SUBSCRIBE NOW!

Vol. 11 No. 9 — October 2003

Shaping the Online Industry

Roger Summit taking a solo on “Cupertino Blues” at a performance of the Daddio band in June 2003.

Ardito: Dialog is certainly one of the first, if not the first, names associated with the online information systems that emerged in the late 1960s and early 1970s. How do you think your work with Dialog shaped the growth of the industry?

Summit: What I think Dialog added was massive, consistently formatted databases and a search language providing flexible yet precise search capabilities. Such multiple, massive, consistently formatted databases allowed cross-database searching.

I’ve sometimes said that the reason broad subject classifications exist is for the convenience of university professors and librarians. In real life, information problems cut across the disciplines. You can’t solve the complete problem within chemistry or environment or some single category like that. A complex problem is likely to cross over several subject boundaries. Very early on, it was my feeling that searching across several discipline areas could be quite interesting in terms of symbiosis and knowledge transfer. To the extent that the same kinds of words might be used in different disciplines, doing a search across multiple disciplines might reveal items in the areas of another discipline that would apply, that you might not otherwise be aware of.

Bjørner: I’m remembering a couple searches I had at MIT when we searched all subjects and did searches with end users by our side. We did it as teams, the academic researcher and the librarian/research specialist. One graduate student came in and actually wanted everything there was on the word “stress.” I said, “Are you looking for psychological stress, stress on building materials, biological stress…?” Another search was on sandwich board, which in this context meant a type of construction material, but my first mental picture was a cafeteria.

Summit: Yes, this process has been termed the search negotiation. The art of defining the area of interest is quite an art — quite an art and quite a skill.

Bjørner: So enabling the user to have those classifications of subject matter but also to move across them was a good innovation?

Summit: Yes. That’s one difference we introduced.

A second advantage of offering a wide range of databases has to do with the way people behave. If one has a problem, one likes to sit down, attack the problem, work it all the way through to solution. People aren’t multi-tasking machines by nature. They don’t like to start work on a problem, do a little bit, set it aside, start another problem, work on it for awhile, then come back and continue work on the first one again. My feeling was that if Dialog could have a comprehensive-enough spectrum of data available, one could indeed attack a problem in one sitting and be done with it.

Youth and College

Ardito: Let’s go back a little in time. We’re asking everyone we talk with about the formative experiences of their early life. We know that you have a bachelor’s degree in psychology but we want to ask a little bit about how you got to that point.

Summit: I was fortunate. I had a Naval ROTC (NROTC) scholarship for college that allowed me to select from any one of a large number of universities in the country, provided I was accepted. After taking the standard admission tests, I considered three schools: Stanford, University of Southern California, and the University of Michigan. I was accepted at all three, and I picked Stanford.

I’m originally from Detroit, Michigan. I grew up in Dearborn. When I came out to California to begin university, I thought I’d arrived in heaven. I wondered, “How in the world will I ever find work here?” There was nothing here — just apricot orchards and a cannery in San Jose. But as time went on, during the 1950s, the Stanford Industrial Park came into being. That was a great innovation. I think it was one of the first in the world. It got the university and industry working very closely together, and it provided a work avenue for graduates.

In growing up, there had been no question about my going to college. My parents were teachers. My dad’s profession was teacher and guidance counselor, so I think we kids never thought that we might not go to college.

Bjørner: Did you know before you went to Stanford what you wanted to do?

Summit: I had no particular direction. Psychology was interesting to me, so I took that as an undergraduate major. My brother, 4 years younger than I, is a psychiatrist; maybe it runs in the family.

Ardito: We know you play the piano well; was yours a musical family?

Summit: My parents were musical, yes. My mother plays by ear, and my dad played the organ; he accompanied some of the silent films on piano and organ. I took up trombone in junior high at my mother’s suggestion and made spending money in college playing in dance bands. I put the horn away when I got married, but I have recently taken it up again and now play in two big-band, jazz orchestras in the Bay area.

Bjørner: When did you actually leave Detroit?

