is interesting in two ways. Not only is it important sort of piece of thinking that I did at a certain point in time in my career but as it's annotated, it represents an evolution of that thinking. The annotations represent an evolution of that thinking. This Library of the Future talk had been given earlier in some other form. And I don't remember right now - I can't right on the fly track the development of this article but a lot of people have referenced this article.


Folder 24: Knowledge Talk Supporting Materials (EAF's Talks)1992

The Stanford How Things Work Project
Then there's an acronym here called HTW. The HTW project and slides. HTW stands for How Things Work. So that was an attempt to get artificial intelligence programs to reason about devices and how they work, very similar to what you see in the books that come out from time to time where they show you how they pick several hundred of the world's devices and they try to explain to the ordinary person how that device works. Well this was similarly trying to get a computer understand how a device works and then could answer questions or can do problem solving about how devices work. So that was called the How Things Work project.


Folder 26: Innovative Applications of Artificial Intelligence (IAAI) Summary (EAF's Talks)1993

Innovative Applications of Artificial Intelligence (IAAI) Summary
The folder labeled "IAAI-93 Summer" was given to me by somebody else who did this summary of the talks that were given at the IAAI inter - Innovative Applications of Artificial Intelligence Conference and does not represent my own work.


Folder 27: Japanese Technology Evaluation Center (JTEC) Slides (EAF's Talks)1994

Japanese Technology Evaluation Center Slides
JTEC stands for Japan Technical Evaluation and the C is probably Committee. JTEC was an operation of the National Science Foundation in which teams of American scientists in certain selected fields were sent over to Japan to make assessments of the state-of-the-art of Japanese science and technology in a particular area that they were interested in. They asked me to do one on the expert systems area and I assembled such a group. And so the group then did a report when we came back, presented it back through the chain, eventually to the NSF and then, of course, once you do a piece of work, it's grist for the mill for giving talks on that piece of work. So there's a folder here called "JTEC Slides".


Folder 29: American Association for Artificial Inteligence (AAAI) Panel : Japanese Technology Evaluation Center (JTEC) (EAF's Talks)1992

AAII panel presentation
And then there's another folder here called"IAAI JTEC Slides" which means that a report on what the JTEC Committee found, which I was Chairman of, was undoubtedly presented to the annual conference Innovative Applications of Artificial Intelligence and that these were the slides from that.


Folder 30: UCLA Talk (EAF's Talks)1992

RSL/HPP Overview
These are overview slides for various aspects of the stuff we were doing. There's the knowledge systems lab as a whole, the heuristic programming project as a sub part of it, the How Things Work project, the Engineers Associate Project, a project called KASE (Knowledge Assisted Software Engineering) which had to do with software engineering .And BB for Blackboard project that we were doing at the time. These are slides for all of those projects. The presentations were made by various people of our project at the Stanford Computer Forum. There was a meeting of the industrial affiliates and we gave them presentations of our work of our laboratory.


Folder 31: Expert Systems (ES) Talk: Software Engineer's Version (EAF's Talks)1989 - 1991

Expert Systems (ES) Talk: Software Engineer's Version (EAF's Talks)1989 - 1991
The folder called "Expert Systems Talk Software Engineers' Version", Carmen McClure's conference just means that I was giving a talk on expert systems aimed at the software engineering community. So I had to change the approach of giving the talk and change the examples and change the lingo a little bit to match up with the software engineers.


Folder 33: Vision of Very Large Knowledge Bases (VLKB) (EAF's Talks)1993

Vision of Very Large Knowledge Bases (VLKB)
Earlier I had been talking about LMKB, Large something Knowledge Base. Well this evolved into a thing called VLKB, Very Large Knowledge Bases and here's a thing called the Vision of Very Large Knowledge Bases. So that was a talk I gave somewhere, but this represents a mid 1990's evolution of that same idea.


Folder 35: Japanese Software Industry (JSW) - Where's the Walkman - CS Forum (EAF's Talks)1994

Japan Software Industry (JSW) - Where's teh Walkman - CS Forum
There's a file folder here called "JSW" which stands for "Japanese Software, Where's the Walkman?" And this is a talk given to the computer science department's forum Industrial Affiliates Program in 1994 in February. Now this represents a shift in the nature of my work. In 1990 -- perhaps in late '92 -- I was asked by the Dean of Engineering Jim Gibbons to look into a project that was supposed to be joint between the Engineering School and the Business School, The Sloan Foundation Computer Industry Project. And as far as he could tell, the Engineering School wasn't having any input into this project. So he asked me to go over and check it out. So I did and found that indeed Gibbons was right and, in particular, there was no one working on the software end of the computer industry. There was some work by Business School and Economics Department faculty on the hardware end of computer industry and Silicon Valley. So I formed a thing called the Stanford Software Industry Project as part of that computer industry project. And then in '93, 1993 in the spring, I happened to be signed up to spend the spring quarter at the Stanford-Japan Center in Kyoto, Japan. And during that period, I did my first batch of work on studying the software industry, mainly studying the software industry of Japan using the students at the - who had come over that quarter for the Stanford-Japan Center essentially as research assistants for helping with the study. And the results of that first study were presented for a while only as a talk that I was giving called "Japanese Software, Where's the Walkman"? That resulted actually in a publication which came out in 1996 in January in a book called "The Future of Software" published by MIT press. That's the written version of that talk.


Folder 41: American Association of Artificial Intelligence (AAAI) "Tiger in a Cage" (EAF's Talks)1993

American Association of Artificial Intelligence (AAAI) "Tiger in a Cage" (EAF's Talks)1993
So there's a big fat folder here that consists of a talk I gave which was a very important one for me and probably for the field. It was called "Tiger in a Cage". The tiger was expert system technology. The cage implied: why hadn't this conquered the world? Why hadn't - you know, we expected a big expert system boom and then the world would be taken over by expert systems, you know, like it was taken over by spreadsheets and things like that. But it didn't happen -- or it happened to some extent but not as much as we expected. So this was an invited speech for the American Association of Artificial Intelligence Annual Meeting called "Tiger in a Cage" and it has been referenced in many places. So the actual talk itself is here. The slides which were 35mm view graphs are also in this folder plus lots of supporting material for backing up this talk about examples of expert systems that were out there in the world. And I used some of these examples and didn't use other examples in my talk.

Trilogy Development Group Opens New Headquarters in Austin, Texas - Computer software start-up moves operations from Silicon Valley
Trilogy is a corporation that was founded by I believe it was six undergraduate students at Stanford. They were all juniors and there were five guys and a woman. The guys stopped out the end of their junior year to develop the software that became the Trilogy Sales Builder software which was the first of many different pieces of software that has made Trilogy a multi-hundred-million dollar a year company. And the woman who was involved here stayed on for her senior year and became my teaching assistant for my Expert System course in her senior year. She's the only undergraduate teaching assistant that I have ever had. And she was absolutely terrific. But her parents insisted that she finish her schoolwork. So then she went onto Trilogy, who had moved to Austin, Texas from Palo Alto, and she became something like Vice-President for Marketing. She then spun off her own company called PC Order and took that public during the boom. And then went back to school for her MBA and she got her MBA at Harvard and then out of the MBA at Harvard, she started the first company that's involved with female egg screening for women who want to postpone pregnancy into their late thirties but want their eggs from an earlier period to be available for fertilization later on. And her first name is Christie Jones. "Forbes" magazine in about 1993 roughly had a cover story. Christie Jones was on the cover of "Forbes" and the story itself was about Feigenbaum students, what they had done over the years in terms of building businesses, going out and building businesses. And there's a funny story the editor told me later about how they had a big discussion about who to put on the cover because they had come over to England when I was there and they took pictures of me near Tower Bridge and so there was a big discussion about whether to put me on the cover because the story was about me or Christie because Christie's a really good looking young lady. And, of course, they chose Christie.


Folder 45: Turing Award Speech (EAF's Talks)1980 -1995

How the What Becomes the How Slide
One of the most important sort of summarizing or generalizing speeches that I gave was the speech that I gave them when I accepted the ACM Turing Award. The speech was called "How the What Becomes the How" and it was published later on in one of the ACM journals -- probably the communications of the ACM published it in I guess 1996. This folder has in it the original slides that I used during that talk. It says that it has in it the entire talk but it seems to be missing the printout of that talk seems to be missing pages 1 through 6. Let me just take a look. It looks like it goes through page 6 through 19 and I don't know where pages 1 through 5 are. There are some other notes here about Alan Turing and some other materials that my friend Donald Michie, pioneer of artificial intelligence in Great Britain and a former colleague of Alan Turing at the - the famous Bletchley Park code cracking institution during World War II, had sent me in connection with preparing this talk. And there are a lot of handwritten notes on this file folder also that was presumably done in preparation for the speech.


Folder 46: Heuristic Programming Project (HPP) Reunion (EAF's Talks)2000

Heuristic Programming Project-- Lessons Learned
There's some very interesting notes given for a talk I gave at a reunion of the people who had worked over the many years at the Heuristic Programming Project. I invited them all back. It was kind of a warm and friendly gathering of all the people who had done all this early research on expert systems, particularly in the 1965 through 1980 period. And I gave a talk in which I summarized lessons learned from that period and speculated on lessons to be learned in the period 2000 to 2015. And therefore it was a quite significant summary of where we had come.
These notes seem to peter out in slide number 22. This is a Powerpoint printout because it talks about lesson number one continued but I don't see where the other pages are or what the other slides are and I don't know where those extra pages are. But in any case, I think of this as a pretty good talk, sort of a useful summary of where we had come.


Folder 47: Possible Talk / Article based on 1992 Singapore (EAF's Talks)1992

Singapore Talk On: AI: Promise Fulfilled
There's a batch of materials all bundled under the label "Possible Talk/Article Based on the 1992 Singapore" - what does that mean -- "Singapore"? There's a talk that I gave in Singapore at the opening of some center in Singapore. But this particular bundle of material includes a lot of stuff in which people are coming to grips with where artificial intelligence is right now. Namely 1988, 1990, 1992 -- where - is it and where's it going? What were some of the lessons we learned? And like, for example, here's some notes here - I think this is by me called "AI Promise Fulfilled". And so there's a lot of lecture notes and talks on that theme. I can see here in the Singapore AI conference keynote speech, my Roman numeral I begins with "Visions of the Dreamers". That is what did we really think was going to happen when the field first started? Who was thinking what? What did Newell and Simon think? What did McCarthy think? What did I think and several of the other people who were early on? And then there's a long talk about that. My guess is that these are fairly important notes. And, you know, I myself, would like to look through these particular notes and see what I was thinking around that time.

Box 9 - 2005

Box #9 is from the era that Ed was Chief Scientist for the Air Force which was '94 to '97 and this particular box represents the end of that era so sort of the '97 time period.

Folder 7: Feigenbaum Medal Winner - World Congress on Expert Systems - Lotfi Zadeh1997

Letter from Jay Liebowitz to Professor Lotfi Zadeh regarding Feigenbaum Medal
Then there's a letter from J. Liebowitz, a professor at that time at George Washington University. He's now at the University of Maryland, Baltimore campus. He is the founder of an organization called The World Congress on Expert Systems. When They started in the early nineties, they invented a thing called the Feigenbaum medal. The Feigenbaum medal was a career award medal in the expert system area for career contributions to expert systems. And this letter is sent to Professor Zadeh, at the University of California, Berkeley, EE and Computer Science Department telling Professor Zadeh that he has been selected as the fourth Feigenbaum medal award winner and that this will be presented in Mexico City at the International Conference. And then I was there and I've been there at all of these. But they stopped running this conference and therefore the Feigenbaum medal stopped at some point a few years ago.


Folder 17: USAF Chief Scientist - Definition of the Role1995

Air Force Chief Scientist-Definition of the Role
The Office of the Air Force Chief Scientist. There have been many people who have served in that role since its beginning back at the time of Theodore von Karman back in the immediate post World War II era. There is an informal organization of those people who are still alive and the military assistants who helped them run the office. These were all these military assistants were Colonels based on a decision that was made back at the time of von Karman that the scientists coming into this role would need more clout in getting things done in the Air Force than would be available by using the ordinary Lieutenant Colonel military assistants. So a full Colonel was appointed to be military assistant to the Chief Scientist. So this organization called the Sorcerers and their Apprentices is an organization which meets once a year at dinner. I became Chief Scientist in one of those bolts out of the blue. There was a phone call from Sheila Widnall, Professor of Aero, Astro at MIT. I had known Sheila a little bit prior to that because I was assigned to serve on a committee that she was running for in the American Academy of Arts and Sciences. And Sheila called one day from MIT and mentioned that she was being picked to go for an important job in Washington and that she had a need associated with that. And she explained the Air Force job and that she needed a Chief Scientist of the Air Force. It happened that I didn't need any education on what the Chief Scientist's job was because I had already been briefed on that by one of my colleagues who had served in the role quite a number of years ago. And so I knew what the Chief Scientist did and had always been intrigued with the job. So the only real question for me was gee, could I turn over the management of the laboratory to somebody for a period of time? And did I really want to go to Washington to do this interesting thing and how would my wife and I lead our lives in a bi-coastal way or did she want to go with me? She had her own business here so that was going to be a little difficult for her. But we talked about it that evening and next morning I called Sheila and said I don't have to suggest anyone for this job. I'll do the job. Oh Sheila had said in the phone call, if I was interested in the job myself, let her know. And so I called her the next day and said okay. And she asked for a three-year commitment. Normally it's a two-year commitment but she wanted a three-year commitment and I checked with Stanford and they were willing to bend their rules to do a three-year. They did it for sort of a two year plus one year. That is, they said we'll do it for two years and then ask us again later which I did and they granted three years.

Box 10 - 2005

Folder 1: Blackboard--Airforce--Scientific Advisory Board (SAB) 1998

I've mentioned on previous tapes, our laboratory was one of the two leading laboratories in a software architecture or framework for doing AI programs called the blackboard architecture or the blackboard framework (the other place incidentally was the University of Massachusetts at Amherst under Professor Victor Lesser, one of the original developers of the blackboard concept). The first folder here called "Blackboard Historical" is a set of transparency view graphs that were collected over the years in telling about this subject in talks and also in some of the lectures that I gave in my classes. So it represents a kind of historical collection of tutorial material, background material on what is the blackboard architecture and how do you use it. This folder, however, represents a particular use of this material and that's why originally we found it in the same folder but split the folders into two. In either '97-'98 or '98-'99, the Air Force Scientific Advisory Board set up a study called Information Management for the Warrior, later known as the Battlespace Infosphere when it was given a catchy name. It was done over two summer study sessions. One of the components of the Battlespace Infosphere or Information Management for the Warrior of Information Management and/or Battlespace Infosphere was information fusion where information from many different sources of data and knowledge come together and form a unified hypothesis about what's going on in the environment. What kind of objects are out there? What are they doing? What's their plan of execution and so on? So I realized that we had a framework, namely the blackboard framework, which would be perfect to use in this application, this information fusion application in the Battlespace Infosphere. So I put together a series of notes and papers that would help the other members of this Air Force Scientific Advisory Board panel come to understand these ideas. And so one of these papers is called "Knowledge Technology Roadmap: A Guide to the Perplexed" which gives people a basic introductory tutorial to knowledge based systems and expert systems and knowledge representation and so on. And then there's a lot of email here.

Blackboard Concepts for Data Fusion Applications
This paper was written in the early nineties. It says here it was published in 1993 but I believe it was actually delivered and written earlier by Professor James Llinas, State University of New York at Buffalo who, at that time, was an employee of the Cal Span Corporation. Was written by Llinas and Anthony (Anthony is a guy in the Army) and it was called "Blackboard Concepts for Data Fusion Applications". So Llinas had picked up on my blackboard model or our work at Stanford and had written a paper saying how this would apply to this military community represented by the U.S. Joint Directors of Laboratories. JDL is the directors of various service laboratories like the Navy Research Lab and the various components of the Air Force Research Establishment and so on. So Llinas wrote this great paper and I just included it with this package for the understanding of my Scientific Advisory Board friends.


Folder 2: Blackboard - Historicaln.d.

Blackboard - Historical
I've mentioned on previous tapes, our laboratory was one of the two leading laboratories in a software architecture or framework for doing AI programs called the blackboard architecture or the blackboard framework (the other place incidentally was the University of Massachusetts at Amherst under Professor Victor Lesser, one of the original developers of the blackboard concept). The first folder here called "Blackboard Historical" is a set of transparency view graphs that were collected over the years in telling about this subject in talks and also in some of the lectures that I gave in my classes. So it represents a kind of historical collection of tutorial material, background material on what is the blackboard architecture and how do you use it.


Folder 3: Infrastructure Protection Initiative--Misc. 1998

I'm going to discuss my relationship with a research institute at Stanford called the Center for International Security and Cooperation. It used to be the Center for International Security and Arms Control and in the post Cold War period, they changed the "Arms Control", the "AC" part, to "and Cooperation". This center was formed during the height of the Cold War when people were worried about accidental outbreak - or not so accidental - outbreak of nuclear war. And it was started by a social scientist and a physicist together and it was symbolic that the physical scientists and the social scientists were going to get together to try to wrestle with the difficult issues involved with arms control and nuclear war. The center has been going for a long time. Its director during the eighties, up until the time he went to Washington to be Deputy Secretary of Defense and then later Secretary of Defense, was William Perry. Bill Perry had been, prior to his involvement with CSAC, Defense Director of Research and Engineering at the Pentagon during the Carter Administration. Then he came back and took on this role with CSAC. Then when he got back from his Secretary of Defense Job in 1997, he came back as the Chairman of the Board of this organization. And Bill asked me, while I was still at the Pentagon, if I would align myself with this organization and be a member of their Board of Directors or Board of Trustees or something like that, I don't know exactly what they call it but kind of a governing board. And so I said sure, I would be glad to do it. And then I got involved with some of the problems they were interested in - problems defined for them or they defined for themselves - due to the impetus of two things. One is a whole set of post Cold War issues that were coming up. The other was the problems that Bill Perry brought back with him from the Pentagon or people like myself brought back from Washington as needing sophisticated reasoning, thinking about by both physical scientists and social scientists. One of those problems was information security, cyber war, information warfare. There's a whole set of issues about that. And, of course, we now live with them on a daily basis although we don't see them showing up as nation to nation issues right now but we live with them on a person to person basis where the hackers enter our machines and destroy our files and we have to protect ourselves and we're just individuals. Companies have a much worse than we do; denial of service attacks, trying to take down amazon.com in the middle of the day, things like that. So in the early days of this research, when it was going on in a kind of hush-hush way in the services including the Air Force in Washington and the National Security Agency - places like that - we began some work on that here at Stanford, at CSAC. We had some money supplied by the NSA to support this research. This money will not show up in the records of CSAC because CSAC had an attitude about keeping their list of funders pure and taking money from an intelligence agency was not pure. It was some kind of tainted money. So the money actually was granted to ----. What follows in the box are a set of CSAC papers; CSAC being a scholarly institute has research publications. The research generally is published in a preprint form first by the Center and then later on it might come out in a book or professional journal. This is a series of papers and they largely concerns the question of the protection of the information infrastructure of the United States and other governments. Bill Perry organized some workshops in which he brought together many eminent people to survey this area and lay out the dimensions of the problems and what the solutions might be. And some of these papers, for example, here's one dated March 10th and 11th, 1997, "Workshop on Protecting and Assuring Critical National Infrastructure". That represents a report on what happened at the workshop. There are other papers here like this one by Steve Lukasik's, Stephen J. Lukasik, "Public and Private Roles in the Protection of Critical Infrastructure Dependent Critical Information Dependent Infrastructure" which represent landmark papers in the early study of how to protect giant information systems like the internet or like the information system that controls the electrical grid or the telephone grid. So these papers are not worthy of scanning because they're already CSAC papers and they can be obtained through CSAC but they will be important papers someday. People will look back at this as these are among the first fifty things that were ever written about this. There is a very interesting relationship between CSAC and the Hoover Institute that is on the one hand, highly competitive and on the other hand, highly cooperative. As you know, the Hoover Institute is known for its interest in war related issues but from a very strong stance on the right. CSAC is viewed by many, including myself, as being a counter position on the left. That is it was started by people who couldn't stomach the Hoover Institute's right-wing stance on things but wanted to study things from the ostensibly neutral but actually quite leftish point of view. Now a number of things have happened since those early days. One is that Hoover has become a lot less rightish than it was. It has many of its Senior Fellows and scholars write from a very broad point of view, not necessarily a right-wing point of view. I mean, you still find the Martin Anderson's and the Milton Friedman's coming from the right but you have a lot of other scholars who are coming from the middle and sometimes from the left. The other thing is that members of CSAC and other organizations on campus have been invited to be Senior Fellows at Hoover. So, for example, Bill Perry himself, and several others from CSAC, are Senior Fellows of the Hoover Institute. So there's a lot more joining together. So these days, I think what you'd find is that with the death of Edward Teller, Hoover hardly does anything scientific anymore. They are focusing mostly on politics, economics and political science and other aspects of social science whereas over at CSAC, you will find actual honest to God physicists and people like that, biologists. For example, Michael May has had a long term involvement with CSAC. While Bill Perry was in Washington, Michael May was running CSAC and Michael May is the former Director of Lawrence Livermore Laboratory, famous bomb physicist.

FW: Better format?
In 1998, after I had gotten back from Washington, I affiliated myself with a a research institute called CISAC, the Center for International Security and Arms Control which later changed its name (same acronym) but it became the Center for International Security and Cooperation, chaired by Bill Perry who had also come back in 1997 from being Secretary of Defense. Bill thought information infrastructure protection was a very important thing to study. And he included that among many different infrastructures that were important to study like the Electric Power infrastructure so that those could be protected in - in case of- well as it turned out, I mean, the war on terror but we didn't know there was going to be a war on terror at the time. All we knew was that those systems were really exposed. Perry initiated a number of workshops in which many people from Stanford, the Lawrence Livermore Laboratory in Livermore and other places, gathered together to talk about these infrastructure for protection problems. And this folder contains a variety of correspondence related to infrastructure protection and various kinds of institutes that might be created to study this and various materials that were used in some of our workshops.


Folder 5: Intraspect Folder 1997

Intraspect was a company that was started by a former student of mine, Dr. Peter Friedland, former coworker on our MOLGEN Project, and also another person who was working in our laboratory named Dr. Tom Gruber. They, together with a couple other people, started this Intraspect Company which was going to do some rather advanced kind of software for knowledge management. I was an early investor in this company, I was asked and I sometimes win in these investments. Sometimes I lose. This is one where we lost fairly big and Intraspect was eventually bought by a company called Vignette and Intraspect no longer exists (note: folder is not available in online collection).


Folder 6: Great Debates--within and about Artificial Intelligence (AI) 1998

This is a folder of seven different huge posters which call themselves maps of great debates about topics that are either intensely debated inside the field of artificial intelligence or surround the field of artificial intelligence and are debates about the possibility of artificial intelligence. This material was published by something called Macro View Press which I'm sure is a particular person or two with some post office box in Bainbridge Island, Washington. But it is copyrighted material so anyone who wants to look at it probably has to look at it here in the file. This is excellent material and I was moved at the time to buy these maps and the handbook that leads into these maps. What are they maps of? They're maps of the topics of these debates which have been raging for years and years. They're complex, intellectual and philosophical and technical debates about the possibility of artificial intelligence. For example, can computers think? Can the Turing test determine whether computers can think? Can physical symbol systems think which was a term introduced by physical symbol systems is a term introduced by Allen Newell and Herbert Simon very early on in the field. Can Chinese rooms think? That was a term introduced into the debate by Professor Searle at the University of California, Berkeley. Then connectionist networks are these things called neural networks. Do computers have to be conscious to think? Daniel Dennett, the philosopher from, I believe, Tufts University or Boston University, he wrote sort of the key book on that subject. And finally, are thinking computers mathematically possible because there's a rather complex set of technical arguments that relate to what is knowable in mathematical terms. So I think these are very important summaries of the state of the art at the publication date in 1998. So it's a fairly recent; we're within six or seven years of being up to date on this material and nothing much has changed.


Folder 7: Japanese Software Industry - Where's the Walkman?1994

Japanese Software Industry: Where's the Walkman?
This file here is called "Japanese Software Industry: Where's the Walkman?" It's the title of a paper that I wrote dated 1994. I was the founder of a project called the Stanford Software Industry Project which was a subproject of a larger project, the Stanford Computer Industry Project. And I did a study in 1993 with the help of students whose names are listed here in the first footnote to the paper. I did a study of the Japanese software industry and wrote a paper not only reporting what I saw but also commenting on the the lack of vibrancy and significant components in that industry. And that eventually got published in a book called "The Future of Software" published by MIT press in 19 - January 1996.


Folder 9: A Personal View of Expert Systems (ES) (EAF Paper)1992

A Personal View of Expert Systems: Looking Back and Looking Ahead
A paper of mine which is, in a sense, autobiographical but it also relates directly to the field of expert systems. It's called a "Personal View of Expert Systems: Looking Back and Looking Ahead", dated April of 1992. This happens to be the acceptance speech for a medal that was initiated by the World Congress on expert systems called the Feigenbaum medal and I was the first recipient of the Feigenbaum medal.So this was what I would call one of the first deeply autobiographical retrospectives I did on what work we did here in expert systems and where it came from, the previous state of the art in artificial intelligence and where I thought it was going. And so, in that sense, it's an important paper.


Folder 16: Report on Future of Stanford Japan Center Research 1998

The folder is called "1998 Report on the Future of Stanford Japan Center-Research". That's the research component of the Stanford Japan Center in Kyoto, Japan. The Stanford Japan Center research component operates under the auspices of the Institute for International Studies whereas the Stanford Japan Center education component works under the auspices of the Stanford Overseas Campuses office. So there's been a continual problem for the Institute for International studies otherwise known as IIS to house this organization. They don't know what priority to give to it, how much management to give to it. It seems like an oddball organization only somewhat under their control, not fully under their control but for which they are financially responsible. The Stanford Japan Center Research was founded by its founding director and head officer was Professor Ken-ichi Imai, who's just an absolutely excellent economist, Japanese economist and scholar. So anyway, the question was what was going to be the future of the Stanford Japan Center research component under the IIS following 1958? Imai was getting older and on the verge of retiring. So I was asked by Wally Falcon who was the Director of IIS to chair a review committee on the future of SJCR, and I did. And the report was sent to Wally Falcon but he then wrote me a letter and says as you know, I have passed the mantle of the IIS directorship to David Holloway who is now familiarizing himself with the issues surrounding SJC-R. Well it turns out that when David and his Deputy Director, Chip Blacker and one of his Chief Officers, Nancy Okimoto, when they really did study this report, David called me in to tell me that basically sorry. We have a list of priorities for things to do around here with our money and the future of the Stanford Japan Center research was number six on a list of six. So really we're not very interested in that. And that was kind of the - after much work, that was the death of this report was to just be put in a drawer and nothing happen to it.


Folder 22: Cognitive Systems2002

Research Directions in Cognitive Architecture - DARPA Workshop on Cognitive Systems
So in 2002, the main funder of research in artificial intelligence, DARPA, decided to basically put new life into its Information Processing Techniques Office, the famous office that had produced so much advanced computer science over the years. And they decided to start a new project called Cognitive Systems. Yet another attempt to rename artificial intelligence in a way that would be acceptable. We're doing something new. We're doing cognitive systems. We're not doing artificial intelligence. That's what we did ten years ago. That's what we did twenty years ago. We're doing something new, cognitive systems, but it's really just another way of recasting the same set of ideas for another round of research. And it's underway right now and it's actually a very good project, well funded, high morale, excellent research, under the direction of Dr. Ron Brachman who's director of that office at DARPA and he's also currently president of the American Association for Artificial Intelligence. So things are in good shape. But the program had to be defined and in 2002, DARPA decided to get together a group of more senior people in the field to discuss where this program should go, what kind of a shape it should take, what would be its goals. In preparation for that meeting which was held at Airlie Center in Virginia, outside of Washington, D.C., I had a meeting with Raj Reddy and with Guha on November 15, '02 and we discussed possible themes that the grand challenges in AI might have. And I actually put together some notes for a talk that I gave at that Cognitive Systems Planning Workshop. So I'm not exactly sure if mynotes here preceded the actual workshop or not. For example, the workshop itself turns out is dated October 27,2002. So that my notes with Reddy and Guha were made about three weeks later. But that was the kind of thinking I was doing at that time.


Folder 26: Bell, Gordon letter about Applied Artificial Intelligence (AI) 1987

There's a folder here which we have labeled "Post Chief Scientist Era Talks" meaning some talks that I gave after I came back from Washington. I came back from Washington September 1,1997. Now remember I said in the notes that I became affiliated with the CSAC people. So one of the early things I did on returning was to give the CSAC people a talk on lessons learned. It's what the military always does when they return from a campaign is they analyze what they learned and they tell each other about it. So I decided to give a talk called "Lessons Learned" especially IT lessons learned in the job of Air Force Chief Scientist. And there's a whole series of machine outlined notes for this talk at CSAC. Then apparently I used the reason these notes appear in the same folder as a set of handwritten notes of December 19,1997 is that I must have given this same talk at a Navy group's meeting in December of 1997. The group from the Navy held their meeting at Asilomar on the Monterey coast. And I was invited to meet with the Navy Study Group, NSG or Navy Study Board or something like that. But anyway, I took notes on what other people said including a very well known Air Force retired colonel named Warden who was very well known at the time. He was kind of the guy known for having planned the air war of the first Gulf War. He gave a talk there. I see his name on here. I remember it as being a very good conference and these are the handwritten notes from it.

Letter from Gordon Bell to Professor Edward A. Feigenbaum
This is when Gordon had moved from being Vice President of Research and Engineering for Digital Equipment Corporation to being the Assistant Director of the National Science Foundation in charge of the computer science and engineering end of it which I believe is called CISE now. But he was the Founding Director of that subdivision of the National Science Foundation. Gordon, of course, is one of the great people historically and intellectually in the computer science field. He's one of the leading architects of computer systems in all of computer history. But his interests range very widely and he helped me a lot during the time when I was trying to do real world transition of artificial intelligence ideas from the university laboratory to real applications. And, in fact, served on the Board of Directors of a company I helped to found, TechKnowledge. And he served on the Technology Advisory Board of Sperry Corporation which I chaired when I was the Director of Sperry Corporation. Well when I was doing the book "Rise of the ExpertCompany", I had done all the interviews and it was clear that we had really great ideas but they weren't paying off big in the world. The applications seemed wonderful but, you know, somehow we weren't billionaires. Somehow Microsoft stuff didn't seem wonderful and they were billionaires. So what's going wrong? So I asked Gordon the question if AI is so great as measured by numerous companies and organizations making money and doing new computer applications, why aren't the AI companies making money? And Gordon actually took the time to give me along and thoughtful answer to that question which I have often quoted in talks. And this is the letter. I notice it's very interesting that he copies on this letter Sam Fuller who is also one of the computer industry greats from Digital Equipment Corporation, now a defunct corporation. Lee Heeht was the guy that I hired as President of TechKnowledge. Tony Slocum was the guy that I hired as President of IntelliCorp but later he turned out to be inadequate for that job and he became President of a lisp language sales co-developing company and sales company called Lucid. And Steve Squires, who was at one point I believe a Chief Scientist of the Information Processing Techniques Office of DARPA who went onto become something life Chief Technologist of Hewlett Packard here in Palo Alto. And finally, YT Chien who was the NSF Program Manager therefore working for Gordon, he was the Program Manager for the part that included robotics and artificial intelligence.


Folder 27: Donald Michie Articles 1994--2000

Donald Michie, Professor Emeritus of Machine Intelligence at the University of Edinburgh. Donald is still alive. He lives in Oxford, England. He's retired but he's still working in artificial intelligence and he's one of the greats of the field. And certainly if you think of Alan Turing as being the sort of the godfather of artificial intelligence in both the UK and the US, then Donald Michie is really sort of Britain's father of artificial intelligence, the actual pioneer of real work in artificial intelligence in Britain. Now Donald, it turns out, was a young colleague of Turing's during World War II at Bletchley Park where they were doing the famous secret code breaking that resulted in the breaking of the enigma code and later the breaking of a thing called the fish codes. Donald was instrumental; he was a helper during Turing's work on the breaking of the enigma code but he was a central key figure in the earliest computer programming that took place on what is now regarded as the earliest working electronic computer, the Colossus done by the Bletchley Park Laboratory and unknown to the world for many years thereafter. So Donald in later life, (Donald's approximately 81 or 82 right now) something like five years ago or seven years ago, decided to write a memoir of the - the making of the Colossus machine and the breaking of a particular German high command code that was the motive for increasing the power of the machines that Alan Turing had designed because they could no longer break the codes. So Donald recounts this in a memoir. And the memoir appeared in a journal called "Cryptologia". He sent me a preprint copy of this. First of all, I sent the copy to Whit Diffie at Sun, of Diffie-Hellman coding, the Vice President for Security at Sun. And I also have a note here that says to tell Toole and Bell - Toole is John Toole, the Director of the Computer History Museum and Bell is Gordon Bell who is particularly interested in the work that was done at Bletchley Park. I also have a note to show it to John Mitchell who's one of the co-founders of my department's security laboratory. And I told myself to analyze Michie's material for style of presentation as a memoir because since I was going to be doing some memoir writing, I thought this was pretty darn good and I wanted to understand how he did that.


Folder 35: ICES Keynote Talk - Mexico City1998

ICES '98, Mexico City
ICES stands for International Congress on Expert Systems. That's the international congress that established the Feigenbaum medal. And when it held its 1998 conference in Mexico City, I was there and I was involved in bestowing the Feigenbaum medal on its awardee of that year who I believe was Lotfi Zadeh from the University of California Berkeley. This was the keynote speech that I gave and I chose at that point - this was 1998 so I had just finished being Air Force Chief Scientist. So I just described what the job was like, what I did on the job and what results came out of it, plus also describing some expert system applications that I found going on in the various military services.

Box 11 - 2005

Folder 8: Joshua Lederberg1998

ACM Allen Newell Award, Joshua Lederberg citation
I wrote at the time that Joshua Lederberg won the ACM Allen Newell Award. I was on the Allen Newell Award Committee. In fact, the Allen Newell Award was co-founded by Gordon Bell and myself, with the assistance of the Association for Computing Machinery, ACM. I won't go into what the award was supposed to honor but Lederberg was one of the early awardees. And I wrote the short citation and the long citation for his winning that award.

Using Latenet Semantic Indexing for LIterature Based Discovery
It looks like he was sending this to me in the late nineties because the paper itself is dated 1998 and it discusses how a concept called Latent Semantic Indexing can be used for enhancing the intelligence of information retrieval. And Lederberg says in handwriting "Re: New Project" So I wonder what he means by the new project. What new project was I heading into or did Josh think I was heading into when he sent this paper. He says this paper "belabors the obvious but is worth noting".