Summit: 1948. I visited the West Coast in the summer of 1941. My parents arranged the trip, and it was a most fascinating experience. I’d never seen a mountain, I’d never seen the ocean, and I was intrigued. I spent the summer in Los Angeles at a Yoga religious retreat called “Self- Realization Fellowship.” (The group still exists: http://www.yogananda-srf.org.) During the summer another “inmate” (a 16-year-old named Arnie) and I spent many hours fantasizing over the details of potential relationships we were looking forward to in our later lives. We also snuck out to see the Harry James and Woody Herman orchestras live at local theaters. That summer resulted in a much “broader” experience than I’m sure my mother had intended.

Military Service

Ardito: Did you go on immediately to get your master’s degree?

Summit: No. I had 3 years of obligated service in the Navy first. I’d signed up for flight training, so I was assigned to an aircraft carrier. One night, the ship was conducting air operations under darkened ship conditions. It’s tough to land on an aircraft carrier even in broad daylight, so I went up to the signal bridge to watch what was happening.… I’d hear the planes roar in and I couldn’t see them at all. They’d roar in, they’d land, and then they’d take off again. The landing signal officer had darkened paddles. The only lights on the whole ship were four lights on each corner of the aircraft carrier deck, and these planes were roaring in and taking off again. It was all quite frightening. I thought about the pilots, “These are some kind of super-duper human beings, and that’s not for me.” I withdrew from the flight program the next day.

My tour of duty was from 1952 to 1955, right after college, during the Korean War. We were engaged in flight operations off the coast of Korea, and we went into Japan for liberty and so-called “R and R.” We got reports of Chinese submarines in the area, but probably were not under any real threat. When I first went aboard the ship, I was a radio officer, which meant that I was in charge of the radio division of 120 men. Here I was, fresh out of college and suddenly I was in charge of 120 sailors. That was an experience.

Let me tell you about one incident that was very interesting. When you’re in charge of a division as a Navy junior officer, you frequently go see the Chief Petty Officer, and you say, “Well, Chief, what do you think we should do about this?” It was budget time, and we were making our budgets. He came to me and said, “We have a surplus in our budget.” I said, “How much?”’ And he said, “Oh, four or five thousand dollars.” I said, “Well, what do we need?” And he said, “We don’t really need anything, but we’ve got to spend it, otherwise they’ll cut our budget next year.” I said, “Well, if we don’t need it, let’s just give it back.” “Oh, no, no. We can’t do that.” We had a long conversation about what to do with the surplus. I’ve thought about that situation many times since then, because the same practice probably prevails throughout government. If there were some way one could instill an incentive to come in under budget — like you’d share the difference as a bonus — we would save billions of dollars of needlessly spent budget money.

After my obligatory three years, I left the Navy. It was in the Navy that I realized I didn’t understand business and the business process. I remember a conversation with another officer who recommended buying Texas Instruments stock. At the time (1953), this was a new company and would have been a good one to buy, but the reasoning was beyond me. I had no inkling of how business functioned and felt insecure as a result. Consequently, attending business school seemed like smart thing to do..

On to Business School

Summit: When I left the service, I got the GI bill. That entitled me to 4 more years of government-paid education. I went back to Stanford, to the business school, to get an M.B.A.

Bjørner: So there was never any intention of using your psychology degree, or did you think you might use the psychology with the M.B.A.?

Summit: No, it wasn’t all that planned out. When I was studying psychology, I thought I might become an experimental or clinical psychologist. I had a girlfriend in college, a microbiology major, who in my senior year mentioned, “Do you know there’s a new thing called a computer?” We spent a lot of time talking about computers, and I was fascinated. With my psychology interest and this machine that you could program to branch and take different paths depending on conditions, I knew that’s what I wanted to do. My career path was guided by that relationship.

By returning to the business school, I filled in all those insecurity cracks in business and finance that I had felt in the Navy. But I also took a couple of courses in the computer sciences area. They had one of the very early machines, an IBM 650, which almost nobody has heard of. Nobody knew what to do with it. Bob Oakford was the professor for a graduate seminar class where we all just figured out how to program. We talked about how these machines might be used in the future. We thought of point-of-sale operations and how neat that would be, because then you could have your inventory instantly updated … this kind of thing. It was exciting.