Folder 9: EAF's Talks, March to April (EAF Talk)1994

Reinventing AI Applications (IEEE Applications Conference, San Antonio, March 2, 1994
It was delivered to the annual IEEE which is the Institute of Electrical and Electronic Engineers Society. IEEE Applications conference for artificial intelligence. And the title of the talk is "Reinventing AI Applications". And so this is a talk about AI applications, where they came from, what stuff in AI do they use and what are the problems, where are they going? This, it occurs to me, was probably a very important talk and probably I should have gotten around to writing up that talk and didn't unfortunately. That is a kind of a thematic thing in my career,spending a lot of time writing up notes for the talk, very elaborate notes, and then once that's done, I lose energy and don't actually write up the thing for publication.

The Software Industry
Remember that I had added a research issue to my agenda in 1992, 3 and 4, which was to study the software industry. First I studied the software industry of Japan. Then I studied the software industry in the U.S., somewhat. And then together with two people that I hired for this project, Avron Barr, an old colleague who had coauthored "The Handbook of Artificial Intelligence" with me and his wife Shirley Tessler, another computer science person, we -- the trio of myself, Avron and Shirley -- gave a talk at the annual event of the Stanford Computer Industry Project in April of 1994 and this was a review of the software industry in U.S. and Japan.


Folder 16: Draft of Book on Advanced Architectures Projectn.d.

The Advanced Architectures Project
The Advanced Architectures Project is what you might call the high speed AI computer architecture project that I launched here at the Knowledge Systems Lab around, roughly speaking, 1984, '85. And it was part of a sort of a big DARPA project on advanced artificial intelligence techniques that were supposed to keep the U.S. competitive with the Japanese during the time the Japanese were doing the Japanese Fifth Generation Project. This was supposed to be kind of the American response to that. We did a lot of work in this laboratory on that subject. Many interesting papers and sub-projects within the advanced architecture project. First of all, we had to write up all of that research at the end of the government contract but it looked like it would then become material for a book on all the work we had done on advanced architectures. And someone named Brad and I don't remember who that was, was obviously helping me collect that material and put together a draft which we're calling here version two of such a book on the Advanced Architectures Project. It was a goal of mine to get that work out into book form. The work had been published variously in journals and proceedings over the years but hadn't appeared as a book. So this was going to be the book. And this is one of those examples of material for which there was good intentions. There was good material. It just never turned into a book. To turn a draft like this into a real book takes several months of sustained hard work and I just didn't put it in. And this therefore becomes a folder scanned in the archives, not a real book sitting on shelves.

Box 12 - 2005

Folder 12: Challenges in Software Systems Projects: A Report to the Air force1997

Challenges in Software Systems Projects: A Report to the Air Force
The report is called "Challenges in Software Systems Projects: A Report to the Air Force". It represents a kind of final report for a big project that the Secretary of the Air Force had asked me to do when I first got to the Office of Chief Scientist of the Air Force. Her name was Sheila Widnall, Professor of Aeronautics and Astronautics at MIT and Secretary of the Air Force during the Clinton Administration at the time that I was there. She was the one who invited me to be the Chief Scientist of the Air Force. Why did she pick an artificial intelligence scientist and a software scientist rather than the more typical aeronautical engineer or physicist or munitions specialist or somebody like that? Because she felt that the Air Force's most significant problems lay in the complexity and difficulty to build of their various immense software projects. So she wanted me to get my mind wrapped around what the problems were there. And this report was one that was produced over a relatively long period of time. It's in a particular style that the military likes which is a kind of Power Point of slide on one side of the sheet and the story that goes along with that Power Point slide on the other side of the sheet. And that constitutes a kind of a small book for them to read. That's in lieu of having to read, you know, full written material. And this particular thing ends with a to-do list. I went around giving this talk to numerous people at various high levels in the services and the Defense Department and the talk always ended with okay and well now you've heard the talk, now what do you do about it? And here's the to-do list of what your service or your department in the DOD should do about it. So this was one of the most important things that came out of my being Chief Scientist.the subject under discussion here was one of the critical items that I was supposed to be worrying about as Air Force Chief Scientist and as a Scientific Advisory Board member - can't date these notes but it's somewhere in the mid 1990's - mid to late 1990's. The subject was Air Force software and not Air Force alone but also the Defense Department acquisition of software. There was a debate going on about how we were going to get the DOD out of its proclivity to produce custom made software for every application and into the habit of using commercial software which is called COTS. You'll see the abbreviation used on these sheets, COTS, Commercial Off The Shelf Software as opposed to custom software. And also these sheets list some features of why the DOD production and maintenance of its software was so flawed in past years.


Folder 14: META--Intrusion Detection 1998

The file labeled "Meta Intrusion Detection" represents early work on a research project that I started with two undergraduate students, one of whom was David Brunner, who was also a coauthor with me of the book on the Japanese Entrepreneur. But David was a computer science student and this was one of the research projects he was working on. I was using as a mold in which to do artificial intelligence on intrusion detection in computer systems. I was using as a mold the blackboard framework that I had used for detecting very quiet submarines in the ocean off the California coast two decades before that. And David and another person made a stab at it and there are some view graphs here that we made up for a talk that was given at the CSAC Science and Technology Seminar which is a lunch that takes place once a week and people who are doing various research things on the science side -- not the social science side but the physical science and information science of CSAC -- give talks at this Tuesday workshop. And this was a collection of materials related to that talk (note: not available in online papers).


Folder 15: USAF and DoD Software--Jim Hendler 1997

Jim Hendler is a good friend of mine. He is a younger scientist in the artificial intelligence area. I mean, if I'm really an old scientist, he's a kind of middle aged scientist. And at some point in his career, which was around the mid 1990's, he decided to become somewhat of a specialist in applications of computer science and artificial intelligence to the military. So he went to this program which is organized to help well known scientists do exactly that. It's run by the Institute for Defense Analyses, IDA, in the Northern Virginia area. And Jim became part of that program. And as part of that program, he had to basically write a term paper. And the term paper he wrote was more than a term paper - it was like a whole year's paper - was called "Software for the Warfighter: Information Push and Pull on the Digital Battlefield". And Jim came to me in the Pentagon for advice on this paper. And I went around and around with him on this paper. Was pretty good paper.


Folder 16: Air Force Expert Systems (ES) Applications 1995

Another one of the things that Secretary Widnall wanted me to look at besides Air Force software in general, was the application of expert system techniques to Air Force problems. So we got people in the Air Force's laboratory, the one focused on information technology, Air Force Rome Laboratory, Rome, New York, to think about that and one of their better people, Nort Fowler wrote a paper on it for me which then I was able to brief the Secretary of the Air Force on it. And it related to Air Force applications of artificial intelligence which were mostly expert system applications. (Note: folder not available in online papers).


Folder 17: Principal Investigator (PI) in a box 1998

This folder relates to a NASA project. I was not directly involved with it but it's a quite interesting project. The traditional way of doing science in space is to send a science astronaut up on board the shuttle and the science astronaut would do science experiments on the shuttle. But there's only a very small number of science astronauts that can be trained. So what if a science astronaut brought up with him some expert systems which were expert at running other experiments so that instead of sending the principal investigator along which would be very expensive and very difficult, you would send a computer program which was an expert system which would be that principal investigator on board. That's PI in a box. And that was done by NASA and this is a set of documents about that project.


Folder 18: What is Artificial Intelligence (AI) ? Talk for USAF (EAF Talk)n.d.

EAF talk to Airforce personnel rdf:about What is AI?
Subtitled "A Talk for the U.S. Air Force". Now what does that mean? This is what you might call a package talk. Every specialist in every field has some one talk that they can give for a general audience to make a general audience understand what that field is all about. And my field is artificial intelligence. So in this particular era, which was the time I was Chief Scientist, my package talk was what is artificial intelligence. And I was going around giving that talk to a lot of Air Force personnel. Now when I say personnel, the reason I'm not more specific is because it was everybody from going to a war college where they were training colonels to perhaps a super war college where they were training generals or even to lieutenants and captains somewhere at the Air Force Academy or wherever I was asked to go and give a talk and if somebody wanted to know well what is artificial intelligence anyway? Dr. Feigenbaum, you're a specialist. Tell us about it. This would be the talk that I would give.


Folder 19: Naval Studies Board 1999

This folder is called "Naval Studies Board" and it represents the read-ahead agenda material and so on for a study that was done by the Naval Studies Board of the National Research Council which is a subdivision of the National Academy of Science Engineering. And in the late 1990's when this study was being done the military was full of talk about a so-called revolution in military affairs, RMA, similar to what had happened when the English invented the longbow and wiped out the French at Agincourt and the invention of the airplane and its use in warfare and things like that. Well information network related topics for warfare were seen as a revolution in military affairs. So you could transmit anything from anywhere to anywhere else. You had high bandwidth. You could do video images. You could have satellites beaming down images about, you know, what's on the ground. You could have unmanned aerial vehicles transmitting television directly to the Pentagon. You know, all of that. That was Information Centric Warfare or as it was called then Network Centric Warfare. I think they're getting away from the use of that terminology now because it was much overused for a time. But the Navy loved the title of Network Centric Warfare. So I think it was the Chief of Naval Operations, CNO, who decided he really wanted to know what in the hell that was and is this for real? Should he pay attention to these three star and two star generals and colonels and so on, who we're talking about this or is it all a lot of smoke? So he got the National Research Council to do a study of what could Network Centric Warfare do for the Navy. So that study went on for quite a while and had a very distinguished panel that met a number of times about it. And, you know, in the end, it turned out really nothing at all basically, in my view, a useless study.


Folder 29: High Performance Knowledge bases (HPKB)--Intro Slides 1997

High Performance Knowledge Bases (HPKB) was a project sponsored by DARPA, the Defense Advanced Research Projects Agency, the sponsor of most of the great things in computer science research over the years, including all the at least half the work we did here at Stanford and also of the people who sponsored the ARPANET which grew into be in the internet and so on. Anyway when I was in Washington, I was trying to give some impetus with financial support if I could get it, for what I saw as the key problem in making artificial intelligence systems high performance, making them very real in the world. And that was that they would acquire or could be given a very large knowledge base of knowledge about the real world because the key principle in AI which I helped to formulate was called the knowledge principle which is that the artificial intelligence programs become more highly competent the more knowledge they contain about the particular domain of the world's work that these programs are attending to. So what you need is more knowledge to make it short and what you need in these AI programs is more knowledge of the world. DARPA was planning a project of this type to follow some of its other - as it was shifting its focus in artificial intelligence; it was planning a new project. And I got the Air Force to contribute a million dollars of research money toward this problem of large knowledge bases which became labeled High Performance Knowledge Bases.
DARPA put in about nine million, Air Force put in one million. That became a national project for a short time like maybe two or three years before it got transformed into another project: Rapid Knowledge Formation, RKF. DARPA shifts its emphasis as it needs to to convince DARPA directors and congressional staff to support different projects. And as it became clear that the issue in getting large knowledge bases was rapidly acquiring new knowledge, a kind of machine learning, then they changed the name of the project to Rapid Knowledge Formation, RKF. (Note: folder not available in online collection).


Folder 33: The Information Operations, Warfare and Assurance Initiative 1998

There's a report here from the Lawrence Livermore Laboratory, otherwise known as Livermore. That's the place where we design nuclear bombs, hydrogen bombs, but during the Clinton Administration, there was no bomb testing going on and therefore no substantial bomb design going on. So Lawrence Livermore Laboratory turned their attention to many other problems of national security of which information warfare was one of those. And a key project that they were doing for the government in that area was called IOWA, Information Operations Warfare and Assurance Initiative. (Note: folder not available in online collection).

Box 13 - 2005

Folder 4: New World Vistas - USAF Scientific Advisory Board (SAB) Publication1995

Information Technology Volume Abstract
The file folder we're looking at is called "New World Vistas". This represents three paper bound volumes or three paper bound sub-volumes of the massive work that the Scientific Advisory Board of the Air Force turned out in the mid 1990's. The study called "New World Vistas" where they were - the board was attempting to project twenty or thirty years out as to what would be important problems to technology oriented initiatives and re - and applications of the Air Force. But the three he - represented here are the Executive Summary Volume, foil - and in - and followed here by two volumes on Information Science and Technology. One is called The Information Technology Volume and that was a sub-panel chaired by me and The Information Applications Volume which was chaired by somebody else. These three volumes are also included in the bound set that's in my library. Anyone who wants the online version of this should go to a website of the Air Force Scientific Advisory Board or SAB. If someone would go to a search engine and search with the key words Air Force Scientific Advisory Board and then scan the site for previous publications, they'll find New World Vistas in there somewhere. Or maybe just search the Google for New World Vistas Air Force.


Folder 11: Battlespace Infosphere (BI) Working Material 1999

This particular file is Harold Cohen. Harold is a retired professor from University of California, San Diego, Visual Arts Department. He was the founding Chairman of the Visual Arts Department. He was quite a well known British artist in his younger days before he was invited to start that department at UCSD at its beginning. Harold is one of the most famous artists who adopted the computer as a working tool but he didn't do it because the computer was great at drawing beautiful formulas. He did it because he wanted to create an expert system of an artist's behavior. What artist? His own behavior. So since 1972 or 3,1973 he's been doing this. So it's over thirty years and his program is called Aaron. There is a book written about Harold Cohen by Pamela McCorduck. It's called "Aaron's Code". Now why is that interesting in my life? Because I was sort of the one who discovered Harold Cohen. He had written a request for research funding to the National Science Foundation to create this expert system which did creative visual arts. And they didn't know what to do with it because they've never received anything like that before ever. So they sent it to me and I looked at it. It seemed very intelligent to me and I just decided I was going down to San Diego in the next couple weeks for a conference. I was going to check this guy out to see if he was for real. And turned out he was for real and I said, Harold, you don't want this $50,000 from the National Science Foundation. That's not going to get you anything. What you really want to do is come up to Stanford on sabbatical and we'll let you use our machines and we'll teach you how to do this whole thing. And he did that - '73-'74 - and then he stayed for a second year on sabbatical - '74-'75 - then he went back to San Diego and the rest is history.


Folder 15: Scientific Advisory Board (SAB)--General Info 1997

What's the relationship between being the Chief Scientist and participating in the Scientific Advisory Board? The Chief Scientist is an employee of the Air Force or at least a surrogate employee of the Air Force, operating on loan from a university or a non-profit. The surrogates are called EPA's. It's a way that the government can borrow people from universities and non-profits and pay their salary. But in all other ways, IPA's operate as government employees and the Chief Scientist operates as, in particular, as an employee who reports directly to the Chief of Staff of the Air Force. So the Chief of Staff is a Four Star General and he has the so-called general staff of the Air Force which consists of a lot of Two Star and Three Star Generals. And the Chief Scientist of the Air Force has the virtual but not actual rank of a Three Star General reporting to a Four Star General. The Scientific Advisory Board is a civilian organization which does not work for the Air Force. It advises the Chief of Staff and the Secretary of the Air Force but they're regarded as an independent advisory group under the chairmanship of some appointed chairmen. They have their own staff. The Scientific Advisory Office in the Pentagon and their chairman is not an employee of the Pentagon. Their chairman either works for a university or works for a national laboratory. In the case of the particular chairman right now, Dan Hastings, he works for MIT. He's Chairman of the Aeronautics and Astronautics Department. He happens to be the guy who was my successor as Air Force Chief Scientist. So he had a lot of background but it was only after he stopped being Air Force Chief Scientist that they could get around to appointing him as - as Chairman of the Scientific Advisory Board. Scientific Advisory Board has about - at the time I was there, it had about sixty members drawn from the many sciences that participate in Air Force work. There was a question - well a question continues to arise and it arose during my tenure as Chief Scientist as to whether the Air Force Chief Scientist should simply be the Chairman of the Scientific Advisory Board and have the Scientific Advisory Board re - report to him or her. And that was vigorously opposed by the then Chairman of the Scientific Advisory Board and everyone else. I was the only one holding that opinion because the idea was that if the Scientific Advisory Board offers an opinion which it needs to be free to offer an op - any kind of opinion counter to the - the currents in the Pentagon or the thoughts of the Chief or the Secretary without any feeling that there could be retribution. It has to be known and perceived by the world as a totally independent organization and wouldn't be if it reported to the Chief Scientist because the Chief Scientist has a direct reporting relationship to the Chief of Staff.


Folder 21: United States Air Force (USAF) Software Talk Notes (EAF Talk)1997

Challenges in Software Systems Projects
There are three folders here labeled "U.S. Air Force, USAF Software Talk (Summer of'97)". And another one similarly titled "U.S. Air Force Software Talk Notes". And finally there's one called U.S. Air Force Software Project, '94 through '97. Well the project was the one I referred to earlier that the Secretary of the Air Force had asked for, wanting to understand how the Air Force did software systems, what was going wrong in the process, what could be improved in the process and so on.
And that became the theme of the main talk I gave as my term in office was coming to an end, I would go around and present the results of my study to various people ranging from the Secretary herself and various of the generals with many stars on their shoulders, all the way down to Colonels and also civilians working in the Pentagon in the areas having to do with system development. The talk ended up being called "It's a Software First World". And the reason the talk was labeled that is because nobody in the Air Force really believes that. They really believe the Air Force is a bunch of hard stuff. It's like an F22 a new fighter aircraft is a titanium shell full of stuff. Well it isn't. It's really a bucket of software enclosed by titanium. They don't understand that yet: that almost all the functionality of modern devices comes from its software, not its hardware. And that was getting them into a lot of trouble. That's why they were having such difficulty with projects that were software intensive. So this talk was called "It's a Software First World" and these are various notes and presentation materials, various stories that I would interleave about this just trying to make that case credible to the highest levels of the Air Force and the DOD.

United States Air Force (USAF) Software Talk Notes (EAF Talk)1997
During the time I was at the Air Force, I was working on and ultimately completed a briefing on the software problems of the Air Force. Occasionally I would give briefings or talks on the subject. The particular thing I have is called AFSW paper. SW is software. It's not a paper though. It's a series of view graphs with notes attached to them. And it's marked as coming from the U.S Air Force Chief Scientist and it has all the earmarks in terms of the words that are used of words that I would use like the phrase "hardware stuff or "rule oriented". Those are kind of- or "stuff oriented form over function" are things that I would say. I can imagine though that these slides were prepared by my briefing to two younger officers who were assisting me in producing this document. One of them was Shawn Butler of the Army and I can't remember the name of the other one. And I can imagine that they helped me produce these slides and turned my words into bullet oriented view graphs.


Folder 23: EAF's Software Notes for Software Report1995

Innovation and Naval Technology 2020
This was the Asilomar Conference I was mentioning earlier. It's dated the first through the fourth of December 1997. I think this is the same conference. For example, the keynote speech is being given by Admiral Cebrowski who is the main intellectual force in the Navy behind Network Centric Warfare. So it's probably the same conference I was referring to that this Navy study group before.


Folder 29: Information Security--Clinton Administration 1998

We're looking at a report which is the report of the President's Commission on Critical Infrastructure Protection; report is called "Critical Foundations". And then a whole bunch of materials in a folder related to that called "Information Security in the Clinton Administration". Now, while I was there, during Clinton Administration time, many in Washington, including myself, were pushing very hard on the White House to pay attention to the problem of information security, information warfare, the growth of the internet. It was a very vulnerable structure and other things were connected to it like - like the - the control of electric power and so on. A lot of things had to be protected and nobody was interested at all in that in the government. So eventually the Clinton Administration was persuaded to appoint one of these blue ribbon panels headed by a retired general, General Marsh whose report I just mentioned, the report of the President's Commission. And as a result of that, the president eventually on May 22,1998, issued a Presidential Decision Directive which sometimes are secret and, in this case, were made public. Directives number 62 and 63 - President Clinton today ordered the strengthening of the nation's defenses against emerging, unconventional threats to the United States, terrorist acts, use of weapons of mass destruction, assaults on our critical infrastructures and cyber attacks. And so first of all, you can see that the terrorism issue was very much in the minds of everyone and then the person who was a major policy individual in all of this was Richard Clarke who was a member of the Clinton Administration, we used to communicate with a lot concerning this issue. And then eventually he briefed Condoleezza Rice about this when she took over the National Security Director's office and then he claims she did nothing about it.

Box 14 - 2005

Folder 16: Scientific Advisory Board (SAB) Study Group: Much Ado about C2 - Command and Control 2000

The military services have a never ending issue with Command and Control. Command and Control is the information part of being a military officer. It's the information integration and decision making part of being a military officer, as opposed to the part which points the gun and shoots it. This part decides what's going on out there in the battlefield and what's the tactical issues and how to issue commands to the right people to do the right thing and how do we know that the right thing was done and so on. Anyway, it's a never ending issue and during all the years that I was Chief Scientist plus the years that followed, the Air Force continued to worry about Command and Control as being its central issue. They were not particularly worried about were they training fighter pilots well or did they have the right munitions. Those issues were being handled well by the Air Force. What they felt they weren't handling well was the information integration and coupling into decision making using modern electronic computers and communications. So, for example, after me, there followed a person who was in the satellite business, Dan Hastings. And after him, someone who was a specialist in Command and Control from, I believe, from Lincoln Laboratory or MITRE Corporation and then after him, another person in the Systems Architecture and Command and Control world, Alex Levis, and it wasn't until this the Chief Scientist that we have right now in 2004 and 5 that the Air Force got back to someone who was strictly in aeronautics. (Note: folder may not be scanned in online papers).


Folder 26: DARPA Artificial Intelligence (AI) Advisory Board 1999--2000

DARPA like everyone else in Washington was affected by or motivated by the concerns about information security and warfare, the vulnerability of information systems that I've described in previous memos on this tape. One of the ir initiatives then at DARPA which was the advanced research place of the DOD was to start an office in Information Assurance and Survivability, IA&S. And I think it may have been the Information Systems Office and the Information Technology Office in cooperation. They had this program called IA&S. And I was asked to serve on their Advisory Committee. There was a letter of invitation dated December 23, 1999. We actually met at DARPA, gave them some advice and after the first meeting of this committee hearing everything that DARPA was doing, I believe - if I remember right - this committee never met again.

Box 15 - 2005

Folder 1: AF Chief Scientist Office1995

AF Chief Scientist Office1995
First thing in this folder is a list of offices Pentagon telephone numbers for two groups of important people at the Pentagon for the Air Force. One is called the Air Staff which is the military people who report to the Chief of Staff and his name is at the top of the list. That's General Fogleman. And then all the other people who reported to him including myself. And then the other list is all the people who report to the Secretary of the Air Force whose name is at the top of the list, Dr. Widnall. And then the next page lists a variety of other people on the Air Staff. And I don't know what the context of this other list was but it provides a context for who was around in what offices during that time.

Chief Scientist's Visits, Sept. 1 1994 to April 21, 1995
A list called "Chief Scientist Visits" September 1, 1994 which was the date I started to April 21,1995 and I don't know why I cut it off at that time but I probably was keeping this list on my old Macintosh that I brought with me to the Pentagon. But it shows the very great variety of contacts and roles of the Air Force Chief Scientist, you know, visiting places in the Air Force, visiting places that work for the Air Force, visiting universities, visiting key people in the government like the Undersecretary for Technology and the Commerce Department, the Director of the National Institute of Standards and Technology is listed here and many companies, Air Force Laboratories and so on. This lists about thirty visits or forty visits. And that's only for the first three-quarters of a year when I was there. So I was really moving around a lot.

Modeling and Simulation and the Revolution in Military Affairs
This is from a folder called "Air Force Chief Scientist's Office". It's a memorandum dated September 5,1995 from Dr. Andy Marshall whose title was Director of Net Assessment for the Secretary of Defense. It's a peculiar title but what it really meant was that Andy was running a small think tank within the Pentagon. Andy himself is a brilliant military analyst. And this was the beginning of my introduction to a concept that was key in framing defense thinking around the middle of the 1990's called RMA, stands for Revolution in Military Affairs. This document on the one hand, talks about modeling and simulation of the revolution but then there is a broader discussion of what the RMA is that's attached to this document.


Folder 2: Correspondence1995 - 1997

Invitee list for EAF 60th birthday party
There's an email here from my wife, Penny Nii, listing people invited to my sixtieth birthday party, January 27,1996. So the reason this is interesting is it gives a context for what kind of people were I was very close to that would sort of make it to my list a relatively small number to invite to this important event in my life.

Jane's US Information Warfare E-Letter
There's an article dated 1997,14th July, 1997 from the publishing company that publishes about military equipment and events, Jane's. And this article is called "Information-In-War", Information In War versus Information Warfare. And there was some discussion about about what's the difference between the use of information during war and a particular kind of process outside of the use in normal military activities called Information Warfare. That is how we can go to war against an enemy's information systems. So this explains the distinction.

The ABL, briefly
This concerns ABL, the Airborne Laser, and it's addressed to Dr. Widnall, the Secretary of the Air Force. This was just a very brief memo to bring her up to date on something I was doing for the Chief of Staff which is evaluating a project that had only just started but had the potential to be an enormously expensive and highly risky Air Force project, the Airborne Laser Aircraft. And this was just an update to Secretary Widnall and what was going on in my looking into this for the Chief.


Folder 11: National Defense University (1 of 2) 1994

The National Defense University actually is a university. It exists on a small military base on the Potomac River near the south waterfront of Washington. And it has departments that are directly related to military issues and problems that need study and for which military officers need education and where research is needed. One of those departments was established during the time that I was in Washington to handle the subject that was becoming one of great concern and a great deal of discussion which was information warfare, both offensive and defensive information warfare. The National Defense University had a department essentially that I don't remember exactly if it was called Department of Information Warfare but it may have been, in which these subjects were covered, taught. There were seminars, discussions and courses. They asked me if I would be an adjunct professor during the time that I was in Washington and I agreed. So I would go over there from time to time and attend classes, meetings, discussions, etc., on this subject of information warfare. Actually the university was within a very short walking distance of- where I was living in Washington which was an apartment building behind the arena stage.


Folder 15: Department of Defense Enterprise Computing 1995

This is a folder entitled "DOD Enterprise Computing". Now enterprise means when we talk about corporate computing. This is information technology infrastructure for the Defense Department, how to manage all of their information. So what we decided to scan was some of my notes which apparently I was attending briefings on this in the Defense Department since I was known as a computer science and information technology person. These are dated October 19,1995. The rest of the documents in here are plans and descriptions of projects that they have under a couple big headings that they use all the time, DII for Defense Information Infrastructure and then software that runs on that, kind of a big operating system that controls a lot of smaller operating systems called COE, Common Operating Environment. So there's a lot of material on what DII is and a DII master plan and what COE is and so I noticed a name on one of these things is a Lieutenant Colonel named Shawn Butler who ended up helping me a lot in preparing my final report which is about Air Force software. I had her on loan from the Army. She retired, went back to university at Carnegie Mellon University and at about age 46, got her Ph.D. in computer science at Carnegie Mellon University. The final document in here is somebody's report to the Defense Science Board Task Force on acquiring defense software commercially. The Defense Science Board is the major Department of Defense Scientific Advisory Board. And they were assessing the question of what happens if you buy defense oriented software commercially rather than having everything custom built for the Air Force. That was a very critical, important issue and it was resolved by Defense Secretary, William Perry eventually issuing a decision memorandum which said that Air Force software would be commercial software unless you could convince the Defense Department to make a custom system, whereas previously the reverse had been true. Everything was a custom system unless you could convince them that a commercial piece of software was okay.


Folder 16: Defense Science Board Lederberg Task Force on Break through Technologies 1995

One of the key members of the high level Science Board, the Defense Science Board, for years has been my colleague, Joshua Lederberg, who I worked with so many years here at Stanford on the DENDRAL project, SUMEX AIM and other projects, MOLGEN. He led a Defense Science Board Task Force which reported on October 12, 1995 on breakthrough technologies. And so this would be a very interesting document to study to see what in 1995 some of the best minds looking at military science and technology were able to come up with in the way of breakthrough technologies that were possible.

Box 16 - 2005

Folder 1: ISO--ARPA 1996

ARPA is the same as DARPA, the Defense Advanced Research Projects Agency. ISO was Information Systems Office. ISO came about as there was a fission between the early office that produced all the miraculous results that DARPA is so famous for was IPTO, Information Processing Techniques Office. Because it got so large, they split it into Information Technology Office and Information System Office, representing the difference between more basic research and more applied research. ISO was the systems office. This folder contains a number of documents of what ISO was doing in the mid 1990's.


Folder 3: ARPA Software Review Panel 1996

This folder actually contains documents having to do with both of those offices I mentioned, ISO and ITO. Key document here is the Software Review Panel Executive Summary. Why was ARPA doing a software review panel? It's because software continued to be such a key issue for Defense Department correct functioning and acquisition of new systems and equipment and capabilities. And software remained almost in what you might call a pre-engineering state even into the nineties. So people were concerned about getting it into a much more scientific posture. And so DARPA being the main research agency was worried about that, wondering how to expend research funds.


Folder 8: Networked Centric Naval Forces 1999

I've already, in one of my previous memos discussed this committee. It was a committee of the National Research Council which is a study arm of the National Academy of Science and Engineering. And it deals specifically with naval problems. The Chief of Naval Operations, CNO, decided that he wanted the Academy to study this concept of network centric warfare and how would it apply to his Naval forces. So this massive committee was set up that met several times and wrote a report. In one of these folders, I gave a report on the ongoing study that the Air Force was doing on a network centric information architecture called The Battlespace Infosphere. I was on this committee and I diligently attended its meetings. This is the committee that I said really had a very minor result. That is I didn't think much of what it produced and I don't think it had much of an effect on the Navy.


Folder 10: Navy Artificial Intelligence (AI) 1995

In another note, I commented on the Army establishing an artificial intelligence laboratory in the Pentagon but they regarded themselves as an AI proponency organization. The Navy also established a Navy artificial intelligence laboratory. It was a bit earlier than the Army's. It was located on the outskirts of Washington near the Naval Research Lab. It had its own separate building. This is a folder of some of the materials that I got from that laboratory in 1995. So I was already Chief Scientist at the time. This Navy Center for Applied Research in Artificial Intelligence as it's called, was actually quite good. They had hired excellent people who were doing very good applications. And I believe they are still in existence as I dictate this in 2005.


Folder 13: Critical Technology Assessment of US Army AI Sector 1994

Obviously in the middle of the 1990's, artificial intelligence. which we normally think of as being a subsection of computer science and information technology, was regarded in a special way by the Commerce Department as being a sector. Artificial Intelligence standing out there by itself. And this is a very in depth assessment of what is that sector like commercially, who's doing what, what's inside that sector, what are other countries doing? Although I can't comment on the quality of the document because I don't remember what it said, I would say that it does stand for a very good assessment that is a sort of an accounting for what was going in artificial intelligence in the mid 1990's.


Folder 22: US Army Artificial Intelligence (AI) Center--AI Proponency 1994

AI Proponency means that the Army, Artificial Intelligence Center in the Pentagon, was promoting the idea of using artificial intelligence systems by which they meant mostly expert systems technology in various Army applications. And each year, they would basically have a conference in which people who were doing that kind of work and there were dozens of places all over the Army that were doing it would come and report. The Army eventually abandoned the Proponency Center, I think because they felt that they didn't need proponency anymore. That it had taken off on its own and it was established in the Army. But in any case, this gives a good view of what the Army was doing in the mid 1990's in applying artificial intelligence to its war fighting.

Box 17 - 2005

Folder 1: Papers--Office of Science and Technology (British?) 1994

These papers do indeed come from Britain. The British had their own planning process for information technology and electronics. They were looking forward into the late 1990's and beyond as to what the country should be doing and what its various governmental and commercial subunits should be doing to make Britain a leader in information technology and electronics. That study was called Technology Foresight, Progress Through Partnership: IT and Electronics. So it might be that IT and electronics was just a subpart of a much bigger study called Technology Foresight. But anyway, this is the British report. There's also a sub-report called the Technology Demographics Roadmap. It's also in IT and electronics. And there's a miscellaneous circular in here which just comes from a corporate visit that I made and the corporate visit was to the Architecture Projects Management Limited which appears to be a software architectures study group.


Folder 15: KBS Problem Solving 1996

KBS is Knowledge-Based Systems. That was a hot topic in artificial intelligence throughout the seventies, eighties and nineties but it kind of came to a head in the 1990's when there was an effort made by DARPA and also by myself in helping DARPA get some Air Force funding to establish an additional research program in knowledge based systems. And so in this folder are various briefings on different people talking about knowledge based systems research, problem solving, knowledge acquisition which is another way of talking about learning, machine learning or human machine collaboration in learning. And there's an acronym in here which is HPKB which stands for High Performance Knowledge Bases which is what DARPA called its program when it finally launched it.


Folder 33: National Academy of Engineering (NAE) Tak Notes undated

This is a to-do item for me. It says write up this National Academy of Engineering (NAE) talk. So obviously I gave this talk. These are notes. Obviously I gave this talk to some group at the National Academy of Engineering in Washington and it tells me to write this up for a magazine called "Issues" or "Similar", one of the other magazines where one would write up general policy articles on scientific research. This was not a talk about science. This was a talk about how to manage research labs in an era of change and the era of change was the mid 1990's. Things were changing in Washington. I was trying to focus attention on certain changes and suggest some actions that research managers of laboratories could take to meet the current needs. And I don't remember what research laboratories were being talked about. Was this a gathering of Defense Department and service laboratories? That's fairly likely. Was it a gathering of industrial research lab managers? Were they invited also? I don't know. There's nothing in here that says what kind of people are invited to this talk.

United States Air Force Scientific Advisory Board 50th Anniversary 1944-1994
On the cover is some printed icons of General Hap Arnold famous in World War II and Dr. Theodore von Karman who was the Air Force's main Scientific Advisor during World War II. And Hap Arnold wanted to continue as World War II was drawing to a close, he wanted to make sure that the relationship between science and the Air Force continued strong into the post World War period so he got von Karman to establish a group which subsequently morphed into the current Air Force Scientific Advisory Board.

United States Airforce Scientific Advisory Board 50th Anniversary 1954-1994
There happens to be inside this folder an envelope with two stamps, a Hap Arnold stamp and a Theodore von Karman stamped when it was postmarked on this part November 9th and 10th for this particular meeting. And that's probably reasonably valuable, whatever stamp collectors call this first day of issue stamp. Or maybe it's not first day of issue. Maybe it was issued before but it was just stamped in honor of this event. It was an honorific symposium with a very high level speeches including a speech by the Secretary of the Air Force and the Keynote Speech by the Secretary of Defense, Bill Perry.


Folder 35: Software Reuse Executive Primer 1995

I'm making a comment on a variety of separately bound documents which reflect concerns of the U.S. Military, on the DOD level and the Air Force level with Advanced information technologies.The first one here is a software reuse executive primer dated May 17,1995. Software reuse was one of the biggest issues for the DOD because the DOD was accustomed to having a great deal of its software prepared custom made for its systems. And so each time it would have to do something similar, it would have to remake that same software at great expense and introduce new bugs. So software reuse was a hot topic in the DOD and the government.


Folder 36: Commercial Multimedia Technologies for Twenty First Century Army Battlefields 1995

I'm making a comment on a variety of separately bound documents which reflect concerns of the U.S. Military, on the DOD level and the Air Force level with Advanced information technologies. The use of what's been called multimedia technologies which means various vision and voice and interaction capabilities. In this case, there was a National Research Council study on what use could that be in on Army battlefields? And various other things like digital communication in satellites which use the direct PC, the DirecTV satellite which was at that time broadcasting bits as well as video signals may be still. And similarly, the government as a whole was concerned about this and there was a report on the emerging impact of consumer digital systems. That is, the things we buy at Fry's, the things we buy at Radioshack, on the government, how the government could realize savings from that and move faster.


Folder 40: Effective Retrieval of Human Technical Knowledge 1996

There's a folder here entitled "Effective Retrieval of Human Technical Knowledge". And it's a bunch of documents including a letter dated June 11, 1996 from a man named Eric van Nortwick. I don't remember who he is and I don't remember this incident. He just mentions though that he viewed a videotape by me discussion "Tales of a Tiger" pertaining to knowledge systems. Well it's not that at all.
It's called "Tiger in a Cage". And Tiger in a Cage is the title of a talk that I gave to a conference of the American Association of Artificial Intelligence some year, perhaps 1993, on the tiger which is expert systems. The cage is why haven't they broken out into widespread use. And that was made into a videotape by a company that makes technical oriented videotapes and sells them to the public. So that might be of interest to some historian who wants to look for it or perhaps it's included in one of the videotapes that we have here. It's called "Tiger in a Cage".


Folder 44: Brain Actuated Control Research 1995

I'm holding a folder which is entitled "Brain Actuated Control Research" and the information in here is dated April and May of 1995. In connection with New World Vistas, one of the things we were looking at was the question of whether in the future we would be able to control external devices by simply thinking about them. Usually, we control devices by thinking and then the thought moves our muscles and the muscles moves a lever or a switch or something and then the device is controlled. But can you go directly from the brain through the nerves out to a system and control it. For example, can you think about flying your airplane instead of pushing levers one way or the other or can you think about firing a missile or something like that and just have it done. The New World Vista Study took a fairly positive view of the future possibilities of that kind of research and development. This particular folder - apparently I asked someone in the Defense Department or the Air Force to find out what was being done in the various military laboratories on this subject and this is a response of what was being done. This is dated 1995. It's now 2005 and the results of that work over the last decade have been really quite sensational so that if anything, our thirty year prediction for this was too conservative. It's probably going to be something like a twenty year prediction. For example, monkeys thinking about doing certain tasks can actually control devices to do that task. And this has major importance, not just for paraplegics and other people who can think but can't move their muscles.

Box 18 - 2005

Folder 1: Encryption Policy 1994--1995

The debate on government control of encryption was one of the hottest debates that was going on in Washington while I was there. Certain key source articles are present in these files, including a document prepared by some of the nation's best cryptographers, including a proposal by John Deutch himself. John Deutch was the Deputy Secretary Defense and he, himself was an MIT scientist. And this debate, John and Bill Perry, his boss, were bringing right to the cabinet and putting it before Bill Clinton and the policymakers within the administration. Eventually the strong encryption people won out so that when you, for example, open your wireless connection to a computer, you're asked if you want 128 bit encryption and stronger encryptions even are readily available.

meeting with Deputy Sec. of Defense about National Encryption Policy
The debate on government control of encryption was one of the hottest debates that was going on in Washington while I was there. I'm looking at a memo for the Deputy Secretary of Defense which at that time was Dr. John Deutch, dated January 23, 1995. So that gives you some idea of the date of this furious policy discussion. It's all cooled down right now and it's hard to imagine with what heat this was being debated. It was a struggle between essentially the code breakers at the National Security Agency, NSA, whose job it to crack everything versus the who are both privacy advocates and advocates of control of business information so that business secrets don't leak and people who just don't want other people spying on what they write and what they transmit. The problem is that in the digital era in which massive amounts of computation are available to every person in their desktop machine, even in 1995 (and now, of course, ten years later, about maybe a factor of a few hundred greater than what we had in 1995) some of that computation can be pointed in the direction of encrypting everything, encrypting documents, your Word documents, encrypting your email, encrypting your voice over internet type phone calls. Anything can be encrypted with encryptions so strong that it would take forever for the National Security Agencies supercomputers to break that. And then along with this debate about what should be available to the public within the United States by law and also who controls it. There was also the question of what should we be allowed to export? Should we be allowed to export let's say a Microsoft Windows System that gave the user encryption capability with longer than 40 bit keys? When I say 40 bit key, I mean that that was the length that the NSA originally had agreed to and everyone else thought was too easy to break. They agreed to it because they knew they could break it pretty quickly. And cracking a code gets harder as you increase the number of bits in the code. By the time you get to the key structure that most everyone thought would be the right one, 128 bits, the code becomes virtually uncrackable by any combination of computers that you could put together. So the debate was pretty much over who gets to control what and and where in the spectrum from 40 bits to 128 bits are you going to come down? And, you know, there were compromises proposed at 50 bits, 56 bits and other places.

Phil's Pretty Good Software Presents
A very well known document in which a citizen who decided that he would just take things into his own hands, his name is Phil Zimmerman. Phil Zimmerman created an encryption program which was was posted on the internet and was distributed to anyone who wanted to log into MIT which gave people strong encryption capabilities for nothing. This revolutionary program was called PGP which stood for Pretty Good Privacy. And PGP sort of changed everyone's focus because it represented facts on the ground. Here was such a program and it was being distributed all over the world on the internet. So what are you going to do? Was it sort of a challenge? And the government tried to bring Phil Zimmerman to court over this and eventually dismissed the charges against Zimmerman. PGP became a company and was then going to market additional versions of this thing commercially. It's still a company. It's not a very successful company.


Folder 9: Command and Control - C2, Notes and Presentations1996

Advanced Battlespace Information System (ABIS) Study
This folder also includes a document sent to me by Anita Jones who was the Director of Defense Research and Engineering, the so-called DDR&E. DDR&E was essentially the Chief Scientist of the Pentagon. So Anita was essentially one level above me in the hierarchy. She was at the Bill Perry, Secretary of Defense Level whereas I was at the Air Force level. And she did a very fine briefing on the subject of battlefield Command and Control with Admiral Cebrowski, who was Director of Command and Control Communications and something else. It's called C4 Systems. I can't remember what the fourth "C" is for -- the Joint Chief of Staff, JCS. And he and Anita did a very fine briefing on a constant called ABIS, Advanced Battlespace Information System. Anita Jones had been a professor of Computer Science at the University of Virginia before she was asked by the Clinton Administration to be the DDR&E. She's an excellent computer scientist. She's a member of the National Academy of Engineering. And for me, it was a tremendous boost and pleasure to be serving in the Pentagon at the same time as Anita. She's an old friend from computer science from her graduate student days at Carnegie Mellon University. So it was like old home week. It was old home week to have Bill Perry there as Secretary of Defense. It was great to have another academic, John Deutch, as Deputy Secretary of Defense, Sheila Widnall from MIT was Secretary of the Air Force. As Bill Perry had told me before I went to the Pentagon, it's a good time to be a Chief Scientist at the Pentagon.

Chuck's SLIDES 7/14/96
This is a set of viewgraphs, a typical Air Force way of communicating, that deals with what apparently was a Scientific Advisory Board study, probably from 1994 or '93 called a C2 vision, meaning a Command and Control vision for global awareness, global reach, global power. That's a slogan the Air Force was using during that time: Global Awareness, Global Reach, Global Power. That's what we're all about. So this was a Command and Control vision for that. Command and Control was, at that time, and continues to be perhaps the single biggest technical problem of the Air Force. They just can't cope with it. They can cope with aircraft but they really can't get their technical minds wrapped around complex information processing and decision making wrapped up with sensor inputs from all over the world, from all over the battlefield, software driving everything. That remains a tremendously difficult problem for the Air Force. So it was unique to the Air Force that they had a software scientist in the Chief Scientist position when they asked me to be Chief Scientist but subsequently of the several Chief Scientist's who followed me, two of them were deliberately chosen because they were experts in the Command and Control area, one person from the Mitre Corporation and one person from Alex Levis and George Mason University. And it continues to be a critical problem for the Air Force.

Vision Document
General Muellner was the military side Director of Acquisition for the Air Force. Actually I have to hesitate there and say that General Muellner had several top level jobs and it could be that at the time that - of this particular memo, it could be that he was in charge of one of the fighting divisions of one of the commands of the Air Force. That's possible. This was a vision of Air Force Command and Control, 2010. And apparently he asked for some kind of a briefing on what the Scientific Advisory Board and the Chief Scientist thought should be done for Command and Control. This particular document is a summary document for him plus some backup notes for that.


Folder 10: Anita Jones: Statement to Subcommittee of Senate Armed Services1994

Defense Science and Technology (S&T) Briefing to the Joint Requirements Oversight Council (JROC)
This was a briefing to the DOD's Joint Requirements Oversight Council. This is a very high level body within the Department of Defense, I believe it's chaired by the Deputy Secretary of Defense, that looks over all systems that the DOD wants to acquire or thinks it might want to acquire to see whether it fits in the overall vision of where the military should go. It's a very important control function of the DOD level of the Pentagon and these are two documents by Anita for that committee.


Folder 11: Anita and EAF : Lessons learnedn.d.

Notes for joint paper by Anita Jones and Feigenbaum on lessons learned on science and decision making in the DOD and Air Force
As we came into 1997, Anita was going to retire from her job with DDR&E.
I knew that my job was going to come to an end at the end of August, 1997. So she and I sat down to kind of debrief each other on the question of what did we learn about the military? What did we learn about the Pentagon? We called this set of notes "Lessons Learned, Anita and Ed". And I typed it up, wrote things on it. Anita commented on it.
So this is massively documented handwritten document and was intended to be a paper that Anita and I would write jointly and publish. And somehow we just never got around to publishing the paper but the material seems as fresh as the hours that we sat down and worked on this outline of a paper.


Folder 12: Thank You Letters, Congratulations Letters1994 - 1996

Letter from Hans Mark to Professor Edward A. Feigenbaum
This letter is from Hans Mark who also at one time was Secretary of the Air Force and then I believe he was head of NASA, a very important figure. He ended up as being, I believe, Anita Jones' successor as DDR&E in 1997 or maybe shortly thereafter. But he's one of the truly great American scientists.

Letter from Sheila E. Widnall to Dr. Edward A. Feigenbaum
I'm looking at a folder called "Thank You Letters" but maybe the phrase "Thank you" is not the right title for this thing because the first letter in here is a congratulatory letter from Sheila Widnall, the MIT Scientist who was Secretary of the Air Force when I was Chief Scientist of the Air Force. And she writes me to congratulate me. She heard that I won the ACM Turing Award. Turing is a famous British scientist often thought of as the first computer scientist. And her secretary obviously got it wrong because she must have thought that she was correcting a mistake because she said Dear Ed, congratulations on being named the recipient of the Turing Award, which should have been the Turing Award. And anyway, I remember this letter. I didn't know where it was but I'm very glad to see it again.

Box 19 - 2005

Folder 2: Heap H1 to H101986

Proposal for a Center for Medical Informatics, sent to Dean Korn of Stanford Medical School
A Lane Library memo that is from Ted Shortliffe and a couple other people. "Proposal for a Center for Medical Informatics", dated June 3,1986. And this particular memo obviously was a copy sent to me by Tom Rindfleisch who much later became the head of the Lane Library. But at this point, Tom was the Managing Director of our Knowledge Systems Center' and he's a computer scientist who was running a major computer facility for artificial intelligence in medicine for Joshua Lederberg and myself, the SUMEX AIM project. Now as a result of the work in artificial intelligence in medicine, there was considerable interest in creating a Center for Research in medical applications of computer science and indeed in basic research in computer science related to medical needs. They didn't use the word computer science. They used the word informatics which is a term that the French came up with when the French decided that the usage of the words computer science was ugly and not French. So the medical community, when it steered off in this direction used the same terminology, informatics. So this is medical informatics. In the end Dean David Korn to whom this memo was addressed apparently chose not to set up such a center because we never had one. But there was established in the medical school - I'm not sure if it's before or after this memo - the section on medical informatics of the Department of Medicine, probably slightly before this memo which housed the faculty members interested in this topic.


Folder 3: Heap H11 to H201987

Letter from Raj Reddy to Dr. Saul Amarel
A 1987 letter from Professor Raj Reddy at Carnegie Mellon who, at that time, was President of the American Association for Artificial Intelligence to Dr. Saul Amarel who was teaching at Rutgers University Computer Science Department. But at that time he was working in Washington as director of the office in DARPA that was giving out computer research funds. And Saul Amarel was also Chairman of a FCCSET subcommittee. It's normally referred to as "fixit" but it's a coordinating committee - federal coordinating committee and the SET probably is Science and Engineering Technology but I'm not sure. This is the subcommittee on computer research and development. This is to coordinate the activity of government agencies. Saul apparently asked the American Association for Artificial Intelligence to come up with a list of grand challenges that the society would consider important grand challenges for the future of artificial intelligence research. And Reddy gives them a long list and a few paragraphs of discussion of this list. What's interesting about this is that this is a never ending task. This came up several years before this and before that, several years and then it was a hot topic as of January of 2005 when I went to Washington to attend a meeting called by DARPA for exactly this same purpose. So it's a never ending job to create challenges that people would consider to be exciting challenges for intelligent systems.

No Silver Bullet - Essence and Accidents of Software Engineering
This is a reprint of a Xerox copy of a paper by a very famous computer scientist Fred Brooks from the University of North Carolina called "No Silver Bullet". Fred was one of the key people in managing the IBM System 360 project. He also did a famous book called "The Mythical Man-Month" about creation of software. And this paper "No Silver Bullet" is probably his next most famous thing. And I happen to categorize it under expert systems because it occurred to me that the things he was talking about might be doable if we acquired enough knowledge about software engineering and programming to create systems to do it automatically. Now unfortunately this idea was tried several times and has never been completely successful. It's been only partially successful. So it still remains a great challenge. People do programming quite well and programs need a lot of assistance in order to help create other programs.


Folder 4: Heap H21 to H301987

Artificial Intelligence: A Rand Perspective
This is a paper written by Phil Klahr and Don Waterman, two employees of the Rand Corporation. Don was a Ph.D. student of mine who started with me at Berkeley and then transferred down to Stanford with me in 1965 when I came down to Stanford and got his Ph.D. here at Stanford, was my first Ph.D. candidate who graduated here at Stanford. Might be my first Ph.D. candidate that I ever graduated. Don unfortunately died many years ago at a very early age. He died of lung cancer. But anyway, this is entitled "Artificial Intelligence, A Rand Perspective". Rand is a famous think tank organization in Santa Monica, California. It had a very important role that it played at the beginning of what we would now call computer science and also at the beginning of artificial intelligence. Herbert Simon was a consultant for Rand during the early fifties in some work that led up to inventing AI. Allen Newell was the person he collaborated with and J.C. Shaw, Cliff Shaw, were actual employees of the Rand Corporation. I was a consultant for many years - from 1957 to 1964 or actually even much longer than that - but I would spend every summer at the Rand Corporation. And they were sponsoring research in Artificial Intelligence. And subsequently in the seventies, picking upon work that we did here at Stanford and maybe even picking up on stuff that I transplanted to them from here at Stanford, they did a a great deal of work on expert systems and expert system development tools and expert system applications to the military. And Don Waterman was at Rand and he was on of the what you might call the carriers of the virus. He knew what we were doing and he helped to do it down there at Rand. And Phil Klahr was a very smart colleague of Don's and others at Rand.This document purports to tell the history of artificial intelligence as seen from theRand Corporation point of view. And the first paragraph starts off by talking about in looking at the first published book on AI, the 1963 "Computers and Thought Anthology of Feigenbaum and Feldman", one realizes that Rand's contribution at the beginning of AI was very substantial, blah, blah, blah. It just tells the story. My guess is it tells it quite accurately. Now the question is why did Phil Klahr and Don Waterman write this article? The answer is I don't know. Was it for a journal article or did Rand published a book on its great contributions to science? I'm not sure.

MCI Mail
This is a brochure from what at that time in late 1980's was a relatively new phone company which got started after the AT&T breakup, a phone company called MCI which was bought by WorldCom and then there was the WorldCom bankruptcy and then now the company reverted to being called MCI. But the computer science research community had email from the time that our community invented it back in, roughly speaking, around 1970-2 as an early application of the first digital network, the ARPANET. And so email was nothing to us. We had already been doing most of our communication with each other by email; we had sort of given up regular postal mail. But it was frustrating in that the world didn't seem to want to come around to this point of view, that no one was getting their act together to offer a service like that to the rest of the world. In fact, the rest of the world sort of got dominated by fax machines in the 1980 much to the frustration of all of us. We thought why are they dealing with fax? Fax is analog paper and anyway, turns out MCI got the idea - they may have gotten the idea by hiring one of the co-inventors of the internet, Vint Cerf. Or Vint may have gone to work for them after they got this idea but, in any case, Vint ended up as a Vice President and Senior Vice President for MCI and he's still there. However, this brochure says after 208 years, the nation finally has a new postal system. And that's announcing the availability of email.

Notes for Siglunch on knowledge representation logic
There's a little package of electronic printed material here, H21. The reason for scanning this and commenting on it is that it represents the kind of discussions that were taking place among AI people around about middle to late 1987.
There's some discussion here on knowledge representation. There's another email from David Wilkins that is talking about two different kinds of AI researchers, particularly the young AI researchers - the scruffies versus the nests, the neats being the people who work in the mathematical logic approach to artificial intelligence and so on. The reason for including this here is these different topics are representative of the kinds of things people would be talking to each other about.


Folder 5: Heap H31 to H401987 - 1988

Limits of Artificial Intelligence
A Xerox copy from the Encyclopedia of Artificial Intelligence Volume I, edited by Stu Shapiro. This encyclopedia in itself is an interesting development because it was in the late seventies through about '84 or 5 that I and my coworkers were putting together a massive collection of books called "The Handbook of Artificial Intelligence". So it wouldn't seem likely that there would be a need in the world for a thing called the Encyclopedia of Artificial Intelligence. We actually, in our book,eschewed the word "encyclopedia". Instead we chose handbook because we wanted to indicate that we were offering more in the way of detailed methods. It wasn't just small encyclopedia articles about something. These were actual methods to do something. But anyway, it provoked a competitor by Shapiro called the "Encyclopedia of AI". And one of the articles in that encyclopedia - the editor obviously solicited it - was an article called"Limits of Artificial Intelligence". And it was not done by a gadfly in the normal sense or the kind of people who had been writing books and articles like this. And it wasn't done by a philosopher like Hubert Dreyfus at Berkeley. It was done by one of the most eminent mathematicians in the world, Jacob Schwartz of NYU who at the time was taking an interest in the field. He may even have, by that time, have become Director of the Information Processing Techniques Office of DARPA. But he wrote this lengthy and well-reasoned article about limits of artificial intelligence although there's no a priori reason to believe that he had any particular insight or ability to think about that particular issue. But then that didn't hold back Roger Penrose, the famous British Physicist, from writing his book called "The Emperor's New Mind". What does Penrose as a physicist know about artificial intelligence? But he wrote yet another article - in this case a book -attacking artificial intelligence.

Texas Instruments microExplorer Computer System
A promotional material from Texas Instruments announcing the availability of their Micro Explorer Computer System. Now the Micro Explorer is basically the combination of the LISP machine and the Macintosh. They had already produced a fairly expensive and powerful LISP machine called the Explorer and when they had gotten enough of the LISP machine apparatus down onto a chip as technology moved forward, they were able to integrate it in with the rest of the computational architectural framework of a Macintosh. So this is Micro Explorer combining the power of the Explorer and the Macintosh II. Now this is really a landmark.


Folder 6: Heap H41 to H501987 - 1991

Comments on Logic at MIT Workshop
This is handwritten by me. It's comments on logic at an MIT workshop in June 28, 1987. The reason that this is important is that it's one of the places where I come out strongly and I noticed in the middle of this page here somewhat abrasively about my irritation with how much effort was going on in artificial intelligence on the sort of perfection of techniques and methods for mathematical logic and their potential application to artificial intelligence which seemed to me at the time and has seemed to me ever since, to be relatively sterile. And this is not a popular view or a widely shared view. So that's why I think this particular piece of paper's quite interesting.


Folder 8: Heap H1000 to H10051985 - 1986

ICOT Journal 1986 No. 11
The institute for fifth generation computer system technology that the Japanese started with a national project in 1983, the institute otherwise known as ICOT published a journal and these particular two small issues of the journal are significant in that they talk about intermediate results and intermediate plans of that project. The project started around 1982 and ran until it was supposed to finish in '92, finished a little bit later with additional government funding and was the subject of that book that I wrote with McCorduck called "The Fifth Generation".


Folder 9: Heap H1006 - H10101985 - 1986

PROTEAN: Deriving protein structure from constraints
This was an invited paper for the International Joint Conference on Artificial Intelligence, 1977 at MIT. It represented a kind of a coming out for expert systems and knowledge engineering as a main theme in artificial intelligence. So therefore was quite an important paper for me. I worked on it quite hard and got it to be I thought a very good paper. In fact, when I presented it, I believe I presented it one more time later.


Folder 10: Heap H1011 - H10151986

Annual Report - Medical Information Sciences Program
What's particularly interesting here is that in the annual report, it talks about May 1986, the Medical Computer Science Group of the Knowledge Systems Laboratory in the Medical School at Stanford. Well, of course, the Knowledge Systems Laboratory is the umbrellalaboratory that I had started and that we were running here. And Ted's group was still a part of that which we were calling the Medical Computer Science Group. We weren't calling it the Medical Informatics Group. When Ted spun this group off to be its own thing in the medical school, it became Medical Informatics and it became part of the Department of Medicine. So that would clear up any confusion based on earlier notes when I was referring to Medical Informatics.


Folder 12: Heap H1018 to H10201986

"Heap" H1018 to H10201986
A panel report from a panel that was discussing something at the National Academy of Public Administration in Washington (might be subtitled, you know, one of Ed's brush with the greats). This was sent to me by one of the secretaries or assistants to Frank Carlucci, who was a member of the Board of Sperry Corporation while I was a member of the Board of Sperry Corporation. Subsequently Frank Carlucci became Secretary of Defense and subsequent to that, he became the founding President of the Carlyle Group which is now quite well known as an investment group. Anyway, this may be the only thing in here from Carlucci to me.


Folder 17: Heap H1039 to H10401986 - 1987

The Development of Logic
Thi is an example of the kind of informal communication I would relatively often get from Joshua Lederberg. This is dated July 31,1987. So Lederberg has been away at Rockefeller University as President for some nine years by that time. And yet he still will be looking for things that are interesting to our work and sending them on. In this case, he's sending it to me but I guess he really meant it for myself and Bruce Buchanan and he said I think you and Bruce, if you don't already know this, will find these reflections and challenges interesting. So it happens to be an article from Oxford from the Clarendon Press on the development of logic. But here's Lederberg looking out for things that we ought to be interested in.

Box 20 - 2005

Folder 1: Heap 1 to 15--(1 of 3) Spring 1985 1985

This folder is labeled "Heap, One of Three, spring 1985". Heaps are just miscellaneous collections of documents with numbers on them, accession numbers. Spring of 1985 is an interesting time. It's right in the middle of this so called AI boom. The AI companies are going strong. Lot of international interest in AI that's due to my book "The Fifth Generation" and also the Japanese Fifth Generation Project and a lot of success in expert systems. These are miscellaneous papers that just relate to various expert system projects that are being done elsewhere and some here at Stanford. There's even a final exam here from a course that I was teaching. It's sort of a computers for poets, "Computers, Their Nature, Use and Impact". There's a very interesting brochure from Xerox in which they're selling a variety of tutorials and other things related to getting people technically capable in the artificial intelligence field. And there are a lot of projections about chips and what computers are going to be doing for people in the future. Of course, now it is the future. This is twenty years later so it's very interesting to just take a look at that.


Folder 2: Heap 16 to 22--(2 of 3) Spring 1985 1985

Heap two of three in the spring of 1985. The important items from a perspective of looking back. To me the important items are the items eighteen through twenty-one. These are papers from Carnegie Mellon University. Herbert Simon is involved in co-authoring or authoring two of these papers. One relates to his work on human scientific reasoning, how was it accomplished in the early days with Yun Gitkou and Pat Langley - Pat Langley is here at Stanford now in 2005, Jill Larkin who studied human expertise and then apparently was working on an expert system program. And a paper by Carbonell, Larkin and Reif on a general scientific reasoning engine which is very similar to the work that Allen Newell was doing at Carnegie Mellon.


Folder 3: Heap 23 to 31--(3 of 3) Spring 1985 1985

This is Heap three of three 1985. The most interesting document in here is number twenty-three with a cover letter. It's about a computer system that was developed at Columbia University, Computer Science Department. One of our graduates of our program here at Stanford, joined them as a Professor and here he's listed as an Associate Professor so he obviously got tenure. The interesting thing is that it's David Elliot Shaw whose D.E. Shaw. And it turns out that Dave Shaw decided somewhat thereafter from after his work on this computer, that he wasn't going to be a computer scientist anymore. He was going to invent some new financial instruments based on what you could do with computers to project the financial returns on certain kinds of financial instruments. And so David Elliot Shaw founded D.E. Shaw Incorporated which is this very large financial services firm in Wall Street now.


Folder 9: Knowledge Interchange Format version 3.0 1992

This paper is from the Logic Group which is the name that Mike Genesereth; Professor Genesereth gave to his scientific working group here at Stanford. And there are a lot of people's names on this. It's Genesereth and Fikes and a lot of names that everyone would know in the field of AI, rather well known names. I just wanted to comment that this is an example of where my work in the Knowledge Systems Laboratory was fading out. Fikes was taking over as the, what you might call the intellectual leader of this laboratory and my interest was moving over toward the software industry and then later on, I left here for three years to go to the Air Force and be Chief Scientist. So this is a perfect example of where Fikes was the leading contributor and he's co-author of this paper.


Folder 15: Lists of Heuristic Programming Project (HPP) Publications1980 - 1986

Letter from Elisabeth Beller to Susie Barnes
There's a little package of electronic printed material here, H21. The reason for scanning this and commenting on it is that it represents the kind of discussions that were taking place among AI people around about middle to late 1987.
There's some discussion here on knowledge representation. There's another email from David Wilkins that is talking about two different kinds of AI researchers, particularly the young AI researchers - the scruffies versus the neats, the neats being the people who work in the mathematical logic approach to artificial intelligence and so on. The reason for including this here is these different topics are representative of the kinds of things people would be talking to each other about.

Medical Computer Science Publications
This is from the medical computer science group of the Knowledge Systems Laboratory and it's listed as jointly the Departments of Medicine and Computer Science at Stanford. And so the list is July 1986 which means that in July 1986, what is now the section on Medical Informatics of the Department of Medicine was still part of the Medical Computer Science Group of the Knowledge Systems Laboratory. So the spin off of this into the Department of Medicine apparently took place after July 1986. This is a very interesting list of publications.


Folder 20: Artificial Intelligence (AI) Research in Knowledge Systems Laboratory (KSL)1989

Artificial Intelligence Research in the Knowledge Systems Laboratory
The folder called "AI Research in July 1989". We at the Knowledge Systems Lab liked to produce sort of an up to date brochure of what it is we were doing so that we had handouts to give to people when they visited and that we could send to the numerous people who asked and remember that we didn't have websites at the time. So having a brochure like this was virtually like publishing a website is now. And this one was an updated version from July 1989. So it's quite valuable as a snapshot of what we were doing in the lab at that time.


Folder 22: The EMYCIN Manual and Tutorial1981

EMYCIN Tutorial
EMYCIN is the granddaddy of all expert system and development tools. The E standsfor Empty MYCIN or essential MYCIN. It's the software abstraction of the MYCIN expert system which was one of the earliest of the advisory expert system - maybe the earliest advisory expert system. So by 1981, this was quite a famous program and we had pretty much wrapped it up here at the Heuristic Programming Project but there was a manual here which was HPP 81-16 or Stanford Computer Science, CS81-885 in which Van Melle and Carli Scott, Jim Bennett and Mark Pierce lay out the details of EMYCIN plus there's a tutorial. Presumably that tutorial was given at some annual meeting of the American Association of Artificial Intelligence or similar. But these are historically very important.


Folder 48: Schlumberger

Box 21 - 2005

Folder 2: Heuristic Programming Project (HPP)1980

Heuristic Programming Project 1980
We at the HPP, by 1980, were quite possibly the most famous artificial intelligence project in the country or maybe the world. You know, at one time, it was the MIT Artificial Intelligence lab and the Stanford Artificial Intelligence Lab and so on. But for Andy Warhol's famous fifteen minutes, we were superstars. And we were really feeling good about it. And in 1980, we decided to tell the world in wonderful terms what we were doing. The cover of this brochure that we produced that was widely distributed in the field is a mural that was painted by the artist Harold Cohen based on drawings that were done by early versions of his expert system, AARON. Harold was a very well known painter. I helped to mentor him in his skills in artificial intelligence and expert systems in 1973 and 4 when he came up here. Subsequently, a book has been written about Harold Cohen's book called "AARON's Code" by Pamela McCorduck. So this brochure has a photograph of Harold's mural at the San Francisco Museum of Modern Art and it also has descriptions of all the work that we had been doing up to that time with some wonderful photographs, including a collage of photographs in the center, the original of which exists somewhere, perhaps in that storage bureau that we were talking about. We've been missing it for a while but there are early pictures of myself and Buchanan, Peter Friedman who's now the Chief Technologist of NASA AIMS. There's his picture as sort of a young college student. There's Paul Cohen who's a full professor at University of Massachusetts Amherst and many other people - Joshua Lederberg, of course. There's my wife Penny Nii who was an important part of the project. This is just full of the great people of the time. There's even a picture here of Motoi Suwa who is now a real big shot in the Japanese government in the science administration part of the Agency for Industrial Science and Technology in Japan. So this was quite a document and then at the end, it lists the key publications of the project. And I think maybe it also lists the doctoral dissertations that were done, all the funding sources that had helped us and all the people who up until 1980 had been affiliated with all the various divisions of our project. So this is truly a superb source of information.


Folder 25: American Federation of Information Processing Societies Headquarters Newsletter, May/June 1978

American Federation of Information Processing Societies Headquarters Newsletter, May/June 1978
This is the newsletter of the National Computer Conference, 1978, published by AFIPS. I'm on The Best Paper Award at the show, and so there's a picture of two of the officials of the conference and myself in the presentation of the award, and I have the award over in my office. Apparently this is the year when Bob Noyce and Gordon Moore won the Harry Goode Award of the society because their picture's here too. And the award is for big contributions for the field.


Folder 26: Expert Systems (ES) Course: Selecting an Application1985 - 1990

Expert Systems (ES) Course: Selecting an Application1985 - 1990
It's the selecting an application part that's the important part. There was a time when I was sort of the world's expert on that whole question of how do you pick the right applications for the artificial intelligence technology. Applications that were more or less guaranteed to be successful so that if you were a research team or you were a corporate development team, you would want to know how you spot these, what are good ones and what are not good ones? And I was the first one to give a talk on the subject. In the early 1970's, I was going around giving that and it was later picked up by other people and they were also giving basically the same talk. But this is an important piece of the history. I don't exactly know when these slides go back to but the thoughts go back to the early days of DENDRAL and MYCIN, the earliest expert systems.


Folder 29: CS 226--Preperation materials 1984

There are a set of quotes here. I used to use a lot of quotes in my lectures. One of them is from the "China Daily", Saturday, June 2,1984.
Well it turns out I was in China at that time on a National Science Foundation exchange, one of the early exchanges with the Chinese. There's a quote here about knowledge, an educated person, etc. There's a quote from Walter Wriston of CitiCorp. Walter just died recently, I think in 2005 or late 2004. There's a quote here from Lester Thurow of MIT about working smarter rather than working harder. And then there's a quote that has to do with Chinese medicine, the knowledge of doctors and so on. So I think those quotes are interesting.


Folder 31: CS 226 View graphs undated

There are two file folders here that have "CS226" on them. CS226 expert system view graphs, general and another one called CS226 view graphs. CS226 often was a course in which well known people from the outside came to Stanford to give the students reports on what they had been doing. And so a lot of the view graphs here are view graphs that they produced for the students for their talks or the students produced based on what they heard from these lecturers. Consequently, these two folders give a rather good picture of what was going on in the expert systems world that was relevant to student training and teaching at the time. That was roughly 1990, '91 and so on.

Box 22 - 2005

The back end of the box is just all materials related to the discussions going on about expert systems and all the different applications that were being developed through the eighties as well as the early nineties. And this might be one of the better collections of miscellaneous applications articles about expert systems.

Folder 3: CS101 Lecture Notes 1975-1985

This folder apparently holds my lecture notes back from 1975 through the mid 1980's when I perhaps stopped teaching this course. This was the course which, in the background, we called "Computers for Poets". It's CS101 Computers, Their Nature, Use and Impact was the official title of it but it was the course designed for the rest of the university so to speak in intelligent layman's language trying to get people from all over the university from the English Department, the German Department, the Social Sciences, you know, wherever they may come from, Law School, Medical School to take this. And I was very proud of this course because it was teaching in a very simple way some very advanced concepts for its time and making projections as to where things would go which ultimately did turn out to be correct and so in retrospect; this was one of my more successful teaching adventures. And one of them here, for example, in 1975, introduces telling the students about electronic mail when electronic mail had only been introduced a few years before that and there were very few places in the country that even knew about electronic mail or could receive it. But here were the students finding out about it or text editing. The personal computer hadn't been invented yet in 1975. And so there was no little computer and no Microsoft Word and none of that. And yet here was the students finding out about it.


Folder 11: Bennett, J., et al, "A Knowledge-Based Consultant for Structural Analysis," HPP Sept. 1978

SACON: A Knowledge-Based Consultant for Structural Analysis
This is an early expert system that created quite an excitement in the field when it came out. It's now dead and no one ever refers to it. It's called SACON/ It maybe the first move of AI out of chemistry and medicine into engineering. This has to do with structural engineering. Advising a structural engineer. And so it created quite a wave of excitement. It showed how we could use MYCN-like ideas for medicine and translate them into engineering, and help engineers.


Folder 13: Lederberg DENDRAL (pre-HPP)

Computation of Molecular Formulas for Mass Spectrometry
This is what Josh Lederberg was doing in Structural Chemistry on a computer before he met me. He did a book published by Holden-Day on computing the molecular formulas from some mass spectral data. That was pre-DENDRAL. A bunch of tables that Lederberg generated using ALGOL Programs on the sub-ALGOL monitor on the IBM 7090 at Stanford University. This was one year before DENDRAL. This is numbered HPP 64-2 but there was no HPP at the time. We had a numbering system where that's the year and that's the item number within the year. So this is identified as 1964 work. But of course I didn't get to Stanford until '65, and there wasn't a DENDRAL project in '65.