After I got my M.B.A., I went to work for Arthur Andersen, the accounting firm, in their systems department. They were doing computer applications. I worked on the Bank of America/Visa application that used to be BankAmericard before it became Visa, and did some very elemental systems work. It turned out that Arthur Andersen didn’t have enough systems work, so they put me on the audit staff. That was the most deadly, dull, awful job I’d ever had. I quit after a year.

I had some other mentoring. A woman that I knew, Ellen Urbach, worked at the business school. I’d known her as we were getting our M.B.A.s. She said, “Roger, you ought to come back because there’s a guy here, Dan Teichroew, who’s sponsored by IBM to foster the teaching of computer technology in the business school.” On application, I was awarded a scholarship from IBM, and I still had 2 years of the GI Bill to cover the tuition for my Ph.D. work. I didn’t return simply to get the degree, but rather to fill in the mathematical gap in my background and to get some computer science. Most of my courses were taken outside of the business school. Dan Teichroew became the faculty advisor for my orals and chairman of my dissertation committee. After the orals, I went to work for Lockheed as a summer hire, and that’s what started this whole information excursion.

Ardito: Was the Ph.D. in operations research?

Summit: Management science was the major, but a Ph.D. is a Ph.D. It is a doctor of philosophy degree, but my specialization was in an area that was called management science, a combination of operations research and statistics.

Bjørner: I’m sure I could look this up, but what was the dissertation on?

Summit: I will give you the exact title. Of course, you can find it in Dissertation Abstracts too, on Dialog. The title is “Simulation of a Management Decision Process Utilizing a Computer Model of the Aerospace Industry.” So much for short titles!

Business games were big things in those days. I didn’t like the ones that were around, so I decided to do one that would behave more like real life — the psychology comes in here as well. The model, called The Aerospace Business Environment Simulator (ABES), is actually a business game that allows management teams to compete in bidding on contracts, receiving awards, hiring and firing employees, setting pay scales, and the like. It is kind of a manufacturing/scheduling operation.

One of the interesting interactions in the model is a function called, “employee satisfaction.” If management set pay scales too low, relative to pay in the industry, it affected the employee satisfaction function and some of the employees quit. The model was rigged so that employees who were fired came from the bottom of the productivity scale, whereas employees who quit (to go elsewhere) came from the top of the productivity scale. Lowering pay meant that you lost productive employees and were likely to overrun contract schedules, which reduced chances for future contract awards. Players felt the game to be quite realistic.

I taught a decision-making class to Lockheed senior management utilizing the ABES model in 1962. Lockheed later licensed ABES to IBM and gave me a “Special Invention Award” as a result.

I recently did a Google search and noticed that ABES was (and may still be) used at Washington State University in 1999, in an engineering class called: “Strategic Management of Technology and Innovation.” What a nice surprise!

Those were my simulation days — simulation and information retrieval — and I thought the ABES model would be sufficient to base my dissertation on, but Teichroew said, “No, that isn’t enough. What you need to do is to build a decision model, a computer model of decision-making that will operate one of these companies, in competition with humans, without any more information than what the humans have, and see what you have to build into this model to make it work, competing against human beings.” That’s really the dissertation: the decision-making model.

Early Computer Experience

Ardito: I’m thinking about the computers ... how early were you using them?

Summit: I was actually using one in 1958, at Stanford, but to a very elemental degree, because it was the IBM 650. When I went to Lockheed, we used IBM 7090s and finally the IBM 360 series. That’s important in another way, because I got acquainted with IBM people and the IBM philosophy, which proved to be extremely useful later on in running the Dialog business. Their philosophy was one of complete support for their customers. Because computers were new at the time, managers felt insecure making decisions on buying computers and software. IBM assured their customers that if you buy an IBM machine you’re going to be supported. The stuff’s going to work, and you’re never going to be embarrassed in your professional position.