CS121 Prospectus, Readings & Outline
The hanging file that we're looking at is "CS121" taught in the winter of the school year, 1990-'91 by visiting professor Donald Michie. Donald Michie is Britain's great pioneer in artificial intelligence research. Michie has had a great career. When he was a young man, he worked for Alan Turing at the Bletchley Park Encryption and Decryption facility that did this incredible work during World War II. Well, there are massive amounts of notes and papers from Donald that he used in teaching CS121 which is an undergraduate course because it has the one label, 121. And it might be interesting to future historians as to how one of the greats in the field saw the field at that time in approximately the winter of '90-'91.


Folder 23: PreHistoric Expert Systems (ES) materials from Herb Simon1991

Letter from Herbert A. Simon to Professor Edward Feigenbaum
In 1991, apparently I had visited Herb at Carnegie Mellon. His letter is dated November 4,1991. It said Dot and I were delighted to see you yesterday, so I presume I was there on November 3rd. He and I, at that time, apparently talked about the kind of work that had been done at the very earliest days of artificial intelligence that by the late eighties and early nineties, would be called expert systems except that terminology and that concept hadn't even been around in these very early days. And Herb was very careful about who did what when and what was historically interesting and important. So he sent me those materials. He says here I have in mind to send them along with a brief forward to the annals of the history of computing. And I don't know if he did that. He said you will note that at least one of these systems was operative by July 1956, about the time that LT actually produced its first theorem. LT is the Logic Theorist which is normally credited with being the first heuristic program in artificial intelligence and therefore essentially LT was the birthplace of artificial intelligence. So here's Herb being very careful about the history and giving me the original materials. I think this is a really important historical document, aside from the fact that it comes from Herb Simon who's also one of the key figures in the whole history of AI.


Folder 24: Expert Systems (ES) - Tools Survey / USA1991

Tools Survey USA
The word "tools" here for expert systems means software development tools that help people develop expert systems quickly. This tools survey was put together - it says copyright 1988, Harmon Associates. The Harmon is Paul Harmon and Paul Harmon was producing a newsletter, "Expert Systems Strategies Newsletter". And that's one of the things he would keep people up to date on is what the various expert system tools were. What's astonishing here is how many there were. Hundreds of them around the world being sold. Of course, there just wasn't a market for hundreds of these things so no one of the expert system tool vendors did very well.


Folder 25: Expert Systems (ES) - Societal Implications1993

Societal Implications & others
During this period of late 1980's, early 1990's, I was teaching a variety of courses that had to do with introduction to artificial intelligence, introduction to expert systems. I was still teaching some general introduction for Stanford students into computers. And one of the things that you want t o teach students is to think about the implications of this technology for society. So apparently I cumulated notes on that topic into a folder called "Expert Systems, ES, Societal Implications". But it's actually notes from a variety of courses put together into a topic folder like that.


Folder 29: Expert Systems (ES) - Reading Material - General1989

Trilogy Development Group - Corporate Fact Sheet
What's interesting about this is that Trilogy is a corporation that was founded by students of mine that were taking my undergraduate introduction to expert systems course. And they had found out about this particular kind of expert system which is called a configuration system and they went out and started a company, started in 1989 in Palo Alto then moved it later to Austin, Texas and produced an extremely successful company out of that idea. Joe Liemandt was a student and he's really the entrepreneur who did a wonderful job in building up this company. It's now several hundred million dollars a year in software, and they're still in existence doing a very good job. On the Trilogy story, there was a cover story of "Forbes" magazine which was published in the 1990's sometime which tells the story of the various students of mine who went out started successful companies, including one on the cover who was involved with this company, Trilogy.


Folder 30: Expert Systems (ES)--Stanford Business 1987--1988

It's sometimes said that the shoemaker makes shoes for everyone else but not for his own children. Well this was an issue here at Stanford. We were the world's leaders in expert systems. We knew how to build them. They seemed to be enormously productive for companies that used them but Stanford wasn't using them. So I went and talked to my friend Bill Massey who was Vice President of Finance who's a Business School Professor and also a Vice President in charge of finance and business things for the university. Bill knew all about what I was doing. He knew the topic. And I just said hey let's look at what Stanford could do. Where are some of the applications that we could do at Stanford? So he sent someone to work on this. I guess it's a person called Richard Whitmore. And Whitmore is responding here on February 18th saying that he asked around, he's received and reviewed ideas from business and finance departments about what they could do. There are many promising ideas. He sketches out those ideas that could be done at Stanford. And then Bill Massey responds to Tom Rindfleisch and myself saying that he is happy to accept my offer. He thanks me for the offer of doing a SWAT review (SWAT review is what we used to call like this like the SWAT team in these television police films). We'd come in there and we'd do a review of whether there was a real expert system application there or not. You know, should we do this or not? We used to do that commercially in a company that some people and myself out of this laboratory started - TechKnowledge. We used to do SWAT reviews and here we were preparing to do SWAT reviews for Stanford. Bill Massey says we are looking forward to meeting with you soon to discuss the potential projects in more detail.
The outcome was we started a pilot project or two down the road at Stanford and assigned students to work on them and as part of the Master's Degree program. And essentially nothing ever came of it.
Although we had a Master's Degree program in AI. And a Master's Degree is viewed as a practicum or it was viewed by me when I started it as a practicum. That is instead of a person writing a Master's thesis, the person would do a significant project that would qualify that person for going out and being a specialist in the field. And we needed good problems. We needed to work with experts in various areas. So doing that work for Stanford was a way of getting ourselves some projects and associated expertise. So it wasn't totally with a view to make Stanford better. It was partly with an educational point of view on how to get good problems for our Master's Degree students.

Box 23 - 2005

Folder 23: Rand Mailbox 1974

This is the version of CS226 that I was teaching after I got back from the Air Force in 1998. So it has a bunch of new notes in it. Apparently I reworked some parts of the course and introduced some new things in it. I'm not exactly sure what it is but it's worth taking a special look at. Computer Science 226 which is the expert system course. This is the 1999 addition of it which has some additional view graphs in there and some additional notes that obviously I introduced specially into this course. So this is an example of what was being taught in 1999 about expert systems by me. I'm not sure when I last taught this course. In fact, I'm not exactly sure when I retired. It might - it'll be on my CV but it's somewhere around the millennium. I just sort of took that as a signal to retire and stop teaching the stuff.

Box 24 - 2005

Folder 2: Advanced Architectures Project: Proposal1983

Advanced Architectures Project: Draft Proposal
The Advanced Architectures Project which was our parallel computing for AI project ran two proposal terms which means probably somewhere in the neighborhood of seven or eight years. And then it came to a halt when I began to understand that, although we did master in CAGE and POLIGON the art of parallel programming of blackboards, the programming was exceedingly difficult and would not be practical, probably wouldn't be used and, in any case, was not credible because it seemed as if the world was only recognizing as credible those things which had been realized in hardware, not those things which had been realized in hardware emulation. And so we just decided we weren't going to do that anymore and we stopped the project somewhere in the 1990-ish timeframe.


Folder 6: Parallel Architecture for the Blackboard Environment - DARPA Proposal1983

Parallel Architectures for the Blaackboard Environment
Several years after we had become very comfortable with the blackboard architecture as a way of organizing AI knowledge based system knowledge and problem solving; there arose a national effort to build a better computer hardware framework for doing artificial intelligence. In spirit, this was very similar to the Japanese Fifth Generation Project. The goals were quite similar. The American goals were a little broader. The project was sponsored by the Defense Advanced Research Projects Agency and was called the Strategic Computing Initiative. A history of that initiative has been written by an official historian. I've read it. I don't really think it's a very good or accurate history, especially as regards the kind of work that I'm going to be mentioning in a moment. But in any event, in 1983 and 4, this project was being planned and proposals were invited for developing new architectures, particularly highly parallel architectures because we were beginning to enter the era of chips, Intel chips, TI chips, workstation chips, that made it feasible to have a multi computer inside the box. But you needed hardware architectures for organizing the multi computers in the box. And we thought of organizing them using the blackboard framework because there were some inherent features of the blackboard framework and blackboard framework was inherently largely a parallel framework. So we submitted a proposal to DARPA of which there are drafts and copies in this file. And then we proceeded to develop two software versions of the blackboard architecture that would help organize this parallel machine and we also developed some ideas for the parallel hardware. We never built parallel hardware because I kept insisting that the Heuristic Programming Project, later the Knowledge Systems Laboratory, was essentially a software and AI project. It was not a hardware project. We didn't really know about building hardware but we could plan the building of the hardware, which we did. We had a hardware engineer that we borrowed from Digital Equipment Corporation, Bruce Delagi, And Bruce helped us a tremendous amount in organizing our ideas for the hardware. Some of Bruce's inputs are present in this box of papers. We developed two software architectures as I've mentioned. One was called CAGE, which stood for Concurrent Age. Age was the previous but not concurrent blackboard development framework. And then the second one was called POLIGON and I don't remember what the acronym POLIGON stood for at the moment. The prime mover on CAGE was Penny Nii. And the prime mover on POLIGON was James Rice.


Folder 23: Computer Industry Project Progress Report Documentation 1993

In 1993, I joined an ongoing project called The Stanford Computer Industry Project, a project sponsored by the Sloan Foundation and run jointly by the Business School and the Engineering School with participation from the Department of Economics. It was supposed to be a comprehensive industry study similar to the one that the Sloan Foundation had sponsored at MIT concerning the automobile industry. I think that resulted in a book called "The Machine That Changed the World". In any event, we were supposed to do something like that for the computer industry. The project was not moving very well at the time and later was taken over by one of our Stanford hotshots, Professor William Miller, former provost, former President of SRI and after SRI, returned to Stanford in the Business School but he was a computer scientist by training. So he was the ideal person to take on the management of this project. Shortly before Bill took over the project, I formed a sub-project of the Computer Industry Project called the Stanford Software Industry Project. I did, as I've mentioned earlier, numerous interviews in Japan using help from the Stanford students at the Stanford-Japan Center and we did a report on that which is somewhere in the archives and it's published in a 1996 MIT press book called "The Future of Software". When I got back to Stanford, having warmed up on the Japanese industry, I decided it was time to look at the American industry, taught a course on that in probably the fall of '93 and got the Stanford students who were taking the course here on campus to help with interviews for the American industry. We laid out quite a number of people that we wanted to interview. There was a big trip that was involved in interviewing some of these people. Not all of them were willing to be interviewed but we did extensive interviewing. The students wrote some papers having to do with the American software industry. I did some comparisons between the American and the Japanese software industry and was ready to write that up as a second and third article; the American Industry and then a comparison of the two. But those two articles simply never got written because then I got sucked up in this enterprise of becoming Chief Scientist of the Air Force and there was no time to do any articles about the software industry. My attention got focused on other things. So this is the residual material. I had two helpers, really terrific people. One of them was Avron Barr who was a former student and the co-editor with me of four volumes of "The Handbook of Artificial Intelligence". Avron, by that time, had married a Carnegie Mellon computer science graduate named Shirley Tessler. Shirley was another terrific researcher. And Avron and Shirley were running a consulting company and they were hired as consultants by the Stanford Computer Industry Project to help with the software industry work. In fact, they were supervised by Professor Miller who took over the project while I was in Washington.


Folder 25: Avron / Shirley Software Slides1995

The US Software Products Industry: Success Factors In An Era of Rapid Change
In 1993, I joined an ongoing project called The Stanford Computer Industry Project, a project sponsored by the Sloan Foundation and run jointly by the Business School and the Engineering School with participation from the Department of Economics. It was supposed to be a comprehensive industry study similar to the one that the Sloan Foundation had sponsored at MIT concerning the automobile industry. I think that resulted in a book called "The Machine That Changed the World". In any event, we were supposed to do something like that for the computer industry. The project was not moving very well at the time and later was taken over by one of our Stanford hotshots, Professor William Miller, former provost, former President of SRI and after SRI, returned to Stanford in the Business School but he was a computer scientist by training. So he was the ideal person to take on the management of this project. Shortly before Bill took over the project, I formed a sub-project of the Computer Industry Project called the Stanford Software Industry Project. I did, as I've mentioned earlier, numerous interviews in Japan using help from the Stanford students at the Stanford-Japan Center and we did a report on that which is somewhere in the archives and it's published in a 1996 MIT press book called "The Future of Software". When I got back to Stanford, having warmed up on the Japanese industry, I decided it was time to look at the American industry, taught a course on that in probably the fall of '93 and got the Stanford students who were taking the course here on campus to help with interviews for the American industry. We laid out quite a number of people that we wanted to interview. There was a big trip that was involved in interviewing some of these people. Not all of them were willing to be interviewed but we did extensive interviewing. The students wrote some papers having to do with the American software industry. I did some comparisons between the American and the Japanese software industry and was ready to write that up as a second and third article; the American Industry and then a comparison ofthe two. But those two articles simply never got written because then I got sucked up in this enterprise of becoming Chief Scientist of the Air Force and there was no time to do any articles about the software industry. My attention got focused on other things. So this is the residual material. I had two helpers, really terrific people. One of them was Avron Barr who was a former student and the co-editor with me of four volumes of "The Handbook of Artificial Intelligence". Avron, by that time, had married a Carnegie Mellon computer science graduate named Shirley Tessler. Shirley was another terrific researcher. And Avron and Shirley were running a consulting company and they were hired as consultants by the Stanford Computer Industry Project to help with the software industry work. In fact, they were supervised by Professor Miller who took over the project while I was in Washington.


Folder 26: Expert System Workshop

EXPERT SYSTEM WORKSHOP
Now, there was a major workshop held in the Expert System area, just called Expert System Workshop. And it eventually became a book called Building Expert Systems. That is perhaps the most influential book of the Expert System's commercialization. The book is edited by Waterman, Lenat and Hayes-Roth, and this is the workshop material that was handed out to everyone at the time. So it's kind of a historic document in that respect.


Folder 27: Notes on Computer Industry Project Book1994

Notes on CIP Book and March 10th meeting
This is an extended outline. 'm sure these notes were written up by Avron Barr and Shirley Tessler based on a meeting that we had with Bill Miller and myself. And these notes were set up as if they were for a meeting with Bill on March 18, 1994. So - and they're notes from the March 10th meeting and this was a follow-up meeting. What we were talking about was doing a book. The Sloan Foundation really wanted a book. They loved that book from MIT called "The Machine that Changed the World". And the idea was the computer was another machine that changed the world and so this would be like son of machine that change the world and, you know, why aren't you guys at Stanford doing one. So we thought we would get our ideas together for what a book like that might look like. This is a very interesting document because this is the sketch of such a book.
And, in fact, it ends with the last bullet is conclusions about where is the computer industry going. So guess what, this book never got written. Didn't get written by me or Miller or by the Computer Industry Project. And eventually the project just folded I think because Sloan wasn't getting what it wanted and Stanford didn't want to get pushed in that direction. And I was in Washington so things didn't move in the direction of this book and the Computer Industry Project ended after a while.

Box 25 - 2005

Folder 14: Expert Systems (ES) Implimentation Group--Workshop 2 1988

When a field is booming and everyone wants to know about it and all the companies want to understand how to perform well in the area, they want to to collect and use the great stories of companies that are developing the technology; in this case, the technology being expert systems. There are always entrepreneurs, particularly academic entrepreneurs who will organize some workshops. They'll organize sort of a "We'll have this workshop. Come and pay a thousand dollars to hear this workshop event". And I don't know what the price was. I'm just guessing at a thousand dollars or five thousand dollars or something like that but that's what I believe the Expert System Implementation Group was about. There was some kind of Cambridge, Massachusetts academic entrepreneurial organization that brought in well known people. For example, I was the lead-off speaker in this one. And my talk was called "Expert Systems, Payoff and Pointers". And then I held a discussion of key issues. But then a famous guy I believe from the Harvard Business School, Michael Scott Morton, gave a talk on addressing user needs for expert systems. And several other talks. Randy Davis, one of my former students but a Professor at MIT Artificial Intelligence Laboratory, Computer Science Department, gave a talk on the knowledge acquisition process. Anyway, this really does give a good reflection of where things were in 1988, roughly speaking -- or the period 1986 through 1988.


Folder 19: PC--Artificial Intelligence (AI) Publication--Issues regarding Expert Systems (ES), Knowledge-Based Systems 1993--1996

Very early in the history of the LISP language a man named Edmund Berkeley, who had written a very famous book about computers in the early days called Giant Brains, was doing consulting for a company called Information International Incorporated and produced this little textbook on LISP long before the world had recognized how important LISP was. It was published by this company itself in March of 1964, and is a historic little book on early LISP. I'll bet there's hardly anyone in the world who has a physical copy of this book. McCarthy probably threw his away years ago.
This is probably the first book that there ever was on LISP other than the LISP manual itself.

Box 26 - 2005

Folder 18: Future Generation Computer Systems1992

Future Generation Computer Systems
The Japanese invented the term Fifth Generation Computer Systems.
And, in fact, the book that I wrote with Pamela McCorduck was called "The Fifth Generation" just taking off on the Japanese theme. Now of course, when there becomes a sort of international banner that can be flown, somebody's going to take advantage of that for commercial uses. And North Holland Publishing Company (which I'm not even sure exists anymore, may have been bought by Elsevier I think) started publishing a journal called "FGCS" which stood for Future Generation Computer Systems because I guess they weren't allowed to copy the terminology Fifth Generation Computer Systems so they just said FGCS. And this special issue for the conference in Tokyo in 1992. So what is this special conference that this issue about? Ten years after the Fifth Generation Computer Project started in Japan, it was to have finished. And so, for example, this thing starts out with a guest editorial titled "Fifth Generation Computer Systems: Success or Failure". So it's trying to evaluate whether this project was any good or not and the rest of the journal summarizes what was found out in this project. For example, here's somebody from the University of Oregon named Evan Tick giving an appraisal of parallel processing research that was done in this project and so on. Here's an evaluation of the language KL1 and the inference machine that was done to implement that language in hardware. This is an important paper because it's by me and Howie Shrobe of the MIT Artificial Intelligence Laboratory. The "Japanese National Fifth Generation Project, introduction, survey and evaluation". So people were looking to me to give an evaluation of this project since I had written this elaborate book on how great the project was going to be. And so Shrobe and I did a fairly substantial evaluation of this project.

Box 27 - 2005

Folder 8: Japanese Software industry : Siglunch1993

Japanese Software Industry: Where's the Walkman
Cig Lunch was the name of the lunchtime talks that we used to have at our project, a tradition that had gone on for twenty or twenty-five years, no longer takes place. But this was where I gave a talk on coming back from Japan. This was my first chance to talk to my own group about what I found when studying the Japanese software industry. So this is probably one of the first of the talks that eventually became the paper that appeared in the MIT book, "The Future of Software".


Folder 14: Computer Industry Project (CIP) - JSW Talk (EAF Talk)1993

Japanese Software Industry: Where's the Walkman
JSW is Japan Software talk in the name of this folder. This was the time when I was focusing on the Japanese software industry. I had done a study, as I mentioned elsewhere, with the help of students at the Stanford-Japan Center in Kyoto in the spring of 1993. And so I knew a lot about the Japanese software industry at that time. And when we had our annual conference for the computer industry project, kind of an external conference that involved insiders talking to outsiders, I gave a talk on that. And so these two folders contain materials on that. Eventually that was published in a paper about Japan software, "Where's the Walkman"t. It was published in a book called "The Future of Software", by MIT press in January of 1996.


Folder 15: Expert Systems (ES) in Japan--Mizoguchi's Tutorial 1992

ES is expert systems in Japan, 1992. This was a tutorial presented at the World Congress on Expert Systems held in December 1991. And I was quite interested in keeping up with everything that was going on in expert systems in Japan because the Japanese work constituted maybe one third of the book "The Rise of the Expert Company". The guy who regarded himself as the guru of expert systems in Japan was Riichiro Mizoguchi, Professor at Osaka University. And he gave a talk on expert systems in Japan for a tutorial at this conference. And so I kept a copy of it.


Folder 17: Fifth Generation Computer Systems1981

Innovation and Symbol Manipulation in Fifth Generation Computer Systems
When the Japanese launched their famous now infamous Fifth Generation Computer Systems Project under the auspices of a semi-governmental agency in Japan called JIPDEC, Japan Information Processing Development Corporation, they launched it with a big conference at which I was asked to be one of the main speakers. The Japanese Fifth Generation plan that was so much commented upon by many others including myself in a book I wrote called "The Fifth Generation" with Pamela McCorduck. There was such demand for it, it ended up being published in hardback by North Holland Publishing Company which I think doesn't exist anymore. I think it was perhaps bought by Elsevier. The editor of this is the nominal director of the project, Professor Moto-Oka. But it contains not only Moto-Oka's sort of ceremonial introduction but also the introductory speeches of the project director, Dr. Fuchi and some of Fuchi's Chief Lieutenants on this project. And it includes my own paper at this conference which is included in this book.
One additional note about this book: This book is a quite rare thing. I don't know how many copies of this book were printed but I've never seen many around and this is likely to become sort of a historically valuable book.


Folder 18: Fifth Generation - Related1982

Fifth Generation Computer Systems Project
When the Fifth Generation Plan was first floated by the Japanese to many of us, it was floated as a thing called outline of research and development plans for fifth generation computer systems, May 1982. I don't know if this document was produced by an American summarizing what the Japanese were trying to do or by the by the Japanese project institute itself (ICOT). It probably is the latter. But it's a kind of a summary of the outline of their plan.

Outline of Research and Development Plans for Fifth Generation Computer Systems
When the Fifth Generation Plan was first floated by the Japanese to many of us, it was floated as a thing called outline of research and development plans for fifth generation computer systems, May 1982. And that's an original document. This is quite rare. This was a document that strongly motivated my interest in the subject.


Folder 19: Japan--Sixth Generation 1991

This folder contains material on a project that was intended to, as you would expect, follow the Fifth Generation Project. The Sixth Generation Project never got called that, in fact. It got called something like Soft Computing. And the intent of this project was to give research support money over a relatively longish period, I think maybe ten years, to a large group of researchers who were not whetted to the idea that the technique for AI was mathematical logic. That was the idea of the Fifth Generation Project and they were creating special machines to do computation using mathematical logic. Soft computing was meant to be a contrast to that. The logic stuff was hard computing. And the Sixth Generation Project got launched and there were many projects within it. This folder contains descriptions of some of that work.


Folder 22: Kansai Silicon Valley Forum1998

Letter from Shingo Yabuuchi to Mr. Edward Feigenbaum
There's a folder called "The Kansai Silicon Valley Forum". The Kansai area is the area around Osaka, Japan, Kyoto, Japan. Nara, I guess is in the Kansai. There was a group that was organized to advance the state-of-the-art in entrepreneurship in the Kansai area. Somebody had funded that. I guess some Osaka prefecture or somebody had funded the Kansai Silicon Valley Venture Forum. Forum means it's a collection of people who were exchanging ideas about how to do this and they asked some Americans to be consultants to this forum. And I agreed to do it. And this is a letter which said this person expresses his appreciation that I joined the forum and they'll be sending some small honorarium payment.


Folder 24: Japan Technical Evaluation Comittee (JTEC) Substantive 1991--1992

JTEC was a project of the National Science Foundation but the National Science Foundation is a funder of things, not a doer of things. So it had funded a center called Japan Technical Evaluation Center. It eventually moved in under a university called Loyola University, but there was somebody in charge of the computer science part of this center, Dr. Shelton. I think the prime mover behind the whole idea was Dr. George Gamota of Mitre Corporation. The list of these people is in the binder. They had asked me to, as I mentioned in previous notes, run a panel study of how expert systems and knowledge based systems were going in Japan. This study was, I think, in 1993. There is a file folder here called JTEC Substantive which is the notes of my group as it went around and divided up the task of seeing lots of Japanese corporations and finding out what they were doing in the expert systems area. In the end, there was a report which was about a quarter of an inch thick. There was an executive summary of it which appears in this pile of information. There are other panels, one that dealt with machine translation of languages and another which dealt with advanced computing technology in general but had two artificial intelligence people on it, Stan Rosenschein and Marty Tenenbaum.

News from Kahaner/Japan
David Kahaner was an employee of the Tokyo office of the Office of Naval Research. He took it upon himself -- in fact, it was his job -- to be familiar with all kinds of scientific and engineering activity going on in the information technology area, perhaps even a few other areas related to computers in Japan and in Asia. So David would travel around, interview people and then write them up in really an excellent way. And many of us were not only on his email loop but we were really fans of David Kahaner. This thick folder is a bunch of email printouts of David Kahaner's kind of state-of-the-art and information technology in Japan and other places in Asia around 1992 plus or minus a year.

Box 28 - 2005

Folder 7: Applications of Artificial Intelligence for Organic Chemistry: The Dendral Project - by R.K. Lindsay, B.G. Buchanan, E.A. Feigenbaum, J. Lederberg1980

Applications of Artificial Intelligence for Organic Chemistry
Yes, I'm looking in Box #28, there's a book by Lindsay, Buchanan, Feigenbaum and Lederberg, "Applications of Artificial Intelligence for Organic Chemistry, the DENDRAL Project". This is a very important book. There were many, many papers that came out of the DENDRAL Project. The DENDRAL Project, of course, was a landmark project, not only in my own career, but also essentially the field of expert systems was invented with the DENDRAL Project and it was a landmark project in the field of artificial intelligence. And indeed, there is currently in this period of 2004-2005, a Ph.D. thesis underway by a graduate student in Princeton University on the history of the DENDRAL Project. Well the DENDRAL Project was reasonably old by around 1975-6. An artificial intelligence scientist who was a college friend of mine, a graduate school friend of mine and an undergraduate school friend of mine, Bob Lindsay, University of Michigan, came out to do a sabbatical here at Stanford in the DENDRAL Project. And Bob was intrigued by the project. He thought it was truly wonderful and got excited enough to want to do a book on it. He collaborated with myself -- he basically did the writing of the book but he collaborated with Bruce Buchanan, myself, Joshua Lederberg and the result was this book on the DENDRAL Project which is sort of what you might call the last word - even though it's *incomplete*


Folder 8: Chemometrics Tutorials1990

Dendral and Meta-Dendral -- The Myth and the Reality
In here is a debate between a fellow Neil Gray, and some of DENDRAL scientists who responded to Neil. The title of Neil's paper is "DENDRAL and Meta-DENDRAL, the Myth and the Reality". So it's kind of a tutorial on DENDRAL but as it says in the title, it's kind of a downer. Part of it is what's the myth about it. Well we didn't really think that there was a myth about it and we were given an opportunity to answer Neil and the comments were by Buchanan, Feigenbaum and Lederberg and then Neil gave a response to the comments by Buchanan, Feigenbaum and Lederberg. So it was quite a debate about the meaning and importance of the project. Our title is called "On Gray's Interpretation of DENDRAL Project and Programs: Myth or Myth Understanding" which we thought was cute. But anyway this is a very interesting exchange of views.


Folder 9: Darpa Interviews: Saul Amarel1986

DARPA Interview of Saul Amarel by Robert Engelmore
My colleague, Bob Engelmore, who is now deceased, worked at our project for a very long time, since late 1960's or early 1970's, with some interruptions. And one of his interruptions was that he worked at the Defense Advanced Research Projects Agency for a couple years, two or three years. Later on, Bob was asked as a consultant to help do some interviewing for - I'm not sure if this was a book or what it turned out to be as a publication but DARPA was really proud of its role in helping to support and shape the history of computer science and technology, not just in the U.S. but worldwide. And therefore it wanted to capture the history of that in interviews with key people. Bob Engelmore participated in that interviewing and some of the interviews are here. This is interview material with Saul Amarel who is a well known artificial intelligence scientist who, at one point, was Director of the Information Processing Techniques Office or whatever it morphed into at that time. The name kept changing over the years but he was Director of it.


Folder 11: Darpa Interviews: Cordell Green1986

Darpa Interview - Cordell Green
I'm not sure if this was a book or what it turned out to be as a publication but DARPA was really proud of its role in helping to support and shape the history of computer science and technology, not just in the U.S. but worldwide. And therefore it wanted to capture the history of that in interviews with key people. Cordell Green was a program manager at DARPA probably during the seventies. But Cordell was also a Ph.D. student of mine and a Stanford graduate and an Assistant Professor here at Stanford and the founder of a research institute called The Kestrel Institute.


Folder 12: Darpa Interviews: Robert Kahn1986

Darpa Interview - Robert Kahn
I'm not sure if this was a book or what it turned out to be as a publication but DARPA was really proud of its role in helping to support and shape the history of computer science and technology, not just in the U.S. but worldwide. And therefore it wanted to capture the history of that in interviews with key people. There's also an interview here with Kahn, Robert Kahn, the co-inventor of the internet and director of this office for a long time and the 2005 winner of the ACM Turing Award.


Folder 13: Darpa Interviews: J.C.R. Licklider1986

Darpa Interview - J.C.R. Licklider
I'm not sure if this was a book or what it turned out to be as a publication but DARPA was really proud of its role in helping to support and shape the history of computer science and technology, not just in the U.S. but worldwide. And therefore it wanted to capture the history of that in interviews with key people. There's an interview here with Licklider, JCR Licklider, who was the Founding Director of that office and the subject of a book on the work of that office. Actually this whole story was told in a superb book by Mitchell Waldrop, called "The Dream Machine" and ostensibly that's the life of JCR Licklider.


Folder 14: Darpa Interviews: Larry Roberts1986

Darpa Interview - Larry Roberts
I'm not sure if this was a book or what it turned out to be as a publication but DARPA was really proud of its role in helping to support and shape the history of computer science and technology, not just in the U.S. but worldwide. And therefore it wanted to capture the history of that in interviews with key people. This is an interview with Larry Roberts who, as office director, supervised the introduction of the first packet switched network, the ARPANET which was the predecessor to the internet.


Folder 26: NLM - Board of Regents1986

Letter from Donald A. B. Lindberg, M.D. to Edward E. Feigenbaum, Sc.D.
In the period 1986 through 1990, I served as a Regent of the National Library of Medicine. The governing board of the Library is called the Board of Regents. It consists of eminent people from various areas, not just medicine, not just library work. In my case, it was computer science. My computer science seat on the board, so to speak, was taken after I finished in 1990, it was filled with Bob Kahn, the person I mentioned before as being the co-inventor of the internet. So I served for four years which meant traveling to Washington a couple times a year for Board of Regents meetings and having a lot of correspondence and interaction with the National Library of Medicine. The closest part of the interaction was with Dr. Donald Lindberg, the Director of the National Library of Medicine who was an old friend of mine from the early days of artificial intelligence and from early work with peer review for the NIH. So I was pleased to serve when Don was the Director.
During that period, Don led an effort to put together a long range plan for the library. Now that was '86 through '90. So by now, we're approaching twenty years into the long range plan, believe it or not, and just recently as of April of 2005, there was another meeting, a new long range planning committee, and I was invited by Don who's still the Director of the library to attend and help to plan for the 2006 through 2016 period.

Box 29 - 2005

Folder 6: LISP--List Processing--Hofstadter "Scientific American" Articles 1983

There are three issues of "Scientific American" here in the folder. They're the February, March and April issues of "Scientific American". There's a series of articles in here by a well known writer, a Pulitzer Prize winner, Douglas Hofstadter. Doug Hofstadter was student here at Stanford, took a course from me, and was involved in many projects in artificial intelligence and became a professor in that area elsewhere, not at Stanford. But he's a very, good writer and he wrote three articles called "The Beauty of LISP". LISP is a programming language. The LISP notation is due to Professor John McCarthy of Stanford University, actually did it I think when he was at MIT. It's based on ideas that were developed at Carnegie Mellon University and Rand Corporation by Newell, Shaw and Simon. But it became the most popular programming language for artificial intelligence programs. And it turns out to be very beautiful as a programming language, probably the most beautiful, elegant programming language. So Hofstadter wanted to tell people that and in what sense it was a beautiful programming language. And he chose the "Scientific American" where he had a relationship with them in writing as his vehicle.


Folder 8: SUMEX - AIM - "The Seeds of Artificial Intelligence"1978 - 1980

The Seeds of Artificial Intelligence - SUMEX-AIM
The acronym SUMEX AIM I've mentioned before, Stanford University Medical Experimental Computer System, of course, Artificial Intelligence in Medicine. That was a national resource supported by the NIH for not only supporting the Stanford work in artificial intelligence in medicine but also a national community. The NIH Division of Research Resources who supported this was proud of the excellent work that it was doing and put out a pamphlet, brochure, on the nature of the work. It was called "The Seeds of Artificial Intelligence" and it related to what work the SUMEX AIM Project was doing to support artificial intelligence work. That brochure was very widely distributed and contains excellent historical material on the projects we were doing here at Stanford plus other thing going on around the country, rather important document.

TV-EDIT Reference Card
This was a reference card for the users of a word processing system.This is what we used to have before people had Microsoft Word. This was the kind of thing. It was these interactive editing word processing systems in a timesharing mode like TV-edit. TV-edit may have been done by our SUMEX AIM people themselves or it may have been adapted from some other work that people had done on the Digital Equipment Corporation PDP 6 or PDP10. I'm not sure. It says here TV-edit was developed by the Institute for Mathematical Studies in the Social Sciences at Stanford University. That institute no longer exists. I think it merged with the Center for the Study of Language and Information. I'm not positive of that. And they even named who were the writers of it, Brian TolliVer, John Preyis and Pentti Kanerva. And it was supported by SUMEX AIM. So yes, it was developed here at Stanford. But that's what the world was like before Word.


Folder 10: Alto User's Handbook 1979

This is a very famous machine. The Alto was essentially the manifestation of the first personal computer. This was developed at Xerox, and some of our Stanford people were consultants over there. We, of course, had Xerox Alto envy. We wanted machines like this. We actually were able to buy a few of these Also machines. And we were given some Ethernet parts that were also developed at Xerox PARC. Our connecting up of these Altos with other machines via the Ethernet actually led to a spin-off company which is now called Cisco. This Alto user's handbook plus its insert which is a little addendum written to update Stanford people on the Alto for the use on campus probably is a historically very important document but it's probably available also at the Computer History Museum and undoubtedly at Xerox Palo Alto Research Center. It's no longer Xerox incidentally. It's the Palo Alto Research Center -- Xerox spun it off.


Folder 11: TENEX Executive Manual 1973

The TENEX was a project supported by the Defense Advanced Research Projects Agency to provide a modern operating system and an operating system research environment for the Digital Equipment Corporation Time Sharing systems that were produced on either the PDP 6 or the PDP10 or both of those. And TENEX then became an operating system used by most of the users of these advanced users of time sharing during the period of the late sixties and through the seventies. SUMEX AIM was one of those computer facilities that was using the TENEX operating system. So this was the manual and then some people produced additional material called "Joy of TENEX", "The Joy of TENEX". And here's another manual which describes some additional things. For example, additional text editors that had been added or mail systems. When almost nobody in the world knew about electronic mail, here's a description of a couple different mail systems. XXLL formatting program is how you would translate from the word processor to a format that was printable on these new fangled things called optical printers which later on became laser printers. There's a thing called the Telnet User's Guide. Telnet is the way that you would interact from your own terminal with a very early version of what later became the internet. It was called the ARPANET. And how you would send files around was called File Transfer Protocol. This is a description of all of those. So this is really amazing stuff.


Folder 13: The Heuristics of George Polya and its Relation to Artificial Intelligence Allen Newell1981

The Heuristic of George Polya and Its Relation to Artificial Intelligence
There's a paper here by Allen Newell dated July 1981 and it says that it will appear in a collection, Groner and Bischoff, "Methods of Heuristics". He wrote this for an international symposium on the methods of heuristics at the University of Bern in Bern, Switzerland. But although I think I know most of Newell's stuff very well, I just have completely forgotten this paper. The paper is called "The Heuristic of George Polya and its Relation to Artificial intelligence". And it's likely to be a very important paper.
Polya, in his later life, was a famous Stanford mathematician. In his earlier life, he was one of the great mathematicians of the twentieth century. Came to Stanford for the last third of his career and he popularized the idea of heuristic reasoning. Allen Newell was an undergraduate student at Stanford and learned about this from Polya and then went on to apply these ideas in founding artificial intelligence.

Box 31 - 2005

Folder 8: PUB Manual (Stanford Artificial Intelligence Project Operating Note 70) 1973

PUB The Document Compiler - PUB Manual (Stanford Artificial Intelligence Project Operating Note 70) 1973
Interesting that the documented author is Larry Kessler, who was at the AI Lab at the time, and Larry is now one of the big hotshot scientists for Apple. Larry helped to bring the Apple Macintosh into existence.

Box 32 - 2005

Folder 9: Raj Reddy Letter1998

Letter from Raj Reddy to Ed Feigenbaum
In a letter dated June 25, 1998, Raj Reddy of Carnegie Mellon is thanking me for participating in a symposium called "Inventing the Future", June 4 and 5,1998. And thanking me for giving a stimulating opening talk.

Box 33 - 2005

Folder 14: Jonathan King - Thesis Proposal and Notes

USING PRAGMATIC KNOWLEDGE TO INCREASE DATABASE RETRIEVAL EFFICIENCY
These are the notes of the thesis by Jonathon King. Jonathan King's thesis was on influential processes related to getting things out of the big conventional database. That's a hot topic now in the commercial world of AI, but it was new at the time. Jonathan King works at Tech Knowledge right now. These are the thesis proposal and notes being built up along the way for his thesis.

Box 34 - 2005

Contains Ed's papers that he published and towards the back of the box are biographical materials about Ed

Folder 3: Photos - EAF1992-96 and n.d.

Various Feigenbaum photos, during Feigenbaum's tenure as Air Force Chief Scientist + photos of Harold Cohen's art images+ other
The photo of the computer scientist, Fred Brooks: Fred is one of the most famous computer engineers and computer science people who have ever lived. He's the winner of everything. He's one of the top figures in the IBM 360 system development and he wrote a wonderful book called "The Mythical Man-Month" which is part of the collection. Well anyway, it's in that collection of books that Henry was looking at. Second comment was the picture of me in Antarctica was undoubtedly taken in I believe January 1996 or December 1995, when the family and I took a tourist trip to the Antarctic Peninsula. I was also in Antarctica a second time on a official government trip to the South Pole and to the McMurdo Station. That was on invitation of the National Science Foundation and on that trip, I accompanied the Chairman of the National Science Board, Professor Richard Zare, who is also a Professor of Chemistry at Stanford. So we knew each other from Stanford. Finally there's a picture of me in the second seat of an Air Force jet fighter fully adorned with flying gear and oxygen mask and helmet and all that. I can't remember if this was my trip in the back seat of an F15 jet fighter or whether this was in the F16 trainer version. F16 normally has one seat but the trainer versions have two seats. But as Air Force Chief Scientist, they gave me a ride.


Folder 19: EAF - Large Knowledge Bases - How Things Work1990

Large Knowledge Bases for Engineering: The How Things Work Project of the Stanford Knowledge Systems Laboratory
This was a project for which we were well funded by DARPA, NASA and other sources that are listed in these papers. The 1990 paper appears to be what you would call a white paper.
That is, laying out in a relatively brief way a sketch of the ideas that were motivating this research and how the research would be carried out. It was co-authored with Robert Engelmore and two younger people we had hired, Thomas Gruber and Yumi Iwasaki, dated November 1990. Well apparently things worked out pretty well because by '92 we're issuing project reports, in this case written by Richard Fikes who's the current Co- director of the Knowledge Systems Laboratory. And then there are some view graphs that Fikes put together for some presentation. So things were moving along pretty well in the How Things Work project and I think that project accomplished a substantial amount. It finally ground to a halt when Tom Gruber left the project to join what was, at that point, clearly the beginning of the dot com revolution. And subsequently Tom's co-founded a company with another one of my students, Peter Friedland, Introspect. Yumi left to go to Seattle. She married one of the assistant professors, Anoop Gupta who went to work for Microsoft and they moved to Seattle. So the project really came to a halt around that time.


Folder 22: EAF - A Personel View of Expert Systems: Looking Back and Looking Ahead1992

A Personal View of Expert Systems: Looking Back and Looking Ahead
The paper is footnoted as an acceptance speech for the Feigenbaum Medal presented at the World Congress on expert systems at Orlando, Florida, December 1991. That was the first presentation of the Feigenbaum Medal which was started by this organization, The World Congress on Expert Systems. And this organization ground to a halt sometime in recent years. So the Feigenbaum Medal was awarded several times but as of the middle of 2005 which is now, it is no longer being awarded. However, I was the first winner of the Feigenbaum Medal and so this was my speech. And again, because it was a very prestigious thing, I worked pretty hard on this speech. It was in the form of an intellectual memoir. For example, one of the subheadings is "Stumbling in on the Birth of AI, The Miracle Years at Carnegie" and so on. So it's a memoir-like paper. The Chairman of this conference was Professor J. Leibowitz whose father subsequently went on to be President of the National Academy of Engineering. And he was there at the conference and he told me afterward he thought it was one of the best talks he had ever heard.


Folder 23: EAF - Dendral and Meta-Dendral: Roots of Knowledge Systems and Expert System Applications1993

DENDRAL and Meta-Dendral: roots of knowledge systems and expert system applications
In choosing the title of this paper, I and Bruce Buchanan made sure that the reader would immediately make the connection between this earliest of the knowledge based systems, DENDRAL and Meta DENDRAL and the subsequent developments in knowledge systems and expert systems that had happened in the roughly twenty-five years from the time DENDRAL was started till 1993. We wanted to make sure that the reader understood the connection. This article was published in the main journal of the field, "Artificial Intelligence" and that was a special collection put together by the co-editors of the journal, one of whom was Daniel Bobrow. The co-editors looked back over the history of the field and asked themselves what were the landmark papers or what were the landmark projects in the field and let's get an article on each of the landmark projects. And that would be the territory of AI laid out in terms of these landmarks. They asked Buchanan and me to write this article on the DENDRAL and Meta DENDRAL project as one of those landmarks. And so this paper appeared in that volume. That volume is an orange covered volume in a very nice cardboard box that we have in the collection that is edited by Bobrow and Pat Hayes.


Folder 24: EAF - DENDRAL: A Case Study of the First Expert System for Scientific Hypothesis Formation (1 of 2)1993

Final: DENDRAL: A case study of the first expert system for scientific hypothesis formation
There's a thick folder here which contains materials related to a paper in "The Artificial Intelligence Journal" which is this main journal of the field called "DENDRAL: A Case Study of the First Expert System for Scientific Hypothesis Formation". It almost coincides with the previous paper that I discussed "DENDRAL and Meta-DENDRAL" but this one doesn't appear to be an invited paper. This appears to be a submitted paper which was written by one of our long term collaborators, Robert Lindsay, University of Michigan. As - as you know from the archive and has been commented on before, there's a book in the archive, what's called "The DENDRAL Book". And its first author is Robert Lindsay who had come out on sabbatical and took as his task the DENDRAL project as a book. Now what it is that motivated Bob Lindsay to put this together as a paper in 1992, I don't know because the DENDRAL project was started in 1965. It kind of came to an end in around 1983 sometime and had become part of history. So I'm not exactly sure what it is that got Lindsay to write this. Now it could have been at the invitation of the editors of the "AI Journal" who were just coming to the realization of how important DENDRAL had been. So maybe they stimulated it.


Folder 26: EAF - The Japanese National 5th Generation Project: Introduction, Survey and Evaluation1993

The Japanese National Fifth Generation Project: Introduction, survey, and evaluation
The Japanese invented the term Fifth Generation Computer Systems. And, in fact, the book that I wrote with Pamela McCorduck was called "The Fifth Generation" just taking off on the Japanese theme. Now of course, when there becomes a sort of international banner that can be flown, somebody's going to take advantage of that for commercial uses. And North Holland Publishing Company (which I'm not even sure exists anymore, may have been bought by Elsevier I think) started publishing a journal called "FGCS" which stood for Future Generation Computer Systems because I guess they weren't allowed to copy the terminology Fifth Generation Computer Systems so they just said FGCS. And this special issue for the conference in Tokyo in 1992. So what is this special conference that this issue about? Ten years after the Fifth Generation Computer Project started in Japan, it was to have finished. And so, for example, this thing starts out with a guest editorial titled "Fifth Generation Computer Systems: Success or Failure". So it's trying to evaluate whether this project was any good or not and the rest of the journal summarizes what was found out in this project. For example, here's somebody from the University of Oregon named Evan Tick giving an appraisal of parallel processing research that was done in this project and so on. Here's an evaluation of the language KL1 and the inference machine that was done to implement that language in hardware. This is an important paper because it's by me and Howie Shrobe of the MIT Artificial Intelligence Laboratory. The "Japanese National Fifth Generation Project, introduction, survey and evaluation". So people were looking to me to give an evaluation of this project since I had written this elaborate book on how great the project was going to be. And so Shrobe and I did a fairly substantial evaluation of this project.


Folder 27: EAF - Knowledge Based Systems Research and Applications in Japan, 19921994

Knowledge-Based Systems Research and Applications in Japan, 1992
Summer 1994 issue of "AI Magazine". This just represents the published form of the study that I chaired and a number of us went on which was that study I've been calling the JTEC study, the Japanese Technical Evaluation Center, an NSF supported operation that was evaluating Japanese scientific and technical efforts.

Box 35 - 2005

Folder 3: Expert Systems and Artificial Intelligence Symposium: Proceedings - contains "Knowledge Processing from File Servers to Knowledge Servers" by EAF1988

Knowledge Processing: From File Servers to Knowledge Servers
This is the proceedings of the third Australian conference on applications of expert systems. It's the 13th through 15th, May 1987. The conference Chairman was Ross Quinlan. There were other people listed here, some of whom are now quite well known in Australia, like Professor Claude Sammut. This was a paper that I had been publishing in various variations about the AI dream, how we tried to achieve that dream and what the long range future might be. And I call the far side of the dream, the library of the future. And so this is my so called library of the future paper that has appeared in several places.


Folder 9: Expert Systems and Artificial Intelligence (AI) Toolkits1988

Knowledge Engineering Environment: KEE
There's a book here called the "Expert Systems and AI Tool Kits". It seems to be edited by a person Tod Loofbourrow, who is still an active participant in the knowledge based system field in East Coast area around Boston. He edited a volume for a thing called the James Martin Productivity Series. James Martin was a famous producer or writer up of technical information, lots of great industry trade books, textbooks, things like that. And so Jim Martin was a - had a very big name and organized a lot of flow of information and knowledge about computing stuff through is - using his name and through his companies. This one was about expert systems and the different software development tools that were available. For example, I'm looking at page 1 of a tool called ART. That's Inference Corporation's Automated Reasoning Tool. That was one of the top products that were available at the time. And then it gives product information and then on page 5, it gives the view of the product that the Martin Report had, the Martin Report being this James Martin industry newsletter, telling people what's new and what's good and almost like a consumer report for technical people. This book is just full of everything that was available at the time. It's just an amazing collection. Here's one for the AION Development System, a system done here in Palo Alto by Harry Reinstein who initially worked at IBM in Palo Alto and collaborated with us at Stanford on an early project on hardware disk diagnosis. So for any scholar who's interested in finding out exactly what was available and what the best reviewers of the time thought of it, this volume is a treasure trove.

Box 36 - 2005

Folder 5: Randall Davis Papers 1975--1983

One of the most important Ph.D. theses done at the Heuristic Programming Project was the thesis by Randall Davis, Professor Randy Davis now at MIT. He's been at MIT since he left Stanford. And a copy of Randy's Ph.D. thesis is here in the box. And I think it's an important document.

Box 38 - 2005

Folder 10: Building Blocks Artificial Intelligence book project 1986-1987

Once upon a time, there was supposed to be a collection of papers. It was supposed to come out somewhere around 1987. It was supposed to be a collection of papers by people whose signatures were included on the cloth of a quilt called "Building Blocks" that my wife, Penny Nii, who's not only an AI scientist but also a quilter, had done for the thirtieth anniversary of AI which was in 1986. And this quilt eventually hung at the computer museum in Boston for several years. Then when the computer museum moved out to Palo Alto and became the Computer History Museum, the original computer museum returned the quilt to Penny Nii and she still has it. Well there was supposed to be a matching anthology to this book and there was actually a poster made of the quilt and advertising the book to come. And there are still come copies of the poster. A lot of people have the poster. But the book never came out. That's because I never finished it. This in here is one letter that responds to requests that Penny made to the various people who signed the quilt for what is the most important paper that you have written so far? This is the one that we're going to consider including in the book, "Building Blocks". And Ross Quinlan who is one the leaders in machine learning research is a Professor of Computer Science at Sydney University in Australia but has been at various places at various times, including visiting at Stanford, wrote this letter on March 24th, 1986 saying which paper he considered to be his most important paper for including in this volume.

Box 39 - 2005

Folder 8: DENDRAL-64 Part 1: Topological Mapping of Organic Molecules1964

DENDRAL-64: A system for computer construction, enumeration and notation of organic molecules as tree structures and cyclic graphs
There two papers: DENDRAL-64 Part I, and DENDRAL-64 Part II, authored by Joshua Lederberg. They're NASA reports from Lederberg's NASA project in the Department of Genetics. They're dated December 15th, 1964, and they deal with the computer enumeration of organic molecules as tree structures and cyclic graphs. Now these papers were written before I arrived at Stanford and before Lederberg and I began our collaboration on the very well known AI program called Heuristic DENDRAL, or DENDRAL for short. Now Lederberg invented the term DENDRAL, and this is the work he had done on DENDRAL before I got here.

Box 40 - 2005

Folder 14: Washington D.C. - Testimony to Congressional Committee on Science and Technology1983

Congressional Testimony on Science and Technology
The hearings were June 29 and 30,1983. '83 was the time when we in the U.S. were beginning to panic about the Japanese taking over so many different sectors of our economic wealth. And I had achieved a certain visibility by talking about a new Japanese project in the computer science area that followed some Japanese projects that were done in the 1970's. During that month of June 1983, this turned into a book called "The Fifth Generation" and I was asked to testify before congress on that subject.

Box 46 - 2005

Folder 6: An Information Processing Theory of Verbal Learning - EAF's Carnegie Mellon Thesis, published by RAND Corporation1959

"An Information Processing Theory of Verbal Learning" - EAF's Carnegie Mellon Thesis, published by RAND Corporation1959
My thesis was finished at the Rand Corporation that summer. It was defended in September of '59, on the day before I was scheduled to go to England on my Fulbright Fellowship. I flew from Pittsburgh to New York to catch the S.S. United States to go to England on my Fulbright. That means, from my point of view, and from Simon's point of view, I was finished in September of '59. But since Carnegie Institute of Technology did not have a graduation until June of 1960, they think I'm in the class of 1960. Hence, that's why the discrepancy appears on various curriculum detail between me saying that I was there from '56 to '59, and Carnegie Mellon saying I was there until 1960. And sometimes I use one date and sometimes I use the other, but it's one of these pieces of confusion in my life.


Folder 31: Faculty Meeting

I think for the historical record I ought to say that we had a very extraordinary collection of young assistant professors. During the time when computer science was emerging from the woods and becoming a discipline, George Forsythe recognized the appropriate timing and started our department as essentially the first, although there's some argument about who's exactly the first. But if it wasn't the first it was one of the first two departments. And so at the time that computer science was coming into its own we had an extraordinary collection of young people.
Klaus was one of them. What can you tell at the time other than he was very bright, but then he turns out to be one of the major figures in the history of the field. Raj Reddy was another one. The Director of the Robotics Lab at Carnegie Mellon, and a truly prestigious and extraordinary computer scientist. Joyce Feldman, currently at the University of Michigan, another one. In fact one of the few female faculty members that we ever had. We ran into a tough problem with these young people. The field was bursting out. It was like a baby grows very fast at the beginning of life. So you had computer science area was bursting out, growing very fast. We had people who two or three years into their assistant professor career were being offered tenure jobs elsewhere. There was no way Stanford could cope with that. Stanford simply can't cope with the idea of promoting someone to tenure after two to three years. So George Forsythe kept telling these people hang in there. We'll try to accelerate it early, but accelerating means you accelerate it one year early around Stanford. Well, that wasn't good enough for these aggressive young people, and one by one they just took these other offers and left. We lost that collection of young people. Of course we replaced them, but you look back on that. Raj went to Carnegie Mellon because that was clearly another great place. But Joyce Feldman went to Michigan, and that was a place which was waning at the time, and waned significantly more afterward. And Klaus went back to ETH in Zurich, because he wanted to get back to Switzerland and they wanted him to help get it started.

Box 47 - 2005

Folder 26: EShell 1986-1987

What's interesting in here is the fact that Fujitsu Laboratories licensed some software that we did at the Heuristic Programming Project called AGE. And there's the Office of Technology Licensing and Fujitsu have a correspondence exchange about AGE. Fujitsu paid us for the use of that software; I think $50,000, one time payment. I met with the people at Fujitsu who are very good friends and they were using our ideas in a piece of software called E-shell. And E-shell was based on AGE and so I just talked to them and said, you know, we really did this here and you just used our material and we're not going to press you for a payment but I think it would be a good idea if you did. It would be nice. We use the research money in our laboratory. And so they did. It was a very gentle form of "we think you ought to do the right thing and license the software".

Box 51 - 2005

Folder 3: EAF Printed CorrespondenceJan.-Dec. 1988

Letter from Edward A. Feigenbaum to Gloria Waterman
Gloria and Don Waterman: Don Waterman was a student of mine when I was teaching at Berkeley and then he came down here to Stanford with me when I transferred in early 1965. And he finished his thesis here. It was a celebrated and superb thesis on machine learning. Don later went to work for the Rand Corporation but he contracted lung cancer at an amazingly young age. Could easily be because his wife was an avid smoker. His wife is Gloria. Anyway, Don died of lung cancer. And we honored him here by setting up the Don Waterman Memorial Library with gifts from a few individuals and the books from Don's library and some books of some of the people's libraries around here. And we had that at the at the Heuristic Programming Project Knowledge Systems Laboratory for some time.
Then we all moved over to this Gates Building and I'm not sure what happened to the Don Waterman Library, whether it was incorporated as part of the library on the second floor of the Gates Building or not.

Letter from Edward A. Feigenbaum to Masafumi Minami
Masa Minami was a visiting Sony person and a Master's student in AI at the time. And Minami has remained a very close friend. He's now a middle level manager at Sony and I see him usually once or twice a year when I go to Japan.

Letter from Edward A. Feigenbaum to Yumi Iwasaki
Yumi was a superb Japanese student who came here in computer science. Actually she was an undergraduate student in another university before she came to Stanford as a graduate student in computer student. I think it was Miami of Ohio or some other smaller college in Ohio. And then subsequently she was another person I would call a Feigenbaum discovery. I discovered her in the Computer Systems Program here and convinced her her to switch from the Master's Degree in Computer Systems to a Master's in a program that I had originally envisioned and started here at Stanford called the MSAIProgram, the Master of Science in Artificial Intelligence. And I convinced her to switch into MSAI Program. She turned out to be a superb Master student and then went on in the MOLGEN Project which was our initial attempt to apply expert system technology to molecular biology, molecular genetics. I've talked about that before. Yumi then went to work for the company that I started, TechKnowledge, after she got her Master's Degree. She worked for TechKnowledge for several years. Then she went to Carnegie Mellon University to work with Herbert Simon Who was my own PhD mentor. And he was her mentor also. So he not only mentored one generation, he mentored the next generation as well. And she got her PhD working with Herb Simon in the late 1980's and then came back to my laboratory as a research assistant, later senior research assistant, working on the project here that we call the "How Things Work" project. And then Yumi was moving along beautifully until she decided to marry a Professor, Anuj Gupta who had come from Carnegie Mellon. And after a while here, Anuj got a very important job. Bill Gates bought - at Microsoft - his little company that he had started off on the side and he got a very important job at Microsoft sort of running the show in the area that his company was involved with and then Anuj went on to become Bill Gates' special assistant and now is a Vice President of Microsoft. And so Yumi moved to Seattle and started to have kids and stopped doing computer science.


Folder 9: EAF Printed CorrespondenceMar. 1981-June 1982

Resolution of Space Discussions
Yumi was a superb Japanese student who came here in computer science. Actually she was an undergraduate student in another university before she came to Stanford as a graduate student in computer student. I think it was Miami of Ohio or some other smaller college in Ohio. And then subsequently she was another person I would call a Feigenbaum discovery. I discovered her in the Computer Systems Program here and convinced her her to switch from the Master's Degree in Computer Systems to a Master's in a program that I had originally envisioned and started here at Stanford called the MSAI Program, the Master of Science in Artificial Intelligence. And I convinced her to switch into MSAI Program. She turned out to be a superb Master student and then went on in the MOLGEN Project which was our initial attempt to apply expert system technology to molecular biology, molecular genetics. I've talked about that before. Yumi then went to work for the company that I started, TechKnowledge, after she got her Master's Degree. She worked for TechKnowledge for several years. Then she went to Carnegie Mellon University to work with Herbert Simon Who was my own PhD mentor. And he was her mentor also. So he not only mentored one generation, he mentored the next generation as well. And she got her PhD working with Herb Simon in the late 1980's and then came back to my laboratory as a research assistant, later senior research assistant, working on the project here that we call the "How Things Work" project. And then Yumi was moving along beautifully until she decided to marry a Professor, Anuj Gupta who had come from Carnegie Mellon. And after a while here, Anuj got a very important job. Bill Gates bought - at Microsoft - his little company that he had started off on the side and he got a very important job at Microsoft sort of running the show in the area that his company was involved with and then Anuj went on to become Bill Gates' special assistant and now is a Vice President of Microsoft. And so Yumi moved to Seattle and started to have kids and stopped doing computer science.


Folder 12: EAF Printed CorrespondenceMay 1979-Oct. 1980

Letter from Edward A. Feigenbaum to Hisao Yamada
Professor Yamada made several visits to Stanford. In 1979, in fact, he and I essentially exchanged jobs. I went to teach at the University of Tokyo in the Information Science Department for one quarter and he came to Stanford to visit here and work on some research I think on his work on electronic type script for Japanese, and we exchanged homes. My family and I lived in his home in Abiko outside of Tokyo. And he lived in my campus home. Hi Yamada has been a friend for years and years and years. And this is one of the few what you might call intersecting friendships between my wife Penny Nii and myself. Penny was a friend of his when Penny worked for IBM Research back in the period of about 1964. Hi Yamada was at IBM Research working on theory. Penny was at IBM Research working on graphics and also working on the Russian to English translation project for the New York World's Fair. And they knew each other back then as they were among the few Japanese who were in the United States doing research at that time. And then I knew Hi Yamada completely independently in 1972 beginning with my first trip to Japan in 1972. And he was one of the organizers, I believe, of the IEEE Conference - Computer Society Conference that I went to during that year. And we've remained good friends ever since.


Folder 15: EAF Printed CorrespondenceDec. 1985-Oct. 1986

Letter from Edward A. Feigenbaum to Hisashi Shinto
I moved to the front of the folder a letter from an executive at NTT, the Japanese Telephone Company. It's from Mr. Kuroyanagi. This letter says he is sending me two photos he took when he met me at the Hyatt Hotel near the San Francisco airport. Obviously the key to this was his desire to send one of his researchers, Dr. Okuno, to spend time at our laboratory getting educated in what we do in knowledge systems. Okuno subsequently became a very good friend and colleague and now is a full professor at Kyoto University in Kyoto, Japan. And I see him regularly. That position of his in Kyoto University is quite a top position in Japan. It's a very highly regarded university and he has a full professorship there. So he's the young guy in these pictures. And then there's two Japanese people and they're two different pictures and I don't know which one is Kuroyanagi.

Box 67 - 2005

Folder 1: Instructional Guide--Knowledge Acquisition: The Key to Building Expert Systems--Donald A. Waterman and J. Ross Quinlan 1987

Don Waterman was my first Ph.D. student. Then he worked at Carnegie Mellon. Then he worked at the Rand Corporation. He unfortunately died an early death of lung cancer. And J. Ross Quinlan was also working at the Rand Corporation. He's an Australian professor who then went back to Australia and he was a full professor until he retired at Sydney University or maybe at the University of New South Wales or maybe both. So these are excellent teaching materials in that particular era, when everyone was desperate to learn about expert systems, namely the late eighties, early nineties.


Folder 2: Instructional Guide--Knowledge-Based Expert Systems: Planning and Implementation--Randall Davis 1987

Randy Davis is an AI Professor of EECS at MIT and a Professor in the MIT Artificial Intelligence Laboratory, one of my former students here at Stanford. Addison- Wesley planned and delivered a fairly substantial series of guides, workbooks and tutorial material to help people into this new field. And they recruited some of the best people in the field to write this material.

Box 26a - 2005

Folder 20: Fifth Generation Computer Systems: ICOT Journal (11 Volumes) 1985--1993

The Fifth Generation Institute was called ICOT, and I've mentioned that in other notes. The Institute for New Generation Computer Technology, ICOT. ICOT itself published its own journal and the one I have in hand is an example of it. This happens to be number 36 of 1993. Now 1993 is a peculiar date because it's after the tenth anniversary and this is supposed to be a ten year project. So the question is what's going on here? And the answer is the Japanese government gave it an extension. I don't remember whether it was a one year extension or two year extension but ICOT was still alive in 1993. And it shut down sometime soon thereafter, maybe '94. Finally there was a lot of activity in Japan on Japanese expert systems. Recall in one of my previous notes that I had chaired a panel, so called JTEC panel, Japan Technical Evaluation, I don't remember what the C stands for but it's a National Science Foundation sponsored event that sent me and Bob Engelmore and a panel of people over to Japan to evaluate their expert systems work. But meanwhile the Japanese were themselves were reporting on what they were doing. And in this case, the person reporting was Professor Mizoguchi of Osaka University. Mizoguchi regarded himself during that era - during the era of the sort of mid eighties through end of the nineties as Mr. Knowledge Based Systems, Mr. Knowledge Engineering of- of Japan, sort of self-appointed top guy. And he's very good. And recently Mizoguchi as things have shifted in the knowledge based systems world to more complex knowledge based systems and the development of indexing systems and inference called ontologies and other inference systems that use them, Mizoguchi now regards himself as Mr. Ontology or Mr. Ontological Engineer, I think he calls himself the Ontological Engineer of Japan. So this document here is a tutorial that Professor Mizoguchi gave at the World Congress on Expert Systems that was held in Orlando, Florida in December of 1991. In fact, I think this may be the first of these World Congress on Expert Systems hence the one where the Feigenbaum Medal was introduced and where I, myself, won the Feigenbaum Medal.

Box 1 - 1986

Folder 17: Tarjan

Concerning Stanford Training Grant Application for Medicine
The National Library of Medicine had gotten convinced that AI was going to become important to libraries in the future. Therefore, they decided to do some funding; at this time, they did funding not directly for the research, but for a training program that would generate more people to do Medical Informatics research. We said, okay, we're the best in this business, we ought to be getting a training grant.We put together a giant proposal. This is trying to make the best possible case for us getting this large training grant program -- the total for the entire proposed budget period was close to 1.4 million dollars. So it was large for its time. Therefore, we needed an extraordinary justification.


Folder 25: Allan Terry Thesis Proposal

Allan Terry Thesis Proposal: CRYSALIS
This is Allan Terry's thesis. Allan Terry was a special kind of graduate student in that he did all of his graduate work with me, but he was actually a graduate student of the University of California, Irvine. So this was a thesis submitted to Irvine. The thesis topic is the expert system CRYSALIS.


Folder 31: UNIVAC Nice, France Trip

UNIVAC Nice, France Trip
Donald Michie, who was an important pioneer in the history of AI in Britain, asking me to come to a British Journalists seminar at the Sperry UNIVAC Conference Center in Nice and give a talk, which I did.

Box 4 - 1986

Folder 1: Feigenbaum and Feldman, Computers and Thought, Vol. 2, source material

Julia Feldman and I were preparing a second version volume of Computers of Thought that would have come out in the late '60s, and I actually have a massive amount of Xerox's of the papers that would have gone into it over in the other office, which I'll give you. And this is early notes on what we're going to put in. Whose paper should we include. Long lists of possible papers that we should consider for including in Computers and Thought II. The book was never done.


Folder 2: Computers and Thought, Vol. 2, Working Paper Nos. 1-3

Computers and Thought, Vol. 2, Working Paper Nos. 1-3
This is a table of contents for a proposed book called Computers and Thought II done approximately ten years after Computers and Thought I was done. So this represents a first cut at what might be a final table of contents for that volume. Now there are many more papers listed here that could possibly go into Computers and Thought II. So the purpose was to leave big spaces after each one for a comment, and I sent this document out to quite a number of my friends in the field asking them to comment so that I could choose from among these papers which ones would be the winners so to speak for this book. Nothing ever came of Computers and Thought II. It was a great idea, it was all ready to go. Computers and Thought II aborted because I took responsibility for it and then never gave it a high enough priority to actually get it done. July 1972, right in the middle of this, was almost at the crux of turmoil in my personal life, having to do with my first marriage, and I just didn't focus enough attention on this thing to get it done. So it's an aborted volume. Would have been very valuable at the time, and I'm very sorry it didn't get done.


Folder 3: Rational Decisionmaking in Business Organizations by H. A. Simon and "On Cybernetics, Information Processing and Thinking" by M. E. Maron

On Cybernetics, Information Processing, and Thinking
Maron is now a Library Researcher at the University of California, Berkeley Campus Library -- School of Library Sciences. But at that time he was an Information Science Researcher at the Rand Corporation, and a friend of mine. And I kept this article because it discusses the work of a scientist named Ryle, a newer physiologist. But anyways, it's a question of how thinking takes place and how the brain works, and that was a particularly interesting series to me so I've kept it over the years.

Rational Decision Making in Business Organizations
This is Simon's Nobel Prize lecture. And the only reason for keeping this as opposed to getting it out of the Journal where it appeared -- American Economic Review -- is that Herb annotated this, "To Ed, with regards," which makes it valuable.


Folder 6: IBM-CDC lawsuit materials: "A Comparison of the Simplex and Semi-duplex Configurations for the IBM 360/67" by K. R. Graham; "Evaluating Time-Shared Computer Usage," Michael M. Gold, May 1966; and correspondence, 1966-67

IBM-CDC lawsuit materials: "A Comparison of the Simplex and Semi-duplex Configurations for the IBM 360/67" by K. R. Graham; "Evaluating Time-Shared Computer Usage," Michael M. Gold, May 1966; and correspondence, 1966-67
I gave testimony in CDC's lawsuit against IBM. This was different from the lawsuit of the Federal Government against IBM, which was eventually dropped. This is a private lawsuit for damages. And my testimony and all kinds of notes relating to it are bound up here.


Folder 8: GEO (French project, lecture notes) 1980-81

GEO is the name of a program that was done by myself and Penny Nii, while the two of us were on sabbatical at Schlumberger Engineering Laboratory outside of Paris in Clamart. And this folder contains notes from that entire visit, including some lecture notes for some general lectures I gave on AI and about Computer Science in universities in the United States and so on. There are some people mentioned here, Jacques Harry is our Geology Expert for this program. And J. C Picard was the Vice President for Engineering and head of that laboratory. We wrote a note here called "Any Midcourse Maneuvers," that is, should we change what we were doing. There's an example run, handouts for a talk that we gave, view graphs. An example run for a test well that was drilled in Syria.

GEO (French project, lecture notes) 1980-81 - Example Run and Diagram
There was a run here of the knowledge base for geology, December 1st, 1980, and the name associated with it is Friedland, but that may be just simply because Peter Friedland may have printed this out for us at Stanford and mailed it to us, because he certainly was not involved with it at all.


Folder 9: Vuegraphs for TEKNOWLEDGE 1982-83

Teknowledge - Executive Briefing on Knowledge Engineering
In October of 1981, the company Tech Knowledge, formed by several Stanford people including myself, started giving industrial seminars in Palo Alto. And we prepared view graphs for these industrial seminars, and these are the view graphs for the first of those industrial seminars. In fact, as far as I know, the first industrial seminar ever given about knowledge engineering in expert systems. So that's the very first executive briefing. Of course now (in 1987) there are dozens of those. They take place almost every week somewhere around the country.


Folder 12: SUMEX background material 1977

SUMEX background material 1977
This is a correspondence by Tom Rindfleisch, the Director of SUMEX, dated 3rd of February, 1977. That has to do with research highlights concerning the use of what's called the Stanford SUMEX AIM Resource in preparation, I suppose, for an annual report. But it shows you where we were in all these various projects around early 1977.


Folder 14: Art of artificial intelligence

The Art of Artificial Intelligence: Themes and Case Studies of Knowledge Engineering
The document here is the one that actually is the one I used while lecturing it. Shows what I actually chose to include in the public lecture and where the slides would come in and so on. The only other thing to say about this is that there is a portion of this called Epilogue, which is the portion that I never got around to writing because the paper itself got too long. And it says that I'm going to write a Part II of this.
That's why this is Roman Numeral I. And of course, as usual, despite the best intentions, etc., there never was a part II.


Folder 15: Slides, miscellaneous

MYCIN output augmented by rule-learning, from Ph.D. thesis work of Davis
Sixty-one will be a package of slides that have to do with a portion of Randy Davis's thesis, TEIRESIAS. This portion of TEIRESIAS is one called "Interactive Transfer of Expertise." And this is an actual consultation session with an expert in which the program is going through the knowledge base with the expert trying to find out the reason why the MYCIN program is giving the wrong , answer to something where the expert thinks the MYCIN program is giving the wrong answer. That exported program together as collaborators, worked through the knowledge-base looking for the part that would be incorrect. And then when that is identified the expert gives the correct piece of knowledge, and then TEIRESIAS criticizes the expert's knowledge.