Bjørner: To actually work with your first computers, did IBM help and provide that support, or did you end up taking manuals and sitting down and learning by yourself, or did somebody teach you?

Summit: On the first computer we just had manuals … that was the class with Bob Oakford during my M.B.A. We simply sat down and figured it out and we’d cut and try until something worked — that isn’t a bad way to learn. At Lockheed, a guy named Len Fick, who subsequently came to work for Dialog, taught me all I knew about FORTRAN programming, which is what I used to program the ABES model.

Fick was in charge of telecommunications at Dialog until he retired. We had long-term employees with Dialog. We had a voluntary turnover rate on the order of 7 or 8t percent per year. In the Valley, people would tell me, “Well, gee, out here we have 20 or 30 percent turnover.” I could never understand that. We had a very stable employee group that stayed with us for a long time.

Most likely, some of my decisions at Dialog were influenced by my earlier experience with the employee satisfaction, turnover, and productivity relationship in the ABES model. I believe that when a business is successful, the employees and the customers should share in the rewards, not just the owners. I’ve always had that philosophy with our employees. I introduced the first profit-sharing program at Lockheed. Also, we had a nice building to work in. The facilities management people wanted to move us down into south San Jose. We said, “No, we want to be in the Industrial Park, because we want the employees to have a place they look forward to coming to.”

All that’s part of the psychology, but it just seemed right and fair to do things that way. Anyhow, the thing I’ve always tried to establish in a negotiation is that each party base its decisions on a principle of fairness. If the other side is proposing something, you say, “What’s really fair, in this situation?” If you come back to fairness as a negotiating guideline, then you have common grounds to do some good, rather than trying to exploit the other guy. In the end, I stayed with Lockheed and received my doctorate in 1965.

The Dialog Search Engine

Summit: Let me go back to another couple of ideas on how our work at Dialog shaped the emerging online industry. Dialog is a sophisticated search engine — much more so than the Web engines. What makes it sophisticated is that it’s recursive, that is, the output from one search process (one query) can be used as input to the next query. That way, you can compound things. None of the current Web stuff I’m aware of actually does that. They do the funnel, (i.e., narrowing down), but they don’t do the pure recursion. An analogy is driving a nail. If you can only take one whack at the nail, you have to hit it awfully hard, but if you can tap it, tap it, tap it, tap it, then you can get through anything that you want to. So, if you have a complex search problem, you can break that down into pieces and attack the pieces; you can wind up dealing with much greater complexity than if you have to totally formulate the thing that you have to say in one query.

Bjørner: That recursion is what I as a searcher really appreciate.

Summit: And that’s why we did it that way. There was a lot of resistance to that as we were doing the design, because memory was very expensive, and in order to do the recursion, you have to store all those previously retrieved sets.

Ardito: There are searchers who immediately go into Dialog and start searching. There are others who have searched for years and years, and they still literally write out their strategies beforehand. Then you have those who fall in between. With me, it depends on how complex the search is or how ambiguous the terms are going to be. Sometimes I do Select to get one very complicated set. When there are too many hits, then I break down the one set by all the different groups so that I can go and combine sets in any way.

Summit: When you do that, that’s your set 1. Then say you want to narrow your search, you can input “select set 1/ti” and you will get more relevant hits. If I’m giving a demonstration, I use /ti, to limit output to those items containing the search words in the title. In that way, the other person immediately sees why a particular item was pulled up.

Bjørner: I think we all have our different ways and we develop a pattern. But the wonderful thing that was and is with Dialog is that you have all those options. You can go back and do it again.

Summit: Yes. Thank you. This goes back to psychology. There are two things that are apparent from psychology. One is the strength of habits. The other is individual differences. We all know, but may not be cognizant, of the wide spectrum of individual differences in various approaches to problem solving. When we were designing Dialog, one group said, “Oh, well, you just index descriptors, controlled vocabulary, and that’s what you put in the index.” The other group said, “You do free text.” So I said, “We’ll do it both ways, to accommodate the variety of needs and approaches of individuals in searching.