Slides, miscellaneous - CONGEN
Julia Feldman and I were preparing a second version volume of Computers of Thought that would have come out in the late '60s, and I actually have a massive amount of Xerox's of the papers that would have gone into it over in the other office, which I'll give you. And this is early notes on what we're going to put in. Whose paper should we include. Long lists of possible papers that we should consider for including in Computers and Thought II. The book was never done.

Slides, miscellaneous - Data on the structure of Coriolin
Sixty-two is the activity of the DENDRAL program working on a problem taken out of the journal.

Slides, miscellaneous - Program Structure
The slides labelled sixty shows a system diagram for how the HASP program worked. HASP was a sonar surveillance program that was done by myself and Penny Nii in the period 1972 -1973 through 75.

Slides, miscellaneous - Rules
Sixty-three is a package of slides showing some examples of the various kinds of questions that MYCIN can answer, not only about its consultation with the user but also about its knowledge base.

Slides, miscellaneous - The Nature of an Application of Heuristic Programming Techniques
Sixty-four is two slides that are quite interesting. Way back after we had done DENDRAL, but before really there was anything else, even MYCIN, in talks I would give around the country people would ask me how did you pick DENDRAL? What's a good application? How do we know what's a good one? And so at one point, at Carnegie Mellon when I was on a visit, I formulated a list of criteria so to speak, and I called it the Nature of an Application of Heuristic Programming Techniques. Problem formulation, knowledge base, and problem difficulty were the main parts of it. And I apparently made up slides to talk about this. This has subsequently given rise to many articles that people now write about now that it's a big commercial thing. How do I find a good one in my company? How do I know what a good application of knowledge engineering is? This is the very first one, even before the term "knowledge engineering" was invented. It doesn't even use the word "knowledge engineering."


Folder 16: Slides, miscellaneous group 2

Misc. Slides - Incl. Dendral
As I would have many, many lectures at various times, I would produce these overhead projector materials, and each one of those relates to some different topic. So we're starting with viewgraph slides number one — ketoanderstone rules. What's significant about this it’s the information that ultimately went unto a paper representing knowledge induced by META-DENDRAL about fragmentations of chemical molecules. META-DENDRAL was another very well known program in AI, perhaps up through the 70s the best learning program that had ever been done. And this was producing new knowledge of chemistry. And this happens to be some ketoanderstone rules that were induced by META-DENDRAL, Rules M-4, M-5 and M-6. I don't want to go into the details about what the rules mean because you can always look up the META-DENDRAL paper to get that.Number two is MOLGEN's DNA structure editor. It shows some of the activity of the structure editor. Number three is a continuation of that. Number four shows some of the interaction capabilities of MYCIN, which — it says here — is a portion of a consultation session which demonstrates the "why" option that doctors could ask during a MYCIN interaction.Number five is an example out of Randy Davis's TEIRESIAS thesis on a thing called a META-RULE. A rule that indicates how other rules should be used. That was the first time that META-RULES had been used in AI. Number six gives essentially a gross flowchart of how META-DENDRAL works. Number seven shows a META-DENDRAL rule-generating tree. Eight gives the form of Mass Spectrometry rules in DENDRAL. Nine is a similar thing. Those were innovations at the time. That was the beginning of what's called rule-based knowledge, or AI programs. Ten is a very general flowchart of how the Heuristic DENDRAL program worked, and it shows that at the time we were doing this we were already separating the knowledge-base from the inference procedure,which is here, called Processes for Generation and Evaluation of Candidates. It's not called Inference Engine. But we were already separating knowledge base from inference processes. Eleven is another kind of flowchart showing the DENDRAL system and how it worked. Twelve shows the general problem that was on our mind. It doesn't say DENDRAL at all, it just says the data from a measurement processes input and output is one or more hypotheses constituting best explanations. It shows what was on our minds. Thirteen and fourteen show an audience what we mean by mass spectra. Fifteen shows how the META-DENDRAL program coordinates with the DENDRAL program. Sixteen and seventeen are examples of questions that MYCIN is able to answer. Eighteen and nineteen relate to an evaluation of MYCIN'S performance. Twenty shows a very simple example of an expert putting knowledge into MYCIN. Twenty-one discusses experiment planning MOLGEN. Twenty-two and twenty-three show results of DENDRAL solving problems on ketones and thioethers. Now these are different compounds that were given to DENDRAL. The important thing in these charts is to indicate how much selectivity you have in DENDRAL moving away from the very large number of group forms alternatives into the smaller number of plausible alternatives. Now Twenty-four shows an example of the operation of the Meta-DENDRAL program INTSUM on a particular structure. Twenty-five relates to the expert system CRYSALIS, and shows the different so-called "blackboards." These are three different blackboards that relate to inferring from the crystallographic data and the other chemical data what the structure of the protein in the crystal was. This Twenty-six is a very gross rendition of the entire Meta-DENDRAL program. Twenty-seven is a piece of sample dialogue for MYCIN. Twenty-eight is just a flowchart of META-DENDRAL. Twenty-nine and thirty relate to the structure generator of DENDRAL in the mid 1970s. It had been much modified. Thirty-one is an example of rule generation in META-DENDRAL. Thirty-two are two rules for MYCIN. Thirty-three is a "how" explanation for MYCIN. How did you arrive at a conclusion. Thirty-four shows — we're back to CRYSALIS now — actually how those multiple blackboards in CRYSALIS got used. Thirty-five and thirty-six are textual and graphical illustrations of how RULE-GEN worked. Thirty-six shows that one has to stop somewhere short of complete enumeration of molecular structure in order to get interesting rules for DENDRAL to use. Thirty-seven gives in words what the different pieces of RULE-MOD part of META-DENDRAL was all about. Thirty-eight shows more about RULE-MOD and how it actually works. Thirty-nine tries to explain a piece of chemistry to the audience, esters and ketones. Forty and forty-one are more results from the original DENDRAL program. Ethers, alcohols, and examines. Forty-two tries to give an illustration of how MYCIN works. Forty-three shows the beginning of a consultation. Forty-four shows a rule and how it's represented inside the machine. Forty-five, we're back to CRYSALIS and how it uses the multiple blackboards. Forty-six, forty- seven, forty-eight, all relate to that MYCIN consultation that started a few viewgraphs ago. These are mixed up. Forty-nine gives various examples of how learning might take place. This is a rather far reaching slide because we're still nowhere near being able to do any of those or do that very much. Fifty and fifty-one summarize some of those results that I mentioned about before on DENDRAL's work, on ketones, amines and estrogens. Now we're into what? Fifty-two is describing to people how the SUMEX system organization and networking is arranged. Fifty-three shows how MYCIN can answer questions about its consultation. Fifty-four, fifty-five are the same. Fifty-six, fifty-seven are more sample dialogue for MYCIN. Fifty-eight is simply a batch of these slides that have to do with sample dialogue for MYCIN.


Folder 17: All-day seminar on expert systems developed for TEKNOWLEDGE

All-day seminar on expert systems developed for TEKNOWLEDGE - Slides
This would be a sort of Feigenbaum seminar, Expert Systems and Knowledge Engineering. And the seminar director was John Rosati who was, I believe, the guy who just founded Continuing Education Institute and was running it as a side business to his main job which was at TRW in Los Angeles. And he obviously talked me into doing this thing in conjunction with TechKnowledge.


Folder 18: HPP material

HPP material- Slides
This packet of slides labeled "HPP Material" just talks about what is the HPP as a laboratory. And it must be relatively recent because the faculty members are listed as including not only me and Buchanan, but Genesereth, Lenat and Shortliffe, so this would have been in the '80s. We already had our VAX Computer and the DEC 2060. So we're talking about somewhere in the '80s. Talks about the blackboard framework in here, and it also talks about some HPP, what I call mega-trends, where we're really going to go over the long run. Would be interesting for the people who see this later to find out whether we went there.


Folder 22: Puff' slide (medical automated instrument)

Puff' slide
This slide labelled seventy-two shows graphically what happens when a patient does a pulmonary breathing test, which is interpreted by the PUFF system.


Folder 25: CLOT 1980

CLOT 1980 - Possible developments of the AI/COAG system which could draw more heavily on AGE facilities.
This is documentation from the University of Missouri on a system called CLOT, an expert system relating to blood clotting. The reason I have this is that we got involved at a very early stage in helping Don Lindberg at the University of Missouri formulate his application to coagulation.Some additional notes about Lindberg is that he is now Director of the National Library of Medicine, and the person Ueno who was a post-doc working for Lindberg and whose notes appear here, Ueno is back in Japan as a professor in one of the universities there and is one of the best known AI specialists in Japan. It all got started here with this project.


Folder 27: MOLGEN (Friedland's work)

MOLGEN talk: TOWARD A THEORY OF HUMAN PERFORMANCE IN EXPERIMENT DESIGN
There were two theses on MOLGEN. One was the thesis of Friedland, and one was the thesis of Stefik. This file folder is labeled "Friedland's Work", and these are view graphs relating to Friedland's work on experiment planning of molecular biology experiments.


Folder 36: Miscellaneous materials--DARPA, AI, etc., 1973-74

There's a thing here dated November 9th, 73, Plan for Finishing AI Report, What's the Minimum We Can Do, What Additional Things Should Be Done. There are various drafts of this report as it was emerging, commented on by different people. There's interaction with a man named Steve Crocker who was a program manager for DARPA for Artificial Intelligence at the time to various people. And a letter to me from Saul Amarel at Rutgers commenting on all of this. And Saul ironically, is now the Director of Information Processing Techniques at DARPA, right at this very moment, 1987. Historically I think that's very important.

Artificial Intelligence Research: What is it? What has it achieved? Where is it going?
It's fascinating that since this is dated '73, and here's a paragraph that talks about the "'what' to 'how' spectrum." And what's fascinating about this is this same text shows up in a critical part of the most important survey paper I ever wrote which was my invited speech at the 1977 IJCAI -- and this is '73. And yet this very same text shows up, so I must have gone back to this thing and lifted it and put it in there because I liked so much what I wrote there. And then I just try to summarize the whole field as I see it. What the goals and subgoals of the field are. Many things in this folder that I'm surprised to see here this early. DARPA did occasionally stop programs, and there was the possibility of that happening with this program. Lukasik I don't think was the kind of person who would have stopped this program. At a meeting of principal investigators in 1972, Lukasik drew a graph of the normal DARPA program, a five-year program which starts out at a certain high level of funding and decreases gradually over time until at the end of five years there's no funding. And then he drew a horizontal line on that graph and said, "This is AI. AI is my long-term investment. It doesn't decrease." So I don't think Lukasik actually would have cut the funding off. But Lukasik's successor, Heilmeier, actually tried to do it. By this time, however, George Heilmeier is one of the strongest supporters of AI in the world and was on the cover of High Technology Magazine in May in the lead article entitled Artificial Intelligence: The Heavies Get In, which was his Texas Instruments effort. George is also on the Board of Directors of the Center for Integrated Systems. But he's very strong on AI now. But he was very skeptical during that period. I think the weight of all the argumentation that he received changed his mind. He's a very rational, clear-thinking, scientific and technical person. And I think by pushing on these people, myself and Newell and all the other people in the field so heavily for so long, and getting such good responses from us about the importance of all of this that it just got to him that this was real important. It was very soon after Heilmeier left his job at DARPA to take on the job of the Chief Scientific Officer of Texas Instruments. It was actually a matter of just a few weeks or maybe a few months that George called up me and John McCarthy asking both of us, "How do I start an AI lab here at Texas Instruments?" So in spite of all the hassle, George became a kind of born-again Al-lover. He loves it. I mean we've been responsible for this massive effort by Texas Instruments. And I love it. So I don't think Lukasik would have have cut off the funding of the field. This apparently was being written for Lukasik.

Extended comments on a DARPA effort by Steve Crocker to review the AI research it was sponsoring, especially in the AI Labs
A man named Steve Crocker, who is now the Head of Artificial Intelligence, and maybe some other things at Aerospace Corporation, was a young man serving at the ARPA Office in Washington (the Advanced Research Projects Agency). In this period in which these messages are dated, mid 1973, ARPA was conducting a reexamination of the importance of research in Artificial Intelligence, which it had been funding for a long time, and was always subject to periodic re-review. The field was constantly under pressure to justify itself. This may have coincided with the rise to the Directorship of Dr. George Heilmeier, now Chief Technical Officer of Texas Instruments. Heilmeier was skeptical about Artificial Intelligence, so this was a set of questions by Crocker to the wheels of the field to give him ammunition for a document which would be, why AI, why is it important, what has it accomplished, where is it going? It's a planning document. And a justification document. For example, right in the middle of this batch of messages is a message from AI Newell at Carnegie Mellon to Crocker. Re: questions on AI field. And Crocker's questions are listed here, and Newell's answers are sketched out. Ditto for Peter Hart who's now Vice President of Syntelligence. And several other people in the field are in this package of messages. This then is two different copies of an emerging document of that sort to which I contributed. And that's why I have them.

Box 5 - 1986

Folder 1: EPAM Notes, EPAM III1963-64

EPAM Notes, EPAM III1963-64
This file folder contains the EPAM-3 program in a listing dated July 20th, 1964. This listing with a tremendous amount of hand annotation on it involving debugging of the EPAM code is annotated and de-bugged by Herb Simon. All of these writings that you see on here are Herb Simon's writings, and the note here saying, "Lee, I suggest you try it with changes, etc., etc," that is a note from Herb Simon to Lee Greg; a psychologist at CMU who was working with us on the psychological, theoretic and experimental side of EPAM-3. Lee Greg-died of cancer a few years later. Anyway, this is a Herb Simon listing done at Carnegie Mellon.

EPAM Notes, EPAM III1963-64
In addition to that there is a Xerox sheet here of an actual experiment associated with these runs of the EPAM listing here on paired associate learning for a particular list of stimulus response items that we had been doing called The KAGLUK - after the name of the first pair of syllables. And there is a tree shown on the second page of this which is my own understanding at that time of how EPAM went through and learned this particular list of syllables. That is the actual tree that it built in doing this. A tree is otherwise known as the discrimination net in EPAM.


Folder 2: Computer printouts - more EPAM runs, annotated "trouble with Gordon Bower's data"1963

Computer printouts - more EPAM runs, annotated "trouble with Gordon Bower's data"1963
Here's a listing entitled "Trouble with Gordon Bower's Data." Gordon Bower is a psychologist at Stanford, and we were obviously trying to predict some of his data. And a partial listing followed by a run of the program, and it was in some trouble.


Folder 5: Computer printout - EPAM, Dan Bobrow Aug. 24, 1962

Computer printout - EPAM, Dan Bobrow Aug. 24, 1962
EPAM-2 was the predecessor of EPAM-3.
EPAM-2 is essentially the EPAM that was done for my Ph.D. thesis. And so we were still using a variation of it in 1962. This listing here has the name Dan Bobrow, who is now a very well known AI researcher at Xerox Palo Alto Research Center, in 1962 when this is dated, was still a graduate student at MIT. And I don't know how he got a hold of this listing or why a listing of mine is under the name name of Bobrow, but it was probably because he was at Rand for a summer institute or something.


Folder 7: The CAT run, EPAM July 24, 1962

The CAT run - Handwritten Notes
We did a particular experiment with a pattern recognition problem that Herb Simon and I had been working on and suggested by Oliver Selfridge. This was a problem called The CAT.The CAT was a recognition problem in which both the "H" and the "A" of "CAT" were degraded so that the "H" had its side slanted to look somewhat like an "A", and the "A" had its top cut off to look somewhat like an "H", and it was a question of how the program could learn to discriminate among those. Would it learn successfully?


Folder 8: POSTMEPAMB-II-2 printouts to Zvegintzov, March 3-April 17, 1963

Handwritten Notes
Now this is a variation of EPAM to handle some experiment of Leo Postman, a psychologist at Berkeley, but I don't remember the details of that. I was associated with the Center for Human Learning in Berkeley, and Postman was director, and a very well known psychologist of the day. And the experiment is a paired associate experiment of some sort.

POSTMEPAMB-II-2 printouts to Zvegintzov, March 3-April 17, 1963
Anytime the name Zvegintzov comes up, that's a graduate student of mine at Berkeley. And these are printouts that he made for me.


Folder 9: EPAM2, Assembly listing and successful reading test 11-17-60. Copy of nearly successful runs of reading test, 11-11-601960

EPAM2, Assembly listing and successful reading test 11-17-60. Copy of nearly successful runs of reading test, 11-11-601960
There's a listing here called "Assembly Listing and Successful Reading Test" dated November 17th, 1960. And my guess is that the so-called reading test refers to work that was later published by Simon and myself called "Performance of a Reading Task by an Elementary Perceiver Memorizer in Behavior Science" And then this is an associated file called "Copy of Nearly Successful Run of Reading Task", and it lists what the tasks were and that there were some difficulties on the second task.


Folder 11: EPAM Paired Association Runs Sept.-Oct. 1960

EPAM Paired Association Runs Sept.-Oct. 1960
This is some EPAM experiment dealing with paired associates. Notice that September 1960 is after I returned from my Fulbright Scholarship in England. I was in England from the day after I took my Ph.D. in 1959 when I left for England until the summer of 1960. So this was obviously when I got back, because I started work again on EPAM when I got back -- because there was no way I could do EPAM work in England at the time.


Folder 12: EPAM II new listing, made Aug. 18 1964

EPAM II new listing, made Aug. 18 1964
The rest of this box (Box 5) consists of a large number of my own EPAM listings. Incidentally, on these EPAM listings, both on Simon's and my own, all of the programs are written in a now arcane but historic programming language known as IPL-V. So anyone looking at this in the future will have to look up these what's called the IPL-V manual edited by Atyan Newell to find out what these symbols mean.

Box 6 - 1986

Folder 2: W. Wickelgren "The Planning Method in Concept Attainment" Berkeley Thesis, 1962

The Planning Method in Concept Attainment
Wayne is a faculty member at the University of Oregon, or was last time I checked with him. This is one of the earliest theses in what is now called Cognitive Science, which is the Psychology end of Artificial Intelligence.


Folder 3: Lectures on AI, 1963-65 and Jan. 1976

Now here's a packet of material called Lecture Notes, Heuristic Problem Solving. These lectures are from probably the late Berkley days of mine, and early Stanford days which put them around, say, in the '64-'65 era. And I'm talking about here, all the concepts that we were thinking about in the first decade of AI work. This is my postgraduate school early Assistant Professorship, and I was giving lectures on AI, and almost no one in the world knew about AI, and I was talking about the events of the time, so to speak. For example, Element "C" in this lecture is "Trust Investment Process: Decision-Making Under Uncertainty, the work of Jeffrey Clarkson," who was a graduate student colleague of mine at Carnegie Tech. And nobody talks about Clarkson's work anymore. That was just at the time. Here's something about Saint, Jim Slagle's work at MIT about circa 1960 - '61, that would have been prominent at the time. Simon's Heuristic Compiler was '62. MH-1 was the earliest AI shot at robotics, Mechanical Hand, at MIT. So I would date these lectures somewhere in the 1964 - '65 period. This is a similar one. This is called Global Outline for a lecture on Heuristic Problem Solving, and it looks like it accompanies the detailed outline. They look the same.
And then these are pieces that I would want to use in the lecture, maybe even make view graphs of. And I take them from other places. And these happen to be taken from the book Computers and Thought. These are pages actually, they don't even look like Xeroxes. They look like I stole pages from a page-proof copy of Computers and Thought, because they look actually printed. And then I would quote from them.


Folder 4: ISI-EAF's report SAFE 1976

ISI-EAF's report SAFE 1976
SAFE was a project conducted by Bob Bolser at ISI (Information Science Institute) in Southern California for DARPA. SAFE was in deep trouble with the office as regards the funding that it might get for the future, because apparently Bolser couldn't explain well to DARPA the rationale for what he was doing. So Keith Uncapher's strategy for fixing this was to get Ed Feigenbaum to come down as an external consultant, discuss with Bolser what he's doing, do a one-man peer review of Bolser's work, and write this up for DARPA. And if it said good things then DARPA would be inclined to go ahead and fund it based on Feigenbaum's clout. And if it said bad things, well, maybe Keith Uncapher should know about that, and stop the project so he doesn't get a black eye. And so this was the report, what Ed Feigenbaum found out by a two-day study of SAFE in 1976.
It was pretty good.


Folder 6: ARPA - AI climatology proposal 1971

ARPA - AI climatology proposal 1971
Some work that we specked out and never did. It says "Dear Larry", and Larry refers to Dr. Larry Roberts who was Director of Information Processing Techniques at ARPA at that time, approximately 1971. Larry Roberts is now (1987) the President of the part of DHL that is doing the ZAP Mail competitor. He's president of that up in San Mateo. And we were speaking out an AI application to climate predication at the Rand Corporation with Rand climatologists.
It was a very hard problem. And in the end - climatology is what it was called here - we decided we weren't going to do that. Bob Engelmore was key to helping me get that one specked out, because Bob was a physicist who had recently joined the project. The notes on this document in black are Bob Engelmore's handwriting, and the notes in red are my handwriting. And it's a draft spec for Larry Roberts on how we would do this.


Folder 8: DARPA report on AI (outlines, draft, email) 1973

DARPA report on AI (outlines, draft, email) 1973
Message from Licklider. Licklider served two terms as a Director of the Information Processing Techniques Office of DARPA. And this about the same time, and I don't remember whether Licklider was back in office for his second term at that point. He started that Information Processing Techniques Office back in 1964 I believe, and then came back in the '70s for a second run at it, and he may have been Crocker's boss at the time. This is more from Licklider about the politics of this situation. All these messages from Licklider. He's sending it to me I guess saying, "Dislike to dun you, but dislike even more to keep Lukasik waiting when he expects something. Hope you can finish up AI report soon." So apparently I was procrastinating.
So Lukasik was the Director of DARPA. Licklider, in his second run at that job, was the Director of Information Processing Techniques Office, and Crocker was a Program Manager. And we were all people who were receiving contract funds. So this is more on the politics.


Folder 9: Reports on AI for DARPA and COSER

Reports on AI for DARPA and COSER
COSER was a panel convened by the National Science Foundation to be a report on Computer Science. "What is computer Science? Where is it going?" And it actually resulted in a book and its subtitle is "A Report of the COSER's Panel" or something like that. A well-known book in the Computer Science field trying to define the field for the world. And this is a message from AI Newell to me and MM at MIT-AI, which is Marvin Mintzky. Nielsen at SRI-AI. Neifeett's now Chairman of our department. Raj Reddy at Carnegie Mellon and Terry Winograd here. And this is an excerpt from the CMU proposal to ARPA which he claims to be an interesting thing to put into the COSER's AI panel, the sub- panel of this COSER's group dealing with AI. And so this is Newell's view in April of 1975 of what the field is all about, where it's going.


Folder 10: Roadmap--printout of file

Roadmap was another project. It's still yet another one of these, "What is AI and Where are We Going?" For example - and this 1975. Here's a message from Keith Uncapher, Director of Information Sciences Institute at USC. "Subject: In Defense of AI, to myself and Licklider." And it's some ammunition from Keith Uncapher as to how we should characterize the field. Here's a message from Licklider to Nieteen and myself. "George Heilmeier wants me to produce a roadmap for the IPTO - Information Processing Techniques Office Intelligent Systems Program, actually for the union of artificial intelligence and knowledge-based computer system applications. Yesterday I received word that the roadmap should be ready for discussion on April 19th, and that creates some schedule pressure." And he describes what the roadmap should be.


Folder 11: Roadmap--printout of file #2 Fall 1975

Roadmap was another project. It's still yet another one of these, "What is AI and Where are We Going?" For example - and this 1975. Here's a message from Keith Uncapher, Director of Information Sciences Institute at USC. "Subject: In Defense of AI, to myself and Licklider." And it's some ammunition from Keith Uncapher as to how we should characterize the field. Here's a message from Licklider to Nieteen and myself. "George Heilmeier wants me to produce a roadmap for the IPTO (Information Processing Techniques Office Intelligent Systems Program), actually for the union of artificial intelligence and knowledge-based computer system applications. Yesterday I received word that the roadmap should be ready for discussion on April 19th, and that creates some schedule pressure." And he describes what the roadmap should be.


Folder 13: Smith, Reid G., "An Introduction to Artificial Intelligence, DREA Technical Note SP/75/2, June 1975

., "An Introduction to Artificial Intelligence, DREA Technical Note SP/75/2, June 1975
This document is by a student of mine, Reid Smith, R. G. Smith. Reid came to Stanford, he was a EE student, Ph.D. student. He now works for Schlumberger Doll Research Laboratory. He came to us from the Defense Research Establishment Atlantic of the Canadian Government, DREA, where he had been working on problems of sonar signal identification using computers. And he came to Stanford. I had been working on that problem about the same time. Reed had gotten interested in the field of Artificial Intelligence and had written this document for his own people inside.
Copy 16 of 30 for the Canadian Defense Research Establishment Atlantic in Nova Scotia. And so this is in 1975 what Reed thought the field of Artificial Intelligence was all about -- his summary for in-house people.


Folder 17: Remembering, Learning and Forgetting: transcript pages 1-112

Remembering, Learning and Forgetting: transcript pages 1-112
There was a series of conferences held at Princeton sponsored by The New York Academy of Science under the title "Remembering Learning and Forgetting." The organizer of the conference was Professor Karl Pribram, then of the Psychiatry Department at Stanford, now of the Psyche. Department. Pribram's office is next door. Pribram was fascinated by the work that I had been doing on Models of Human Memory, the EPAM work. And he invited me to attend these conferences on Remembering Learning and Forgetting at Princeton. I was a young scientist in the field. However, I was at Stanford by that time. And I gave one of the long talks on EPAM, and my talk was followed by workshop-type discussion with lots of people in the field. Here's Pribram, saying "I am Carl Pribram, and I began my career as a neurosurgeon."

Box 7 - 1986

Folder 2: Computers and Thought - Polish Ed.

Computers and Thought - Polish Ed.
This is the Polish edition of Computers and Thought, for which royalties were never paid.


Folder 15: PUB - The Document Compiler by Larry Tesler, 1973

PUB The Document Compiler
We had a document preparation system which is the one that was in broad view before SCRIBE or TEX was developed. It was called PUB, the first real document programming system ever. So everyone used PUB. PUB was it for ten years. And this was my PUB manual. So it was an important system in the history of word processing.

Box 8 - 1986

Folder 15: Speech Understanding Systems, 1979

Review of the ARPA Speech Understanding Project
This is a document which was produced by the people engaged in the DARPA sponsored national Speech Understanding Project trying to get a follow-on to that. They had done five years and they wanted another five years, and they wrote this document, and this was the time of Dr. Heilmeier's skepticism about AI, and he said, "Well, where have you gotten in five years? No, I don't think that's enough. No you can't have the money." So this lead to the termination of the national Speech Understanding Project.


Folder 25: AI Handbook, Miscellaneous

AI Handbook, Miscellaneous
And this is an outline done by Avlon Barr. It's a handout for a meeting we were having. We had these continuous meetings with the group that was trying to formulate this idea. Barr was a key player and eventually became the first-named editor on Volumes I and II. You see the date on that is May something or other.

Box 12 - 1986

Folder 39: Mathematical Social Science Board (MSSB)--1966-78

There was an organization called the Mathematical Social Science Board, MSSB, and I was on it. I think it was supported by the Social Science Research Council, or spun out of the Social Science Research Conference. And for a period of time I was on it. This was administered by the Center for Advanced Study in the Behavioral Sciences up here on the hill. The list of people on this Mathematical Social Science Board is given here, and it includes me. And this was a proposal to the National Science Foundation to continue granting money to keep this board in existence. Allen Newell was on this board, Simon was on this board, Susi Van Peters was there, and many other big names.

Box 14 - 1986

Folder 26: Flavell, 1979. Includes: "From Social Interaction to Higher Psychological Processes" by J.V. Wertsch. 1978. "Metacognition and Cognitive Monitoring" by J.H. Flavell ¯Materials for Psych 246. "Fragments of a Theory of Human Plausible Reasoning" A. Collins. Notes from Barr and Flavell on artificial meta-cognition. HPP-79-11 "Transfer of Expertise: A theme for AI research" A. Barr, J. Bennett, and W. Clancey. HPP-79-6 "Meta-knowledge and Cognition" A. Barr. Monograph, Kreutzer et al., "IV. Conclusions." "Realizing That You Don't Understand: A Preliminary Investigation" Ellen Markman. "Skills, Plans and Self-Regulation" A. Brown and J. DeLoache. Notes on Markman, April 20, 1979. "Teiresias: introduction to knowledge-based programs" from The AI Handbook.

Box 15 - 1986

Folder 14: Proposal: MOLGEN - A Computer Science Application to MolecularGenetics (NSF Grant MCS 76-11649)
Folder 21: Xerox Contract. Contract for Feigenbaum to work as consultant for Xerox, 7/80.

Xerox Contract. Contract for Feigenbaum to work as consultant for Xerox, 7/80 - Letter From M. Franklin Squires to Edward Feigenbaum
I think this is a personal consulting contract -- a no-cost contract just to allow me to get into Xerox easily without elaborate procedures.


Folder 22: VM Project: Ventilator Manager notes, reprint "Computer-based Medical Decision Making: From MYCIN to VM", L. Fagan, E. Shortliffe, and B. Buchanan, 1980.

VM System status and Futures, John Kunz Thesis Alternatives and "Forming Theraputic Plans in VM/ONCOCIN by Larry Fagan
The VM Project stands for Ventilatory Manager. That was one of the HPP projects. Myself and Larry Fagan who actually did the work. This is dated 1980, and it's a write up of the Ventilator Manager, and a proposal by a student named John Kunz on things he might do for his thesis relating to VM. And there's also a reprinted article on VM from MYCIN to VM.


Folder 27: ONR Panel: List of Names for ONR/NRC Panel on Applied Mathematics

ONR Panel: List of Names for ONR/NRC Panel on Applied Mathematics
This sheet is the list of people who were serving on an ONR sponsored panel called "Panel on Applied Mathematics Research Alternatives for the Navy". And I was on the panel, and so was Jerry Lieberman. Jerry was the one who really got me on that panel. Raj Ready at Carnegie Mellon was on that panel also.


Folder 30: NSF Computer Engineering Division: Copy of report "Grants Supported by the Computer Engineering Program in Fiscal Year 1981 and 1982" and map with general distribution of awards.

NSF Computer Engineering Division - Map with general distribution of awards
This map is interesting from my point only for one reason. This is sort of definite proof of the fact that NSF definitely worries about the geographical distribution of awards rather than to give merit alone. For their Computer Science and Technology Advisory Board of which I was on, they would present the list of awards to us in alphabetical order by states. You know, Arizona, Alabama, that sort of thing. Here's a piece of evidence of that. And this happens to be the awards themselves for that year in case you care. MOLGEN is one of these. That's why I was sent this, I believe. But these are the awards for those years.


Folder 32: HPP Miscellaneous, three folders and loose notes

November 74, a file labeled "An Intelligent Terminal Proposal". DARPA at the time was beginning a program called Intelligent Agents, and there was some discussion about whether we should propose research activity into that program. This is an outline and some text material for a proposal along those lines, and I can't remember whether we ever did decide to propose this sort of thing or not. But this is material from that event.
This would pertain to my capacity as Research Director of the Heuristic Programming Project.

HPP Miscellaneous, three folders and loose notes: Principal Investigators - Information: copy of article on accounting problems on grants between universities and fed, and office memo reviewing federal export regulations. Intelligent Terminal Proposal: notes on intelligent couplers and bulletin board, 1974. Terminal Theft: correspondence, invoices and purchase requisitions re: theft of two terminals 6/75. Loose Papers: memo on space for HPP personnel (7/16/76) and memo re: deficit on NIH grant. (7/19/76).
Myself sending around an article from Science in '82 talking about Federal auditors harassing principal investigators for telling my people to be careful. And this is Gerry Lieberman's memo. Again, similarly, we have to be careful about these new government regulations concerning export control. The government trying to define certain scientific things as being technical data which is subject to export control.


Folder 33: IBM Proposal: Attempts to determine the User's conceptualizations of task domains within Intelligent Tutoring Systemsÿ . 1/25/83, Buchanan - Principal investigator

IBM Proposal: Attempts to determine the User's conceptualizations of task domains within Intelligent Tutoring Systemsÿ . 1/25/83, Buchanan - Principal investigator
This material here is a proposal that Buchanan submitted to IBM in support of the work of a person he brought here named Derek Sleeman, a British guy from Leeds University, who's now returned to Aberdeen University, to support work on Artificial Intelligence applied to tutoring and education.


Folder 51: ARPA P.I. Meeting, L.A., Feb 6-8, 1974

ARPA P.I. Meeting, L.A., Feb 6-8, 1974 - Memorandum for Principal Investigators
This is the person who was in the ARPA Director of Information Processing Techniques job in the period surrounding the Lukasik-Heilmeier Directorship here, the guy one level down in the Computer Research Grants Office was J.C.R. Licklider. This is 1974. Licklider is calling together a meeting of the major principal investigators of his projects, Feigenbaum, McCarthy, Marvin Minsky of MIT, Allan Newell of Carnegie Mellon, Keith Uncapher of Information Science Institute, and others.


Folder 58: ARPA Contractors Meeting: San Diego, March 12-14, 1975.

ARPA Contractors Meeting: San Diego, March 12-14, 1975. - 1974 ARPA Project Summary - Lawrence Livermore Laboratory
The people who sponsored us at DARPA all these years held annual contractors meetings at that time. "Contractors" meaning us and all of our friends who accepted ARPA money. We would all go down there for a two or three-day meeting once a year. We'd bring write-ups of what we had done in the recent year so that people could know each other's results. But we wouldn't get up and give a lecture from that paper. That was the preliminary material. You know, like if it's in here we don't have to talk about it. And then we would spend the time talking about the really interesting stuff. "How do you put this all together? Where is it going next? What is this thing called ARPANET? Why should we pay any attention to it?" All you're trying to do is drain off our programmer's time. Why don't you let us do important things instead of this silly network stuff? You know, how wrong can you be? But that's what this is all about. This is the material that we would be given when we got down there, and we'd read some other time. But this is extraordinarily valuable material. This is really it. This is not just Stanford stuff; this is the great stuff being done around the country as it's unfolding and as the people are talking about it, while they're doing it, to their colleagues. And this is not published papers.


Folder 59: ARPA Contractors Meeting: Washington D.C., Dec 1971

ARPA Contractors Meeting: Washington D.C., Dec 1971 - 1971 ARPA Project Summary - The Information System Theory Project, Applied Data Research, Inc.
The people who sponsored us at DARPA all these years held annual contractors meetings at that time. "Contractors" meaning us and all of our friends who accepted ARPA money. We would all go down there for a two or three-day meeting once a year. We'd bring write-ups of what we had done in the recent year so that people could know each other's results. But we wouldn't get up and give a lecture from that paper. That was the preliminary material. You know, like if it's in here we don't have to talk about it. And then we would spend the time talking about the really interesting stuff. "How do you put this all together? Where is it going next? What is this thing called ARPANET? Why should we pay any attention to it?" All you're trying to do is drain off our programmer's time. Why don't you let us do important things instead of this silly network stuff? You know, how wrong can you be? But that's what this is all about. This is the material that we would be given when we got down there, and we'd read some other time. But this is extraordinarily valuable material. This is really it. This is not just Stanford stuff; this is the great stuff being done around the country as it's unfolding and as the people are talking about it, while they're doing it, to their colleagues. And this is not published papers.