Bjørner: We haven’t talked much about system features. Basically you said that your strategy was that if somebody else had a feature…

Summit: ...we’d copy it. But early on, the systems were fairly different in structure. I don’t remember the particular problems there were at SDC. I know that a lot of these early systems that Carlos was talking about [in Parts 1 and 2 of this series] were experimental systems working with relatively small files. There was the Cyril Cleverdon Cranfield collection of some 18,000 papers that he abstracted and indexed. This was an experimental base that a lot of people used for evaluating relevancy and recall, and although useful, it was not sufficient to prove the operability of some of the systems in an actual environment.

Ardito: Can you tell us about the OneSearch introduction?

Summit: We had many challenges in designing OneSearch. It was one of the most thoroughly reviewed features that we developed in Dialog. Libby Trudell and Sophie Hudnut had a lot to do with that. It was mocked up and tests were done. It went back and forth for paper review several times before the programming was done, to the point we felt it would interface with the actual databases. Well, it’s obvious now, how could it be anything different? But at the time, we had to go through a lot of iterations with it.

Some of the database suppliers were not overly excited with DIALINDEX, and if we offered it, they wanted to be paid a royalty for its use. We explained that it encouraged usage on the databases it identified and so was in a sense a marketing tool. In the end, most agreed to go along with it.

Another tough program was the Remove Duplicates feature. We’d been talking about that for a long time. Finally I said, “Let’s stop messing around with this.” Then one programmer, Mike Barbarino, got together with two other programmers, and came in over a weekend and put it together. Mike was a guy that did research. He was the only programmer that I remember who researched issues, and before he did the coding for Remove Duplicates, he went into the literature to see what people had done before. He was truly a software engineer, as opposed to a programmer.

Ardito: Dialog extended its service internationally very early, did it not? Tell us about that.

Summit: Yes, we installed the Dialog software at ESRO (European Space Research Agency) in 1969, making Dialog the first online network system to be used in Europe. This implementation was managed by our chief programmer, Ken Lew, assisted by the very able director of technical information at ESRO, Noel Isotta, and his chief programmer, Gerd Muhlhauser. Then by virtue of a glitch in Tymnet tarriffing in the early 1970s, European customers of Tymshare found that they could access Dialog with no telecommunications charges. The word got around. We contracted with Roger Bilboul as our agent for Europe. His establishment of the International Online Meeting in London further stimulated the online business. Dialog dominated online usage in Europe for more than 2 decades.

Japan was another interesting story. At their request, we contracted with two organizations, Maruzen and Kinokunia, in the early 1970s, becoming the first overseas service to provide information retrieval access in Japan. Tai Yamanaka, who at the time worked for Maruzen, was instrumental in assisting in the solving of the many problems that were involved in this service.

Australia soon followed in the mid-1970s.

You know, Dialog was sometimes criticized by its competitors as having slow response time, and some attributed it to the data cells we used for master file storage. The fact is that occasionally we did have slower response time than we would have liked, but it was not because of the data cells — it was because we had so many customers.

A New Professionalism

Summit: My third idea on shaping the emerging online industry is that I think online created a new profession of librarians.

Ardito: Can you say a little bit more about the “new librarians”? Do you mean that online created a different profession or a specialized segment of it?

Summit: Or was it just a tool? Online provided a tool to the research librarian. That’s a term I use a lot, as opposed to “information specialist.” I just talk about “research librarians” and people seem to understand that that’s something beyond the custodial librarian. So considering the profession of research librarianship, Dialog gave the research librarian a new tool. You could say that it’s just a new tool, that this person was doing research before, using manual tools, and then the computer comes along, so you get a computer tool, but it’s still a research librarian. It is a question of when a difference in degree becomes a difference in kind. I think that the combination of the computer and the massive databases really represents a difference in kind, in terms of what the profession can do and what the professional librarian can do.

Bjørner: In many ways, I think, because it’s more active, rather than being passive, it also opens up beyond the specialties. When you’re a librarian with books or with printed materials, you’re very often focusing on a certain subject area as an individual. But having the world of information at your fingertips makes you think in a different way, too.