Folder 60: ARPA EAF: Revised Proposal to Defense Advanced Research Projects Agency for the continuation of the Heuristic Programming Project and the VLSI Theory Project. May 1982

Letters from Edward A. Feigenbaum to Philip R. Surra
This file documents essentially the business part of dealing with the ARPA contracts. In particular, memos with a quaint recalcitrant and surly contract officer resident at Stanford named Philip Surra and there's a lot of things that Surra made us do, that we've jumped through hoops that we wouldn't ordinarily have to jump through hoops. He wasn't a Stanford person. He was working at the ONR Office on campus.

Box 16 - 1986

Folder 1: EPAM Vocabulary List/ EPAM Inverse List (8-22-61)

EMPTY FOLDER
EPAM vocabulary list and EPAM inverse list, and this was something which was very important in programming at the time with all these funny symbols that we were using -- non-mnemonic symbols. We had to know not only what each symbol meant in terms of our own understanding of this program - what the D's meant, what the E's meant, what the S's meant, what each nonsense syllable was, and so on, but we also had to have it printed in inverse order as well so that when we looked at something in the code we can then go back and find out what that really meant. So this was both a vocabulary list and an inverse sorted list to find things in the code. This is a very important working document. And there's a loose sheet which apparently has pulled out from here so the question is how to bind that in or hold it in, or it's going to fall out. (Note in 2017: The sheet has apparently indeed fallen out during the archiving process, which is why this folder is titled "EMPTY").


Folder 2: EPAMELTON runs (EPAM Hypothesis for Melton, Shepherd and Teghtsoonian running recognition experiments) Prop. of N. Zvegintzov and E. A. Feigenbaum

"EPAMELTON" runs (EPAM Hypothesis for Melton, Shepherd and Teghtsoonian running recognition experiments) Prop. of N. Zvegintzov and E. A. Feigenbaum
There was another psychologist in the Center for Human Learning named Arthur Melton. Melton and a pair of people named Shepherd and Teksunian from Bell Labs, Roger Shepherd now at Stanford, had done some experiments and published them. And this graduate student, Zvegintzov and I were trying to simulate the data from those experiments to try to explain those experiments using EPAM. And so this was called EPAMELTON. Nothing ever came of POST-MEPAM or EPAMELTON. That is, there's no published paper.


Folder 4: Max Allen - Theses work. Handwritten copy of program, printout of program, printout of EPAM run

Max Allen - Theses work. Handwritten copy of program, printout of program, printout of EPAM run. IPL-5 Programming
This is by a student named Max Allen who did a thesis on a variation of EPAM called FREEPAM - Free Recall EPAM, I believe. It was a Ph.D. thesis in the Psychology Department of Berkeley. And I think I just kept it as a historical document because it was a different kind of EPAM.


Folder 5: EPAM III--2 notes on changes to program, log book, summaries

This folder contains in it a Blue Line document "For Users of EPAM-3," written by this guy Zvegintzov, a graduate student. There's an EPAM log in here of the different experiments we were doing. EPAM's vocabulary. IPL's list of basic processes. And a working paper that got published called "Generalization of an Elementary Perceiving and Memorizing Machine." This paper probably preceded all of these notes and was used as a guide for us. And there's a lecture notes here dated April 13th, 1962. This is basically an experimental notebook. And there are at the back - some thoughts, handwritten items, handwritten procedures, how EPAM really worked, and some IPL5 coding sheets. My own coding sheets are back here, too, in the green. I apparently saved some of these initial programs. Not sure why, but these are the actual coding sheets and a lot of notes.

Feigenbaum's original programming (coding) sheets for EPAM III--ten sheets
Now this is the EPAM listing in IPL-5. IPL-5 is the precursor language to LISP. And this is the program that I wrote, EPAM-3. This listing was made at the Rand Corporation where I was doing some work, and it's a real historical artifact because probably Newell, Simon and myself are the only other people in the world who would still have copies of IPL-5 programs, what they looked like. This is pre-LISP. Underlying it it has everything in common, but at the language level it has nothing in common, so the syntax is completely different.

Box 17 - 1986

Folder 15: AI Lab Memos and Computer Science Reports July 1974

Recent Research in Artificial Intelligence, Heuristic Programming and Network Protocols
This is a DARPA Progress Report for the period July 73 to July '74, which accommodated not just HPP work listed as Feigenbaum and Lederberg work, but also McCarthy's AI Project, and Vint Cerf's Network Protocol Project. These were several ARPA projects which were aggregated into one annual report which was sent to ARPA, and then published as a CS report so that we could send it around. Les Earnest, who's listed here as the editor, is currently Associate Chairman of the Computer Science Department, he's also the founder of Imaging Corporation. After this but before he came back to being Associate Chairman - and this would be an extraordinary source for historical information -he was really the managing director under McCarthy. He kept things running.


Folder 17: AI Lab Memos and Computer Science Reports October 17, 1970

RESEARCH IN THE COMPUTER SCIENCE DEPARTMENT AND SELECTED OTHER RESEARCH IN COMPUTING AT STANFORD UNIVERSITY
This particular technical report, CS 178, would be of special interest because it's what you might call the research brochure of the department for that year -- 1970. This tells what everyone was doing in 1970 in our department. Everyone was asked to write up something. And then all the theses. This is similar to what in a company you'd call the annual report.


Folder 22: AI Lab Memos and Computer Science Reports December 24, 1965

A PROGRAMMING LANGUAGE FOR THE 360 COMPUTERS
Niklaus Wirth was a student of mine at Berkeley. He got his degree there, transitioned down here at the same time I did, was an Assistant Professor here, started a development in different programming languages of which this was one paper in that series, ended up being the language Pascal that everyone uses. His thesis advisor was Harry Husky, and I was his other thesis advisor. There were three of us. And then he came down here to teach at the start of the department. He's Swiss, and he went back right after he left here.

Box 18 - 1986

Folder 6: Lederberg Material on History of DENDRAL

Lederberg Material on History of DENDRAL
This contains a Lederberg Stanford memo. "In the beginning - DENDRAL-What Happened. Dear Ed: If it is not already too late and you still find any of the earliest pieces of paper,correspondence, proposal, draft, etc., it might date and document the beginning of the DENDRAL project. I will dig out what I can, too. When were you at the center What was I doing there?" What Lederberg means is that the first contact that he and I ever had was a meeting at the Center for Advanced Study in the Behavioral Sciences, up on the hill. I came down one Saturday with Julian Feldman, my coworker up there at Berkeley, to attend a monthly meeting that we were doing with Carl Preeber. Carl and I were friends at the time. He had invited me to a lot of conferences on Information Processing Theories of Memory. And Carl and I therefore decided we wanted to talk more about this on a routine basis. So I'd come down from Berkeley, or he'd come up to Berkeley. Well, it happened that one of these days Carl had a conference room reserved at the center and had invited Lederberg to that meeting. Now Why Carl decided to invite Josh, I don't know. Josh had some interest in computer models of thought because he had become friendly with McCarthy. And that's when I first met Lederberg. And although what we eventually worked on was not Computer Models of Memory. That's how the conversation got started. And Josh is asking me here "when were you at the center?" Who knows that was a random Saturday of some month. I would guess it was early 1964. Or - or maybe in '63. And what was I doing there? That's Lederberg. I don't know what Lederberg was doing there. Carl Pieber invited him up to this meeting. And then in '75 Lederberg is writing this note to Bob Merton, the Sociologist of Science, titled "Dear Bob: Memo: Tidbit of CASBS History." That's the Center for Advanced Study in Behavioral Sciences. "Ed Feigenbaum just reminded me what I totally forgot. That our DENDRAL was conceived at the center. He tells me that we met at the CASBS during the summer of 1964 when he was at UC Berkeley at a Saturday conference on mind modeling that Carl Preeber had organized. Ed joined the Stanford faculty January '65. By then we had already exchanged complementary ideas, blah, blah, blah, blah, blah."And then there's a Lederberg Chronology of the DENDRAL project dated 11-1- 77. Here's the Lederberg file number on the machine. SUMEX-Chron. TMP stands for Temp, so maybe he erased it.And this is a chapter in McCorduck's book. Lederberg wants the exact circumstances and dates. There's also a piece out of some document that Lederberg and I wrote, and it's an attempt to generalize what DENDRAL does in categories.


Folder 7: DART (IBM Sponsor)

MODEL-DRIVEN PARI
One of the things we proposed after MYCIN'S work on medical diagnosis was DART work on equipment failure diagnosis, called DART. I got Mike Genesereth involved with that after he came to Stanford. Kind of transferred from Mike. This is a DART Proposal to IBM in which I'm still the PI but Mike is involved. Here's some IBM correspondence about it, which is in fact the letter of contract believe it or not, because it's addressed to S.P.O. John Richie. This is really the letter of contract. And it has in it this diagram which would settle all the ownership questions as regard to who owned what in the system. That diagram was pulled out a hundred times during a subsequent crisis between Stanford and IBM called The Problem with the Umbrella Agreement. It was a big deal between IBM and Stanford. And we were outside of the problem because we had this diagram. This diagram clearly delineated what what wasn't so clear in other things. Anywhere where the dotted line is it says who owns what on what sign. IBM owns this, and Stanford owns this.


Folder 14: MSG Manual

MSG Manual
Here's a now obsolete mail system called MSG that was being used during the era when 10X was the operating system at Stanford. It's no longer used. There's no MSG, but this does give you an idea of what our system was like at the time.


Folder 16: KRL (Knowledge Representation Language)

KRL (Knowledge Representation Language)
I should say the precursor of the UNITS Package was KRL by Bobrow and Winograd at Xerox Palo Alto Research Center. This file is an overview of KRL, so it shows you that it was really on our mind at the time, because this folder is right next to the MOLGEN folder. So we were clearly drawing on that. And there's another paper that Xerox people were doing at the time on GUS. So what you would have to say is that we were influences on MOLGEN.


Folder 17: MOLGEN - Notes and Correspondence

MOLGEN - Notes and Correspondence
This represents a genuine fact-stuffed file of real scientific interactions relating to MOLGEN. This starts with some MOLGEN notes of November 17, 76. These notes written up by Nancy Martin who was a visiting scientist at the time, and it includes scientific working notes. People as they're exchanging information in electronic files.

Box 19 - 1986

Folder 2: Correspondence

Correspondence - Letter from A. V. Napalkov to Edward A. Feigenbaum
This is a letter dated June of 1975 from a man named Anatoli Napalkov and his address is stapled on the second page of the letter. He was what the Russians called a neurocyberneticist. Sort of neuro-computational models of thinking. I came in contact with him in 1960 in connection with my trip to the Soviet Union, saw him in '64. His wife, Nela Napalkov, is a biologist. She came to Berkeley while I was Berkeley. She spent a year there. And this is a letter from Napalkov in 75 about something or other. I don't know. This is probably the last letter I've ever gotten from Napalkov. And his wife says "I'm looking forward to seeing Nancy and you in Moscow again," except Nancy and I were already divorced, so you can see there's been no contact for a long time with these people.

Box 20 - 1986

Folder 38: ROGET

Memos from James Bennett: In-house Review Proposal, Sunday, November 29, 1981; ROGET: a knowledge-based System for the Design of Expert Systems, Wednesday, November 25, 1981
This memo talks about in-house review proposal, dated December 18th, 1981. And it talks about why we should in-house peer reviews of the quality of work going on. As the laboratory grew bigger and bigger, shouldn't we all be looking at each other's projects and making some kind of comments and critiques.


Folder 55: Snowbird 1980

Snowbird 1980
This thing here is a report from a subcommittee of the Computer Science Department Chairman. They meet every year or couple years in Snowbird, Utah in the summer. And a subcommittee was appointed and Denning, myself, Gilmore, Herg, Richie and Traub were appointed to write this document indicating that a funding crisis existed in computer science. A Discipline in Crisis, November 21st, 1980. This turned out to be a quite important report because it led to some action by the National Science Foundation Computer Science Advisory Committee to advise that there needed to be several more centers of excellence, and large amounts of money poured in to create those.

Snowbird 1980
This is one of the periodic kinds of things where computer scientists get together to figure out where their profession is going, and this is a panel shared by Professor Traub. The panel was at Snowbird, Utah, at the biannual meeting of chairmen in Snowbird, Utah. I was Chairman at the time, so I got it.


Folder 62: Stanford University Board of Trustees Meeting - March 7, 1983

Slides for Stanford University Board of Trustees Meeting - March 7, 1983
Here are transparencies made for a meeting of the Board of Trustees, March 7,1983, when they asked me to come and give a talk on what the laboratory does. So it's really a talk about the work of the laboratory and our problems. Who we are, what we do.

Box 21 - 1986

Folder 3: HPP-66-1 - Artificial Intelligence Project and Other Papers

SOME SEOUENCE EXTRAPOLATING PROGRAMS: A STUDY OF REPRESENTATION AND MODELING IN INQUIRING SYSTEMS
The thesis by Steffan Persson, dated September 26, 1966, AI Memo Number 44. That's a special case because Stefan Persson was not a Stanford student. He was a Berkeley student, hadn't finished up his work at Berkeley when I moved. And so I was supervising his thesis sort of long distance. But when it actually got printed up, it got printed up as a memo of us here, in my project.


Folder 5: HPP-64-1 DENDRAL-64-A System for Computer Construction, Enumeration and Notation of Organic Molecules as Three Structures and Cyclic Graphs

DENDRAL A SYSTEM FOR COMPUTER CONSTRUCTION, ENUMERATION AND NOTATION OF ORGANIC MOLECULES AS TREE STRUCTURES AND CYCLIC GRAPHS - Complete Chemical Graphs; Embedding rings in trees
There were two streams of work in DENDRAL. There was what you might call the pure computational stream, which had to do with systematic enumeration of chemical structures. And then there was the heuristic stream. And that's why the stuff that I was involved with we named Heuristic DENDRAL to distinguish it from the other stream which was DENDRAL-- which stood for Dendrite Algorithm, meaning it's a pure algorithm just exfoliating these chemical structures in a systematic way. Now Lederberg was the person solely concerned with pure DENDRAL. We were implementing versions of pure DENDRAL for him. But in terms of who wanted to that kind of work, Lederberg's driving love at the time was to do that. And there came a time when he wanted to write that all up, and he wanted to get it down on paper and he wrote a big series of articles, Part I through Part V, "DENDRAL, A System for Computer Construction Enumeration and Notation for Organic Molecules, as tree structures and super graphs". Doesn't say anything about AI.


Folder 11: ACM

Letter from David H. Brandin to George S. Malindzak about filing quarterly activity reports
It says here, CC: Ed Feigenbaum, Chairman of SIGBIO, A special interest group for biological computing of the ACM. Dave Brandin was the Chairman of these special interest groups, and he's talking about financial accountability and financial responsibility and a lot of stuff having to do with SIGBIO and financial reports.


Folder 12: Quo Vadimus: Computer Science in a Decade, June 1980; "Minutes of the Workshop on the Application of Artificial Intelligence to Electronic Warfare," 6 and 7 August 1975

MINUTES OF THE WORKSHOP ON THE APPLICATION OF ARTIFICIAL INTELLIGENCE TO ELECTRONIC WARFARE-6 and 7 August 1975
If you want to know more about this, Tom Garvey still works at SRI. People weren't talking much about application in 1975, so this is one of the earliest such things. And it summarizes the reason why application ought to be a credible thing.


Folder 13: Computing Reviews

Computing Reviews
This is a letter from a man who just took on the job of editor of Associate Editor of Computing Reviews. And he's not only sending me a form, but he's sending me the classification system for computing reviews. How they're classifying all the items in computing, which is good. I mean that's something important, to see how the field was classifying itself.


Folder 14: Australia Conference Transcripts

I've only been to Australia once so far in my life, and it was for this conference. At Canberra they were holding a conference on AI. Actually, they were quite interested in AI in Australia at the time.
And shortly thereafter the Lighthill Report was published in Britain. The Australians take seriously whatever was published in Britain, and that killed AI work in Australia for a decade. Now they're quite interested again and I'm going to be there next May to give a talk. There's a letter here from the CSIRO, the Government Bureau Division of Computing Research, and it asks me to do some work on these transcripts. It's dated 74, and that doesn't say when this thing was. It could be 72 or 73. And so here's my work on transcripts.


Folder 16: ARPA UPP Project (Basic Research Grant)

ARPA Proposal Outline
These are milestones. This is the ARPA basic research contract for the HPP, sometimes called the 6.1 contract. This is dated July 75, but what it gives is a list of dates related to milestones for things we have to accomplish in this project. This will give you a list of what we're actually doing at that time in the HPP for ARPA.


Folder 30: LISP, March 1972

LISP/360 REFERENCE MANUAL
The 360 means the IBM 360. LISP 360 was done by Stanford for the IBM 360. It was started while I was director of the Computer Center here at Stanford, and this is a March 1972 manual for it. It's part of the history of LISP.

Box 22 - 1986

Folder 1: ARPA Contractors Meeting January, 1973

ARPA Contractors Meeting January, 1973 - 1972 ARPA Project Summary - IMAGE PROCESSING RESEARCH AT THE UNIVERSITY OF SOUTHERN CALIFORNIA
The people who sponsored us at DARPA all these years held annual contractors meetings at that time. "Contractors" meaning us and all of our friends who accepted ARPA money. We would all go down there for a two or three-day meeting once a year. We'd bring write-ups of what we had done in the recent year so that people could know each other's results. But we wouldn't get up and give a lecture from that paper. That was the preliminary material. You know, like if it's in here we don't have to talk about it. And then we would spend the time talking about the really interesting stuff. "How do you put this all together? Where is it going next? What is this thing called ARPANET? Why should we pay any attention to it?" All you're trying to do is drain off our programmer's time. Why don't you let us do important things instead of this silly network stuff? You know, how wrong can you be? But that's what this is all about. This is the material that we would be given when we got down there, and we'd read some other time. But this is extraordinarily valuable material. This is really it. This is not just Stanford stuff; this is the great stuff being done around the country as it's unfolding and as the people are talking about it, while they're doing it, to their colleagues. And this is not published papers.


Folder 14: Callero, Monti, "An Adaptive Commmand and Control System Utilizing Heuristic Learning Processes," SAIL Memo AI-58

AN ADAPTIVE COMMAND AND CONTROL SYSTEM UTILIZING HEURISTIC LEARNING PROCESSES
One of the first students around here, Monti was an Air force officer who was here as a Ph.D. student, and that's quite rare to have that. He was in the Operations Research Department and got intrigued with AI and learning machines, and wanted to do an AI thesis out of that department. Now he was put in contact with me by Jerry Lieberman, and even though he was an OR student he did his thesis under me, and it was a learning program, an "adaptive command and control system". That's the clue that you're talking about learning because it's adaptive as well as saying "utilizing heuristic learning processes". Command and control is a military term.So what he was doing was a military missile targeting system, or attack system, or something like that, which was also very strange for a university. Universities don't sort of do that sort of thing. But that's what he knew best. And so he did his thesis on that. The overall objective of the defense system considered is the mod - in the model is to minimize total loss and target value -- that's a real military thing. Post-attack data file; as the attack develops - really an extraordinary thing. Look at these are missile assignments to targets.As the years went on, particularly after the Mansfield Amendment during the Vietnam War, the DARPA sponsors of all of this stuff for the ONR sponsors and whatever became more military mission oriented. And so if you look, for example, at the work we're doing in the HPP now (in 1986) on parallel computing for AI, has an application which is electronic intelligence, and another one is radar tracking of airplanes.


Folder 15: Mark Stefik Thesis

Mark Stefik Thesis
An important thesis that was done from the project was done by Mark Stefik, now the Director of Knowledge Engineering for Xerox, Palo Alto Research Center. Mark's thesis was on AI Applied to the Planning of Experiments in Molecular Biology. Helping Genetic Engineers Plan Recombinant DNA Gene Cloning Experiments. He would come to see me and he'd bring notes, and we'd talk about them,and I would file them away in a binder called Mark Stefik's Thesis. And this is the unfolding of Mark Stefik's thesis as seen by the supervisor, what he would bring to me, and I would write notes on it. It kind of gives a view of how a student works through the ideas in a thesis and what kind of feedback a student gets from the advisor.


Folder 16: ARPA Contractors Meeting - February, 1974

ARPA Contractors Meeting - February, 1974 - RISOS Project 1973 ARPA Project Summary
The people who sponsored us at DARPA all these years held annual contractors meetings at that time. "Contractors" meaning us and all of our friends who accepted ARPA money. We would all go down there for a two or three-day meeting once a year. We'd bring write-ups of what we had done in the recent year so that people could know each other's results. But we wouldn't get up and give a lecture from that paper. That was the preliminary material. You know, like if it's in here we don't have to talk about it. And then we would spend the time talking about the really interesting stuff. "How do you put this all together? Where is it going next? What is this thing called ARPANET? Why should we pay any attention to it?" All you're trying to do is drain off our programmer's time. Why don't you let us do important things instead of this silly network stuff? You know, how wrong can you be? But that's what this is all about. This is the material that we would be given when we got down there, and we'd read some other time. But this is extraordinarily valuable material. This is really it. This is not just Stanford stuff; this is the great stuff being done around the country as it's unfolding and as the people are talking about it, while they're doing it, to their colleagues. And this is not published papers.


Folder 17: Lenat Thesis

AM: An Artificial Intelligence Approach to Discovery in Mathematics as Heuristic Search
This is Douglas Lenat's thesis. When the history of AI is written it will have in it certain landmark theses that changed the way the field was. This is one of them. And this was the best thesis that I ever supervised, and this is a treasure. This is an immensely important document in my view for the history of AI.


Folder 18: CONGEN

CONGEN is just a later name for a particular part of DENDRAL, stands for Constrained Generator. The explanation of it for the historians who want to know about it is in that DENDRAL book.

COGEN Feasibility Study
CONGEN is just a later name for a particular part of DENDRAL, stands for Constrained Generator. The explanation of it for the historians who want to know about it is in that DENDRAL book.


Folder 19: CONGEN Manual 1977

CONGEN Manual 1977
CONGEN is a part of DENDRAL. Stands for Constrained Generator. It's the structure generation part of the latest version of DENDRAL. Not the latest but the mid-1970 version of DENDRAL. And there was a manual for it, and this is it.


Folder 21: Heuristic Programming Project 1980

Heuristic Programming Project 1980
Heuristic Programming Project is the name of our laboratory as it developed out of the DENDRAL Project. The DENDRAL Project was a handful of people. The key players were myself and Joshua Lederberg and Bruce Buchanan who came here as a young post-doctoral person in 1965. And it grew with several programmers and several students and several chemists. But it was still a focused and relatively small project. As we began offshoots into medicine, primarily with the MYCIN Project, we realized that we needed a broader name. Something in medicine was not DENDRAL. DENDRAL referred to a chemical/molecular program. We needed to have a broader name. And we named the growing laboratory Heuristic Programming Project, or HPP. That name was rather carefully constructed. It was no accident. John McCarthy already had a name - Stanford Artificial Intelligence Laboratory. But in any event the state of my mind at that time was not to call things Artificial Intelligence. That term had gotten us into so much trouble rubbing raw nerves that we had to make a name which didn't connote anything. It didn't rub people's nerves the wrong way. A very technical term. Invent technical language so that people won't get alarmed. Well, programming is what we did. We were software people. And the essence of what we were after was the use of heuristics, so we just called it Heuristic Programming. It's a term that had been used by Newell and Simon in the first two or three years of AI's history, so why not? We just co-opted that term. It's a rather technical term. It went along with Operations Research stuff. Linear programming, dynamic programming, quadratic programming, and here was heuristic programming. You know, aren't we rigorous? And that's how we lived with that term. That was a good decision I think, over the years. Then the laboratory maintained that terminology until approximately 1984 - 85 when so many things had grown up under that one laboratory that we split the laboratory into pieces. And the term "Heuristic Programming Project" remained with one of the pieces, but the overall thing was called Knowledge Systems Laboratory, or KSL. So around late '84 - '85 you'll see in my papers a transition from HPP being the cover story to KSL being the cover story. In HPP, the principal investigators were originally Josh Lederberg and myself. That was our lab for AI. Of course Lederberg had his own laboratory in genetics. And later we quite rightly and fairly brought Bruce Buchanan in to be Co-principal Investigator since he was playing such an important role.By 1980, we were becoming really big shots in the world. The world was really paying attention because we had invented expert systems, knowledge engineering, we were the world's leading laboratory in that we were the focus of everyone's attentions. We were getting a lot of visitors and we were proud of ourselves, and this was a stand up and be seen document. We were going to produce the loveliest brochure that anyone had ever seen. These are pictures of a lot of the people. This was done by an artist that we had helped get started in AI modeling of how drawing is done. Harold Cohen at the University of California, San Diego. This was a mural he did for the San Francisco Museum of Modern Art. And we got the picture of that. It's AI related. And we printed up enough of these so we could give one to each attendee of the first AAAI National Conference held at Stanford in the summer of 1980. I mean if we were going to be proud of ourselves and be a big deal and all that, we were going to give one of these to everyone; this was a party favor. And it went over extremely well. This was one of the hot items in AI for several years. People wanted these.


Folder 23: HPP Recent 1979

AN EXAMINATION OF A FRAME-STRUCTURED REPRESENTATION SYSTEM
"An Examination of a Frame Structured Representation System" is a key early paper on a piece of software we did that had a very big impact in the world called The Units Package. It says here, "Abstract: The Unit Package is..." The Unit Package is a big deal here. Eventually it was licensed by Stanford to IntelliGenetics, was developed by IntelliGenetics into the commercial system called KEE, Knowledge Engineering Environment, which is a main product of IntelliCorp.

Digital Communications and the Conduct of Science: The New Literacy
This paper is one further attempt by Lederberg to organize his ideas on the impact of electronic networks. And this is a very well known paper. This was a special issue of the IEEE. He wrote this on invitation. And even as recently as last weekend in the electronic mail Lederberg was writing up something for an Office of Technology Assessment Study online. We were doing it as a national group of people helping OTA get this report together. Lederberg referenced this article in his comment, and referenced the term "Eugraphy" as the thing you send in electronic mail.l You send a "eugram" instead of an electronic message or something like that. So his analogy is to a telegram or telegraphy.

SACON: A KNOWLEDGE-BASED CONSULTANT FOR STRUCTURAL ANALYSIS
An expository piece on a significant project we did in the early days of expert systems at the Heuristic Programming Project, of course, later called the Knowledge Systems Laboratory. SACON was Structural Analysis Consultant. The paper is called "SACON, a Knowledge Based Consultant for Structural Analysis". Then if the reader wants to find the details of the SACON rule -based system, they have to consult the appendix which is also given here.


Folder 25: Various Proposals

Artificial Intelligence Techniques in Legal Problem Solving - Four Papers
This is an early paper, and seminal in the sense that if you're going to track back a line of thinking, it would start at this place. AI and the Law. Stanford ran a conference on Computer Applications to Legal Research and Analysis in 1972, and Bruce and I and some other people - Tom Hedrick of the Law W School, and Thorn McCarty who was a student here, made some presentations, and they were collected up into this pamphlet here. Like Buchanan's paper, The Current Status of Artificial Intelligence Research as it Applies to Legal Reasoning. So for that reason it's important. Now AI and Legal Reasoning never really took off, so the whole thing isn't all that important, but some day it will be, and this will be tracked back. It'll be an important paper.

Box 23 - 1986

Folder 9: PhD Exam questions 1962-66

PhD Exam questions 1962-66
There's a section of the DENDRAL program called "The Predictor." The Predictor is the part of the program that is active after DENDRAL makes its hypotheses about what the organic chemical structure might be. The hypotheses are fed to The Predictor one by one. The Predictor then makes a best prediction of what the mass spectrum of that molecule would be under our best knowledge of mass spectrometry. The prediction of mass spectra are then matched against the empirical data of the mass spectrum to see which hypothesis fits best. Kind of a classical scientific loop. The Predictor therefore had in it essentially everything we knew in this program about mass spectrometry. And this is the entire Predictor program in LISP.

Box 25 - 1986

Folder 1: The Fifth Generation: Dawn of the Second Computer Age Internazxtional Conference, London. Sponsored by SPL International. (1 of 5) July 1982

This is a volume of papers here called The Fifth Generation, Dawn of the Second Computer Age, held in London by SPL International. The Japanese Fifth Generation Project was announced in October of '81. I was a keynote speaker at that conference. The British got all excited about this project thinking it was a major threat to them and wanting to use it to stimulate British research in AI. And they ran this conference on Fifth Generation. What is it, and where is it going, and what can we do about? And they invited me to speak at it. This was handed out to all the participants - either papers or their renditions of the view graphs that were done at the conference. There's also a list of all the people who attended the conference.


Folder 10: First USA-Japan Computer Conference Proceedings, Tokyo. Co-sponsored by AFIPS (American Federation of Information Processing Societies) and IFIP (International Federation for Information Processing) Oct. 1972

FIRST USA-JAPAN COMPUTER CONFERENCE PROCEEDINGS
You can sort of date the interest that the first time Americans had woken up to the fact that there was anything at all interesting in Japan in computer science and technology. The Information Processing Society of Japan and the American Federation of information Processing Societies got together to have a joint conference, run it every two years. And I don't know if it still exists. I haven't heard of it in a long time so maybe it doesn't exist anymore. As Japan grew up in the computer arena it wasn't novel to go to Japan, or to have joint conferences. This was the first time I was in Japan, October 1972 at this conference.


Folder 12: ARPA Contractors Meeting: Monterrey, CA. April 21-22, 1980

ARPA Contractors Meeting: Monterrey, CA. April 21-22, 1980 - 1979 DARPA Funded Research Summary - Data Systems Division MIT Lincoln Laboratory
It used to be a habit of ARPA's to get its contractors together once a year, or maybe once every two years sometimes, to disseminate information, what everyone is doing. And there's a tab in here for everyone, everybody's project is tabbed in here with a few pages of what they're doing. And everyone would give a verbal presentation of this, plus a list of all the attendees. So this is a history of where the ARPA Computer Science community was in the middle of 1980.


Folder 14: Readings in Artificial Intelligence: Compiled by C. C. Green and B. P. McCune, CS 224 "Models of Thought Processes", Spring 1975

Readings in Artificial Intelligence, Psych 224
This volume entitled Readings in Artificial Intelligence was done in connection with Computer Science and Psychology Joint Course 224, at that time being taught by Cordell Green and one of his teaching assistants. And he put together this volume of the papers he wanted his students to read at the time. So that was another kind of "where does the field stand" view of that time. I would regard that as significant.


Folder 15: The Programming Language LISP: E. C. Berkeley and D. G. Bobrow, ed. March 1964

THE PROGRAMMING LANGUAGE LISP:
Very early in the history of the LISP language a man named Edmond Berkeley, who had written a very famous book about computers in the early days called Giant Brains, was doing consulting for a company called Information International Incorporated and produced this little textbook on LISP long before the world had recognized how important LISP was. It was published by this company itself in March of 1964, and is a historic little book on early LISP. I'll bet there's hardly anyone in the world who has a physical copy of this book. McCarthy probably threw his away years ago.
This is probably the first book that there ever was on LISP other than the LISP manual itself.


Folder 16: AILIST Digest Beginning Vol. 3: issue 74

AILIST is a n electronic journal of little notes that people send each other, or references from bulletin boards -- just one of these electronic communication media. This happens to be targeted at a large AI community. And from time to time I would print out AILIST because it would be too hard to read online and I'd want to look at something in hard copy. And so this is a collection of those.

Box 29 - 1986

Folder 8: Cognitive Science Society

Cognitive Science Society - Membership List
I was on this list because my work in Artificial Intelligence relates to memory and cognitive processes.


Folder 31: National Library of Medicine Program Project Support for Research

Grant Application Form
This is for a National Library of Medicine project. That's another basic research project. It happens that we labeled this one Biomedical Representation of Knowledge. And most of the funds were used to help support Ted Shortliffe and his group as it got off the ground, and there is a whole proposal, and a whole packet of material that relates to that job.


Folder 43: NSF Testimony Feldman Report

STATEMENT OF DR. J. A. KRUMHANSL ASSISTANT DIRECTOR FOR MATHEMATICAL AND PHYSICAL SCIENCES, AND ENGINEERING NATIONAL SCIENCE FOUNDATION before the SUBCOMMITTEE ON SCIENCE, RESEARCH, AND TECHNOLOGY COMMITTEE ON SCIENCE AND TECHNOLOGY AUTHORIZATION HEARINGS U.S. HOUSE OF REPRESENTATIVES
This is a copy of testimony given by Dr. Krumhansl, the Assistant Director of NSF for Mathematical and Physical Sciences in Engineering, testifying before Congress on the need for, in general, more funds for science, but in particular, there is a section in here in which he asks for more funds for computer science. And we thought it was a great breakthrough on our NSF Advisory Committee to get Krumhansl to at least surface a paragraph on us in Congress.

Box 30 - 1986

Folder 32: Information Processing and Memory EA Feigenbaum. 1967

Information Processing and Memory
This is the first version on EPAM-3. The EPAM-2 paper that was reprinted in »re Computers and Thought was delivered at the Fall Joint Computer Conference, 1961. The EPAM were being published in both behavioral science and computer science journals. But they related so much to psychology that psychologists were interested in publishing material about that.


Folder 35: Feigenbaum publications pre-1965

Latent Motives, Group Discussion, and the "Quality" of Group Decisions in a Non-objective Decision Problem
This article is my first publication, written with two young professors at Carnegie Tech in the Graduate School of Industrial Administration. I was a graduate student. I was taking a course on Microeconomics from Assistant Professor Richard M. Cyert. He's now president at Carnegie Mellon. I got bored with microeconomics, but I got intrigued by applying Simon's Theory of Decision Making with the new technology of computer simulation that we were using to problems of decision making in firms, because that's what Cyert was talking about in the class. And so I approached Cyert, and Cyert approached his colleague, March, and the three of us did what we thought at the time was the first computer simulation in economics. And in fact the article that resulted from it, which is this one, was rejected by The Journal of the American Economics Association as being odd - and strange. It was published by the journal of Behavioral Science, and it became the first article in what's called The Behavioral Theory of the Firm, which eventually led to a book called The Behavioral Theory of the Firm edited by Cyert and March. This turned out not to be the first computer simulation ever done. Apparently, the second. There's one we didn't know about at the time by Austin Hoggatt. However, this was a real seminal paper and it was my first public paper. And they actually let me give the presentation form of the paper at the American Economics Association meeting in 1958 in the Christmas meetings. So that was my first presentation of a paper at a conference. Cyert and March said, "Why don't you go ahead and give it?" So I did. March of course is the same Jim March And we we've been very good friends since that time. They knew about organization theory and decision making in firms. I knew about computer simulation technologies. So that's where we fit in.


Folder 36: Feigenbaum publications 1965-68

AN INITIAL PROBLEM STATEMENT FOR A MACHINE INDUCTION RESEARCH PROJECT
Now we're into the meat of the history of one of the most important projects that we did, the DENDRAL Project. This particular file called "DENDRAL Original" has in it an actual copy of the very first thing which I ever wrote about DENDRAL. I got here at Stanford on January 1st, 1965, initiated my collaboration with Joshua Lederberg. We came up with a kind of a plan -- it's dated April 5th, 1965 -- for what we wanted to do. We -- Joshua Lederberg is not even a coauthor of this document -- were still at the Stanford Artificial Intelligence Project. We hadn't formed the DENDRAL Project yet. This is AI Project Memo number 30, April 5th, 1965 and it's co-written with a young colleague of mine, a former student at Berkeley who came with me down here to Stanford, Dick Watson, R. W. Watson. And what it does is list what we think we're going to do, and the purpose of writing this was to attract the attention of graduate students. And it was called an initial problem statement for a machine induction research project. What we were looking for was circulate this among the graduate students and find out who wanted to work on this project. This was the very first thing that was written about DENDRAL.


Folder 43: CS-72-318: Constructive Graph Labeling Using Double Cosets by H. Brown, L. Masinter, L. Hjelmeland. Oct 1972

CONSTRUCTIVE GRAPH LABELING USING DOUBLE COSETS
Double Cosets is a mathematical term. So here was the mathematics coming into DENDRAL. And the journal was discrete mathematics. Not AI, not Association for Computing Machinery, but the Math Journal. The idea of double cosets was really far out. I didn't know anything about that -- it was a giant gap. Harold's work filled a critical need. It gave us a mathematical proof that the underpinnings of DENDRAL were sound. That if we proceeded by the means that his mathematics told us, then we would have a computer program underneath that would generate all the legal structural isomers of a given chemical formula, and it would not miss any, and it would not generate any redundantly. We had to have that proof. Otherwise you couldn't tell whether DENDRAL was doing the right job or not. It could be inadvertently missing some possibilities. We needed that mathematical proof underneath. Or put another way, we just didn't feel comfortable until we had it. It was a nagging thing. We just had to get it.

Box 31 - 1986

Folder 1: Project Reviews: Heuristic Programming Project (1 of 4)1981-83

HEURISTIC PROGRAMMING PROJECT RETREAT: summaries of research projects
Then here's, in '82 the HPP Retreat at Asilomar with all the preparation material that people did for that in '82. Exactly what's going on. A lot of this material was prepared for the reviews, and then it was incorporated in this book. So for example, here's a thing labeled "Confidential. Review of Knowledge Acquisition Research at HPP." And there's the review committed,and they're going over who's doing what, and it's their critical review. Blackboard Project review. So this is not only the Asilomar book but it's also the review book. Gives everyone's material, the current state, everyone's review of everything. Some day that'll be a fascinating document.


Folder 9: HPP Retreat Asilomar 1984

HPP Retreat Asilomar 1984
We had another review at a Silimar, because this one is dated the HPP Retreat at Asilomar in 1984. And it gives the state of various projects in 1984. I guess this material was handed at the Asilomar review on a budget-planning expert system. That's what it's about. And it turned into a thing called Financial Recourses Management (FRM), which is an ongoing project.


Folder 21: Elliot Levinthal Viking Project: photographs

Viking Mars Orbiter photographs
This is historic material, which is not mine but Elliot Levinthal's. He sent it to me because I asked. But when Viking was up and they were producing news bulletins on Viking, and science was coming out with stuff like this -- this is the Viking material. A Landing on Mars. And Elliot would send me photographs like this occasionally from JPL. This is from the orbiter. I have a feeling that these relate to that thing. Let's check the dates on that.

Box 33 - 1986

Folder 1: Jan Aikens - Summary of Thesis

Summary of her thesis
Here's a single sheet in the Aikin's file, which is a summary of her thesis. She now is Chief Scientist of ION Corporation in Palo Alto. Another little AI company.


Folder 8: Doug Lenat - Plausible Mutation of DNA (Unpublished Paper)

The Plausible Mutation of DNA
We had another review at Asilomar, because this one is dated the HPP Retreat at Asilomar in 1984. And it gives the state of various projects in 1984. I guess this material was handed at the Asilomar review on a budget-planning expert system. That's what it's about. And it turned into a thing called Financial Recourses Management (FRM), which is an on-going project.


Folder 9: Dennis Brown - Draft of Uncompleted Thesis

Dennis Brown - Draft of Uncompleted Thesis
Dennis Brown, a Ph.D. student for many, many years, who was also Associate Chairman of the Computer Science Department while I was Chairman. He helped me run the department. He never finished his thesis. This is a draft of an unfinished thesis. He left before finishing his thesis, and he's the Vice President at Tech Knowledge.
Well, he left because his thesis wasn't very good and it wasn't getting anywhere. So now this is a draft up on stage. The problem itself is very good, and that's why we were working on it with him for so many years. But it wasn't worth a damn in the project. But he never made it through. He's still a very distinguished and contributory person.


Folder 12: Reid Smith - Thesis Notes and Plan

Thesis plan--detail
Some material on a draft thesis plan from Reed. Reed's thesis was not really appreciated at the time, and right now it's one of the heavily referenced works in parallel computing. And Reed works at Schlumberger Palo Alto Research Lab.


Folder 13: Peter Friedland - Molgen Thesis Notes

Peter Friedland - Molgen Thesis Notes
Peter Friedland's thesis was MOLGEN. It was a particular brand of MOLGEN. And these are - yeah, again, emerging notes from Peter that I kept copies of. Peter Friedland is now Director of AI Research for NASA as of January 1st, 1987.


Folder 21: BCPL Reference Manual

BCPL Reference Manual
BCPL had a history of coming out of university research lab where they were trying to produce a simple and relatively conventional language that would still have some very nice properties that university people thought ought to be in a language.
I think it may have originated at Harvard. But it never really made it into wide-scale use, but it did run on IBM machines which is why we reprogrammed DENDRAL into it.

Box 34 - 1986

Folder 6: Information Processing and Memory (typescripts, galleys, reprint) 1967-70

Information Processing and Memory
This is the best article on EPAM-3. This came out first in the proceedings of the fifth Berkeley Symposium Mathematical Statistics and Probability, Volume IV, Biology and Health, which is listed here in the footnote, but it was reprinted in a more readable place, Models of Human Memory, a book.


Folder 10: A Model of Short Term Memory by Jack Carroll, Jr.

A Model of Short Term Memory
Lots of people were doing their own variations of EPAM and I would help them out. Either they were doing it as students of mine or students of somebody else in someplace, and I would talk to them and give them a clear idea of what EPAM did and they could go on to then expand it in any direction they chose. One of those developments was called LEPAM, and I think LEPAM stood for long-term EPAM, but I'm not sure.


Folder 12: A Report on the Current Status of the LISP Version of EPAM III

A REPORT ON THE CURRENT STATUS OF THE LISP VERSION OF EPAM III
What happened was when I got here at Stanford I was teaching courses in Information Processing Psychology, and one of the students, as a project, wanted to re-implement EPAM in the LISP language. It had not been done and had been done in the list processing language, IPL5. And so with my supervision he went ahead and implemented it in LISP, and we actually got it running. And that presumably was a LISP called either LISP 1.5 for the PDP-6 at the Stanford AI Laboratory, or it was LISP 1.6 that was running at the Stanford Computer Center. A variation on LISP 1.5. Three is the third version of EPAM which is the - the last version of EPAM before Simon or I quit doing this in 1964. EPAM-3 - there was an EPAM-1, which was a very early thing Simon and I did. There was an EPAM-2, which my Ph.D.
thesis, and there was an EPAM-3 which contained a series of revision that Simon and I from the period 1960 through '64 to make EPAM more powerful in general. And then I didn't do any more on it. Simon did a little bit on it. And EPAM kind of passed down to history. Except for the resurrection of this and a few things that Simon did at Carnegie Mellon. The students who worked on this report are Richard Russell, Wayne Wilner, and Steve Levine. They did it in 1967 in the spring quarter..


Folder 16: EPAM Book project--reprints

Now we come to another book that never got done. Herb Simon and I decided at one point to put together a carefully edited collection of readings of EPAM and EPAM-like models. I met with Herb in Pittsburgh and we worked out the plausible contents of such a book. And then I dropped the ball. I said I would do the next step and I dropped the ball. So the book never got done. And instead Simon collected portions of EPAM in a book that he did for his own writing called Models of Thought. And these are notes relating to that book which never got done. The first and last that was ever heard about this book. Simon went on to do additional work on EPAM, and published that work with Lee Gregg, a psychologist at Carnegie Institute of Technology, now Carnegie Mellon. Gregg 'died several years ago.
They published that additional work on EPAM in this paper, An Information -Processing Explanation of On-Trial and Incremental Learning. So that's an EPAM derivative done by Simon himself. Then Simon continued to think about EPAM as he was considering how expert chess players recognize situations on chessboards. And he wrote this article which is related to EPAM on Perception in Chess Playing Programs. So if you read this there are EPAM-like mechanisms in this program, and that's why I was collecting it, Because if we ever did that book I would need to put this paper in. EPAM is mentioned here as a venerable simulation program first devised by Feigenbaum to explain.


Folder 17: EPAM Book project - papers by others

A model of Concept Learning
This is material -sent to me by George Ernst, talking about some work he did which was similar to EPAM. And you can see that EPAM begins to be mentioned immediately. The program is called Concept EPAM, or CE, to distinguish it from other versions of EPAM. And I must have read it at the time. It's made no impact on my life, my work, or my consciousness, but it's EPAM related

Box 36 - 1986

Folder 2: Addenda to Computer Thought1977

Addenda to Computers and Thought
This is the remnants of a pile of Addenda to the book Computers and Thought that I had mimeographed sometime after the book came out. We were never really going to do a second edition of the book. That didn't make sense. But to my students who bought the book, and anyone else I could figure out who'd had the book I would send one of these things to. It would give expansions and corrections to the book. So I probably handed out hundreds of these.


Folder 3: ARPANET

ARPANET - Various Messages, November 1976
Someone beginning an electronic correspondence with the ARPANET principal investigators, or users actually. At the time you could send this message to all network users. The next generation of an ARPA-like telecommunication network is being contemplated.


Folder 4: SUMEX Renewal1975-80

SUMEX Renewal1975-80 - Letter from Suzanne S. Stimler to Edward A. Feigenbaum
This document gives 1979 information on the budgets that were funding the National Institute of Health Biotechnology Resources. SUMEX, our computer facility was funded through them. And so someone from the NIH, Suzanne Stimler, Director of the program, sent me fiscal information on their program; it's a state of funds.This relates to my SUMEX activity.


Folder 13: Satellite Business Systems: High Bandwidth Communication Link 1977

This is something which never actually came into existence. So it's one of those ideas that gets talked about a lot, you spend a lot of time on it, and then it's ahead of its time and it never happens. This is what's called the Inter-University Satellite High-Bandwidth communications among a bunch of universities. Today a representative at Carnegie Mellon, IBM, Satellite Business Systems, Stanford, and MIT met to discuss the prospect of an Inter-University link by satellite. These are memos on that. And then it just never happened, and we could never figure out quite why it doesn't happen, but that's the file on it.


Folder 14: NSF proposal for equipment 1977

NSF proposal for equipment 1977
This file with the name "King" is for a National Science Foundation proposal related to the Computer Science Department's attempt to get more computer equipment for the department. King is Jonathan King who was a graduate student here from many years, finally got his Ph.D., now works for Technology Incorporated. He was hired by me on a one-year leave of absence from his Ph.D. work in what was really my first full year, maybe at the beginning of my second full year as Department Chairman, in order to help me do fundraising for the department. He was in charge of actually writing the proposals on sketches laid out by me. He would whomp them together and we'd get equipment for that. He worked for one at that. That's why he's the author of these documents.


Folder 17: General Professional (includes photograph of Feigenbaum in an unnamed group portrait) 1976-81

Carnegie Mellon University Alumni newsletter, including section on their new Robotics Laboratory
This comes from the Carnegie Mellon Alumni Newsletter, and it's pictures of the opening of their Robotics Laboratory with Raj Reddy giving the keynote speech, and Alan Newell's here talking to the president of Westinghouse.


Folder 20: National Science Policy/Administration1978-79

National Science Policy/Administration1978-79
The National Science Foundation has a continuing committee in the Computer Science area called the Advisory Subcommittee for Computer Science. It's a committee of computer science wheels who get together from time to time in Washington to advise the person who's in charge of giving out money for computer science. His name is Kent Curtis. In other words, this committee reports to Kent Curtis. I was on it for several years. And this is minutes of the meeting of December 7th and 8th, 1978 with a list of all the people who are on this committee. That's what transpired at the NSF Advisory Committee meeting. We're advising him on what are good ways to spend the NSF money. What good programs should there be. And it also had an auditing function in that it reviewed the reviewing process at NSF for computer science to make sure that the best things were really being funded, to check on the reviewing process.


Folder 22: Computer Science Department

Announcement about Faculty Seminars
This is Fred's attempt to organize a thing called Faculty Seminars to get Stanford to get together to discuss important subjects. This is dated 1979.

Computer Science Department - Mills-Stanford 3+2 program
Computer Science Department arranging a joint program with Mills College for Computer Science. Nobody knows that existed. I worked it out.

STANFORD COMPUTER SCIENCE DEPARTMENT RESEARCH REPORT
The Computer Science Department itself produced each year, and I think it still does, a thing called Research Report, about what's every professor doing in research. This is the 1976 one.


Folder 23: Articles by others

Articles by others
If you were to look around in Expert System software that's available in the world now you'd find some of them called Induction Tools, Induction Software. For example, a name of one is ExpertBase. It's all based on Ross Quinlan's work which he did here when he was a visitor from Australia, and this is an original paper.


Folder 28: Lecture Notes

Lecture Notes on Models of Memory
I give a lot of lectures, and I always write out my notes by hand. So as you know from looking at the other notes I mix and match pieces of lectures. So one piece of note could find its way into ten different lectures in different places. And this is a collection of those notes that have to do with Models of Human Memory. It starts out with notes about EPAM, and you can never tell what's going to end where because of the mixing and matching. It describes a lot more than just EPAM. I see here HAM, which stands for the memory model by Gordon Bower and John Anderson, Human Associate of Memory - HAM. This down here is the thesis by John Anderson. And some other people Ronald Heartland, Don Norman -- these are all various people's models of how memory might be handled. But somewhere I gave a talk on human memory and put these all together. So it's on EPAM and other models of memory. See, EPAM was a model of memory, but it was a model of a particular kind of memory, and as the years went on people developed broader models.


Folder 31: DOLPHIN INTERLISP Technical Characteristics

DOLPHIN INTERLISP Technical Characteristics
The prototypes of the Xerox LISP machine were called Dolphins. The software system running on Dolphins was called Interlisp. And this is a Xerox Corporation informal blurb on the technical specs for Dolphin Interlisp.


Folder 32: Miscellaneous Files - C. S. - University/Industry Relations1979-82

Miscellaneous Files - C. S. - University/Industry Relations1979-82
This is Charlie Bass at Zylog sending a letter to Don Knuth's research guy, Luis Trabb Pardo, about how Zylog is willing to get behind Don Knuth's TEX Project. And Charlie Bass of course was the Charlie Bass of Unger & Bass. Nobody knows that he was at Zylog.


Folder 33: Computer Science Department, administrative matters, 1976-80

Computer Science Department, administrative matters, 1976-80
This is a never-ending problem. We've just gone through a major salary survey for programmers for our group because we think we're underpaying them. We have to convince personnel of that. And here's exactly the same thing. September 1st, 1976. It's a memo from Ed Shaw. "Special Salary Adjustment for Computer Programmers." Now he's a Department Chairman.

Box 37 - 1986

Folder 4: General Memos 1969/1970 (1 of 2)

General Memos 1969/1970 (1 of 2)
This thing, Statement by William F. Miller Accompanying Release of Salary Data. There was pressure from the students during the whole Vietnam War uprising thing to release information about salaries around here. Don't ask me how that arose, but Miller wrote this memo. By that time he must have been Provost. .Right here are all the salaries.

Box 39 - 1986

Folder 1: AI Handbook 1975

Early notes for outline of the subsequent "Handbook of Artificial Intelligence"
These are some notes for CS 225, which was an AI course I was teaching at the time, dated October 9th, 1975. There's a draft here called "Feigenbaum's Proposed Structuring of AI". It has to do with the Handbook of AI; a very early outline. We would sit there and we'd discuss what ought to be there and what the sources are, and we'd write it down like this. Different subjects. These are just different student outlines/proposals of things. Some became articles and didn't get into the book, like Hofstadter's articles on music. Looking at this stuff, I'm reminded once again of what an enormous amount of intellectual work it took to do this handbook of AI. I once told someone that it was the most difficult professional thing I've ever done in my life.

Most Important Contributions of AI as seen in 1975, looking backward, done for COSERS AI panel
Here's how the Top Ten idea happened (see document sn759jy8378). Newell was serving on a committee that was putting together a massive report for the National Science Foundation. The committee was called COSERS, and it came out with a book eventually. It's a couple inches thick, and for its time, it was the statement about what computer science was about. I believe it was edited by Bruce Arden, the Chairman of Electrical Engineering at Princeton. So Newell is asking that in order to put together this COSERS report, the part that deals with AI, each of us wrote down what we thought were the top ten scientific results in AI so far. "Why don't you write down your own set often and ship them to me by mail. I'll let you see the whole set when I get it on file and after I receive yours to keep you uncontaminated."

Proposal to ARPA and "Computer Vision" by A. Rosenfeld, July 28, 1975
This is proposal material. The reason proposal stuff is interesting for the AI Handbook effort is because proposals try to structure work, and that's what people were doing at the time. These are the tasks and what the MIT AI lab was telling ARPA that it was going to be doing. General objective to make machines smarter, to make machines more useful, to understand intelligence, and then they list the details for that. Recent progress.

Responses to "What are the Top Ten contributions of AI?", as seen in 1975 looking backward
Top Ten was a cut at trying to put together the material on AI. One way you can do it is to go through all the different topics and say this is what people have done. Another way to do it is to say what are the ten biggest things ever done in AI? What is "it"? What's the top ten "it"? And so I started to ask people to do that, so that we could highlight what was the most important stuff. We'd take the intersection of a lot of different people's views of what the Top Ten, and that's what this section is. John Gaschnig, SRI, Artificial Intelligence Lab. John was a student of AI Newell's at Carnegie Mellon. John came to SRI and died of cancer there sometime in around 1980. Very smart young kid. And, for example, here's Hans Berliner's -- another researcher at Carnegie Mellon -- view. And then there's Gaschnig's notes on this. Here's Hans Berliner's list. Lanny Forgy, another student of Newell's at Carnegie Mellon, who's now a researcher there. Here's Gaschnig's response, although he gave twelve. This is Elaine Rich, a student of Newell's at the time, subsequently the writer of one of the best selling textbooks in AI. She's at the MCC in Austin, Texas, Microelectronics Computer Technology Corporation. Then some more notes by John Gaschnig, a call for opinion. Then we started to write down the top ten failures of AI.


Folder 2: Handbook of AI 1976

Notes on outline for Handbook of AI
Looking here at a thing called "Domains in Experimental and Cognitive Psychology of Potential Relevance to Artificial Intelligence, written by Peter Thorndyke. He works as head of the AI Group at FMC right now. He was a student at the time in psychology. It has very few notes on it. The notes relate to some references to people's work that are missing from this outline. From the words at the top, which say "Done. Irrelevant. Needed. Selected. Modified. Relevant but postponed," I can only assume that this outline was part of the beginnings, the thrashing around of the beginnings of The Handbook of Artificial Intelligence, asking people what stuff they thought should be in this handbook, what was the relevant work, trying to collect ideas for how the field itself was organized. And this is somebody's version of that. "Done" means one of these articles is already done or irrelevant, or "Needed" means we have to go get one, that sort of thing. There are some notes for the Heuristic Search Chapter of the handbook. And I think these asterisks mean these are articles that were either selected to be done or were done by this time -- this is sort of a tracking outline saying this is what we're going to do. There's this thing called Overview of Heuristic Search, which finds its way into the handbook. Then there's the note that says "listings of .DONE," means that the various articles in the handbook would have a name like "Heuristic-Search." And when it was finished, the version of the finished one would be filed under ".DONE." And so this is a note to myself to make listings of the ".DONE" articles so I could read them. And these are the people that we were contacting with the outline to ask if the outline was adequate. These were our friends that we were contacting. We were online, and we're kind of doing a national debugging of this idea for The Handbook of AI. And here, it says, "Send outline to Amarel, Newell, Bledsoe, Terry Winograd, Bob Bobrow, Nilsson, McCarthy, Buchanan, Thorndyke." These are people we were consulting. Then there's an AI Handbook outline, so this is what we were developing in those meetings, trying to get the plan together. And this is it at some date which is not listed here, but you can see it developing little by little. The notation "Need" and a number follows the proposed focus of the article where number is a number of the interval zero to ten. Low numbers indicate little expected difficulty with the article, whereas high numbers indicate a potentially difficult article. A "need[7]" would say that's going to be a fairly hard article to do, and the overview of vision was going to be even harder. For example, Overview of Memory Models and Representation was a "need [10]". Very hard. We were going to give that to our best people, and wouldn't expect it for a while. An example of [2] would be topics like Breadth-First Search, Depth-First Search, where it had been worked out, like Q-lisp, Safe Planner, Conniver -- we know those cold, write 'em up.And this also gives, in the introduction, who the audience for the handbook was supposed to be, and the suggested style of an article. So from the very beginning we were keeping in mind who we were writing this for, the marketing direction so to speak, and that never changed. And the selected style for articles we were really trying to enforce a uniformity and a discipline to the extent that we could.And then in the end we couldn't, of course, enforce it because people do write differently. So in the end it was an extraordinary editing job of homogenizing the presentation stuff, and that was primarily done by Avron Barr and Dianne Kanerva. Dianne Kanerva just died a few weeks ago (today is 1986). We're going to dedicate Volume IV of the handbook to the memory of Dianne. She was so important to this project. I'm afraid that history will overlook that. And then there's Allen Newell from Pittsburgh sending in his comments, and of course those were critical because we pay a lot of attention to what Newell says. So Newell writes a fairly lengthy review of some of the outline. "Received the outline today. It does indeed look like a worthwhile project." And then he goes on to give details. And we commented on this with some notes on the side here. He says he'd "like to be kept informed for obvious reasons since it is a good foil for my thinking bout the AI book." He must have been thinking about his own book about AI, which never came to be. I want to point that when a message from Newell would come in like that we would pay very serious attention to that. That's like a message from God coming down from Mt. Sinai.


Folder 7: People AI Handbook 1975-78

People AI Handbook 1975-78
Here's a computer listing from October '75. This group was meeting all the time,with all their home phone numbers and their office phones. These were people who were all gathering, students, and we were all thrashing around working on these articles for the AI Handbook. Doug Lenat, who's a famous figure in AI now, was there. Peter Friedland.Doug Hofstadter, who wrote the articles on AI in Music, which we never used -- couldn't quite cope with Doug's material. But this was when Doug was still a student. He was a postdoctoral student with Pat Suppes. So that was the first batch of people who worked on it.


Folder 16: Applications 1978-79

email containing overview of THE RAND INTELLIGENT TERMINAL AGENT (RITA)
Here's an electronic message from Larry Fagan who was one of the students at the time. Larry Fagan is now Dr. Fagan, Ph.D., M.D., and he's Ted Shortliffe's Chief Lieutenant at the Medical Computer Science Branch of the KSL over in the Medical School Office Building. Larry was a student at the time.
He was enlisted to do some of these articles. He sent me these comments. He had read some some early articles. He's pointing out some real problems and saying, "I just hope I don't fall into any of these traps while writing my article." He was writing articles on Speech.

Box 42 - 1986

Folder 54: CIS

CIS - CIS Proposed Policy on Software
In the year 1981 there was a very large discussion at Stanford as to how Stanford should handle software from a royalty and licensing point of view. The issue actually got up to the CIR, which is the main university committee on research. The policy went up to that for approval, and it was much discussed before that. Here's a memo from Jeff Ullman called Proposed Policy on Software, March 1981. This is a memo from Bob Dutton of EE to the CIS Executive Committee on the whole software issue. Dutton was interested in it because of a big piece of software that was very interesting and very much used, called SCALD. That had been preceded by another Stanford program, whose name I can't remember, that was done at the Stanford Artificial Intelligence Laboratory, but was widely used both by Dutton and by DEC. Well, Dutton had an interest in commercializing SCALD. that eventually became one of the Silicon Valley companies. I don't know which one but Dutton could tell you -- VLSI Logic or one of those. And so he was really interested in the whole issue. And I was interested in the issue. And the whole Computer Science Department was interested in the issue of who owns the software and what would be the right to it that the developers would have if they wanted to commercialize it. This was part of the development of that. The policy went up to the Committee on Research. And they approved a real university policy. Jerry Lieberman was in charge of formulating the policy, and there is an official university policy now that's still in place today (1987).

Box 43 - 1986

Folder 5: Computers and Thought (correspondence and reviews) 1963-67

Computers and Thought (correspondence and reviews) 1963-67 - Various
Royalties were never paid. As much as we tried to get them, it's an interesting historical note that all attempts failed.


Folder 6: Computers and Thought Lecture 2 Patrick Winston (text), Aug. 21, 1973

Computers and Thought Lecture 2 Patrick Winston (text), Aug. 21, 1973
We started a lecture series in the field, which for many years was the only award lecture in the field of Artificial Intelligence called The Computers of Thought Award Lecture. From the royalties of the first anthology of Artificial Intelligence, Computers and Thought, done by Julian Feldman and myself in 1962 and '63, the royalties were donated to the Association for Computing Machinery. Eventually the royalties found their way into the control of the International Joint Conferences on Artificial Intelligence to give a prize and a lecture called The Computers and Thought Lecture for young people in the field who had made outstanding contributions. The first of those Computers and Thought award lectures went to Professor Winograd, who's here at Stanford now. And a later one went to Patrick Winston. 1973 lecture went to Patrick Winston, now Director of the Artificial Intelligence Laboratory at MIT. This lecture apparently was tape recorded, and this apparently is a transcript of that lecture. The information here is about it. It's session 13, Computers and Thought Lecture, August 21st, 1973. And there is a front part of this which is a statement by me about the history of this lecture, a more detailed description of what I've just given you. And then it says, "I'd like to introduce Pat Winston, who will present the second Computers and Thought Lecture - Second Computers and Thought Lecture." And then there's a break, and then Pat comes on, and this is a transcription of Pat's talk. The lectures were always given at the International Joint Conference on Artificial Intelligence held every two years. So this is Pat Winston, second Computers and Thought Lecture. These conferences are called IJCAI, International Joint Conference on Artificial Intelligence.IJCAI '73, this would be.

Box 44 - 1986

Folder 38: Lighthill Debate

FIRBUSH Newsletter, including comments on famous debate with Sir James Lighthill
The biggest event in Artificial Intelligence in Britain in the 1970s was a publication of a report critical of the work by Lighthill, the so-called Lighthill Report. This was the Lighthill Report with rejoinders by various people. This is the official document that came out of Britain about the Lighthill Report. This report essentially killed Artificial Intelligence in Britain for ten years, and it was only revived in the new Alby Program under the term Intelligent Knowledge Based System, IKBS. They can't even use the term Artificial Intelligence, even now, because of this report.

Box 46 - 1986

Folder 57: Kansas State Visit

lecture at Kansas State--schedule of lectures by various lecturers
I give a lot of lectures and have always given a lot of lectures, and I always write out outlines by hand. This represents a lecture given in 1968, April 15th. So how these lectures are structured, first four in the area of Computer Models of Thought Processes Fifth. So it must be five lectures that I gave there. And I see here a categorization of different programs that existed at the time.

lecture notes concerning time-sharing and Stanford approaches, given at Kansas State University
I give a lot of lectures and have always given a lot of lectures, and I always write out outlines by hand. This represents a lecture given in 1968, April 15th. So how these lectures are structured, first four in the area of Computer Models of Thought Processes Fifth. So it must be five lectures that I gave there. And I see here a categorization of different programs that existed at the time.

Box 47 - 1986

Folder 40: Molecular Structure Elucidation III (HPP-74-7)

MOLECULAR STRUCTURE ELUCIDATION 111
The problem of enumerating non-cyclic molecules was much, much, much easier intellectually than the problem of enumerating cyclic molecules. The latter became or was tied up with some of the most sophisticated problems in graph theory of the time, for which there were no theorems. That means that although we did not think of ourselves as mathematical graph theorists, we were at the cutting edge of mathematical graph theory trying to prove new theorems to apply this stuff. We weren't' the only ones. Other computer scientists were trying to do that too, but not related to chemistry. Like a young person in this department who's now at Princeton, Bon Tarjan, was an expert on that, Donald Knuth was interested in exactly those kinds of algorithms. Well, one way to make progress on that if you're stymied, if you're a geneticist like Lederberg who's brilliant but not a specialist, or if you're a computer scientist like myself who's not a specialist in graph theory, is to hire someone who is. Knuth had a visitor here during one year. Sabbatical visitor from Ohio State Math Department named Harold Brown. And Harold suddenly found as we began to talk about these problems, Harold suddenly found people who were interested in what he knew how to do, people who loved his work, and that's extraordinary. Most of the time nobody loves mathematicians' work, nobody pays any attention except people who really wanted the stuff and they had a focused need for it. So Harold got attracted to us and we to him because of what he knew. And he worked on our problem, and then he never left, never went back to Ohio State. He became a researcher on the DENDRAL Project, produced some extraordinary work of which this is one, this paper. HPP74-7, Molecular Structure, Elucidation III, one of a series of things on cyclic graph structures. Notice how mathematical it is compared to any of the other stuff. These are the real proofs, and the algorithms for how to do it. Extraordinary. And he left here for a time and worked at NASA Ames, and then he came back and he's one of the key scientists on the HPP Net.. He's a Senior Research Associate. But he gave up a tenure job at Ohio State to do that.


Folder 45: Search Strategies for the Task of Organic Chemical Synthesis (HPP-73-9) (version: SAIL Memo AIM-217 or STAN-CS-73-391) 1973

Search Strategies for the Task of Organic Chemical Synthesis (HPP-73-9) (version: SAIL Memo AIM-217 or STAN-CS-73-391) 1973
We found a computer science student who knew a lot about programs dealing with organic chemical molecules. He was doing a thesis on that. His name was Sridharan, an Indian, and we hired him as a researcher after he got his Ph.D. His thesis was in the area of organic chemical synthesis, not analysis. So he wrote a paper on what his thesis research is all about, but don't get this confused in the historical records with anything we were doing here. It's just that he was here, but this was his thesis work.


Folder 48: An Algorithm for the Construction of the Graphs of Organic Molecules (HPP-73-13) (version: STAN-CS-73-361)

An Algorithm for the Construction of the Graphs of Organic Molecules (HPP-73-13) (version: STAN-CS-73-361)
Larry Masinter, who's named on this papers, was a math graduate student at the time who came over to work with us on these problems. Solved some very difficult problems for us, got his Ph.D., left, and is a Key Researcher for Xerox Palo Alto Research Center. You might say he's the Chief Developer of the Xerox LISP Machine. So this another important DENDRAL article.

Box 49 - 1986

Folder 49: Hintzman, Douglas, papers undated

Explorations with a Discrimination Net Model for Paired Associate Learning
Doug Hintzman did a version of EPAM called SAL here at Stanford, and it won the prize for the best piece of work done by a graduate student in Psychology in that year.
He started this project as a graduate student working in the course I was teaching. And his main advisor was Gordon Bower. The connection to EPAM is described in the introductory section of this article. The word "discrimination net" was invented by me for my Ph.D. thesis. So when they talk about a discrimination net model of verbal learning, that's just what EPAM was. He did another version of it.

Box 55 - 1986

Folder 1: Correspondence: Weizenbaum, Joe to Feigenbaum, 23 December 1970

Correspondence: Joseph Weizenbaum to Feigenbaum,
Now this is a very interesting letter. This is dated December 23rd, 1970, from Joe Weizenbaum. Joe Weizenbaum is the author of Computer Power and Human Reason. It was the first intellectually powerful blast at the field of Artificial Intelligence. And Weizenbaum has emerged over the years as AI's major critic with credibility that arises from being inside the field of computer science and AI rather than outside. The other critic is Hubert Dreyfus, but he's not well thought of because he's seen as an heir in philosophy and no one should pay any attention to him. But Weizenbaum is inside and this is a long handwritten letter. It says, "Thank you for the very kind letter. That my mood saddened you grieves me. But then you yourself gave evidence that my feelings were based in reality." Anyway, it's probably the beginning of Weizenbaum's souring on AI, and I was probably trying to convince him otherwise. "I remain very impressed that your program has the same gross structure as Moses' SIN Program."


Folder 2: Preliminary report to DARPA, 9 November 1973

Artificial Intelligence Research: What is it? What has it achieved? Where is it going? (another early draft with comments)
This document was prepared on the Stanford AI Lab computer. I can tell by looking at the name of the file. So this was pre-SUMEX. The draft of this report as it is emerging is called "Artificial Intelligence Research: What Is It, What Has It Achieved, Where Is It Going?" This document is a return from somebody who who was commenting on it and left some comments for me. For example, one comment is "If anyone but Lukasik is to see this you will have to pay more attention to work that isn't yours and your students." I can't tell whose handwriting this is. I have a feeling it's AI Newell, because I remember - well, first of all, this top comment is very Newell-ish in style. "As you say, it is very good." So I suspect these are Newell's comments.

Box 56 - 1986

Folder 1: Printout: EPAM II, code LIPLV, 1961 (1 of 3)

Printout: EPAM II, code LIPLV
These pages are the code for EPAM-2. So this is the program the lead to my thesis. This one is dated August 1961. And this is everything. I mean now I could probably resurrect how EPAM worked just by looking at all the subroutines of this code. This is a giant program for its time, one of the largest programs in AI in existence at the time. So this is what evolved from - from about '57 to '61, including the point of my thesis. Then this changed into this. EPAM-2 became EPAM-3. So I kept them together.
If you want to look at what EPAM-2 is you'll have to look in the Anthology Computers and Thought, is what I did in '62. And there's an article thereon EPAM-2. And EPAM-3, there were several important articles on that.
Simon and I actually had to go back to the code for EPAM-3 to settle an argument that was started in the Journal Cognitive Science by a Stanford graduate student and Gordon Bower. Together they wrote a piece which was simply wrong about EPAM-3. And Simon got really agitated and wrote a draft article as a rebuttal, and I had to bolster the draft article by going back to that document and unscrambling the code and saying this is exactly how it was done.