Summit: Yes, and there is a skill associated with it. It’s a different skill that needs to be learned. It is an extension of other skills in terms of the interpersonal, in terms of a more expressive ability. I think we’ve created a different professionalism within librarianship. When we started seeing in job postings Dialog searching ability as one of the requirements, we thought that online had arrived. And we saw that people putting out their resumes would state experience in searching Dialog.

Keeping the Business Alive

Ardito: What were the major challenges in your work with Dialog?

One was keeping alive at Lockheed during the early days and growing the business within an aerospace company. After I retired in 1991, I was invited to give a talk to the Lockheed management group on entrepreneurship and how you can grow a business within a company like Lockheed.

Bjørner: And you showed them how to do it before then!

Summit: Yes. And we were heroes. People would get mad at us because we took over the cafeteria with our computers; we were displacing folks, and getting a lot of attention. When other managers complained we were crowding them out, my boss would indicate, “The best thing that’s ever happened to you is Dialog, because it has proven that we can put together a successful business.” I liked that.

Another time I walked into one of the offices and someone had drawn on the blackboard, kind of a World War I airplane dogfight scene, with little double wingers flying around, shooting at each other. One of them goes whoosh, and another was on the ground in flames. Each one of these airplanes had a project name on it. I was at Palo Alto Research Laboratory and there were a lot of independent research projects that were competing and shooting at each other, many of them failing. But the plane labeled Dialog was flying along and maintaining its own. I wish I had taken a picture of that.

Contract Negotiations

Summit: The other really tough aspect of work was negotiating supplier contracts. I had some very positive experiences. One example is the Commonwealth Agricultural Bureau (CAB), and its managing director at the time, Mr. Running. A British guy. We sat down to solve problems and we’d talk about how the business should be to meet the objectives they had for dissemination; i.e., getting information out to the farmers. Then we wrote a contract that was just a terms sheet. “I’ll do this, you do this; you do this, and we’ll do something,” and so on. Then we gave that to the attorneys, and they wrote up a contract and we signed it. That was fun, because it was problem solving. We had that same kind of relationship with NTIS (National Technical Information Service), with NAL (National Agricultural Library), and with several of the early suppliers.

But there were unpleasant contracts, where the philosophy seemed to be “We want to get as much from you as we can; if you die, we don’t care,” — those kinds. There are some big names that we didn’t get much support from, where it was a “you give and I’ll take” kind of thing. There was that kind of negotiation, a zero sum negotiation, rather than the attitude, “How can we expand the pie so that more customers can be served?”

Bjørner: I imagine that at the beginning of online, some of the suppliers were very concerned about erosion of print sales...

Summit: Yes. And reasonably so.

Ardito: As a matter of fact, one of the suppliers we’ve talked with said he did lose money from the corporate market as they bought into online — that the corporations would cancel print subscriptions. Academics generally kept the print, but companies would cancel or not renew because they would go on Dialog to search those as well as other databases.

Summit: Now what the publishers tend not to talk about is the new subscriptions they got by virtue of people becoming familiar with their product through Dialog. There had to be subscriptions like that. If they had analyzed their new business and asked, “Where did you learn about us?” or “Why did you subscribe?” I’m sure there would have been some off-setting subscriptions. Also I’m sure there was increased use in general, from people who had never been subscribers, never would be subscribers, but just wanted occasional access — at least they would pay some royalties.

There’s another thing that some of the print and online publishers did, which is just poor business practice. Look at what their subscription prices were over a period of time, say from 1970 to 1980, or from 1970 to 1985. Take a look at what they charged for a subscription to their printed publication and how they escalated the subscription price over that period. What they would say, of course, is “Well, you know we were losing subscriptions, so to maintain our revenue, we had to increase the price.” That’s a way to drive business away, not the best way to sustain a business. Raising prices should be the last resort, not a first resort.

Ardito: Another practice I remember is to have a two-tiered pricing scheme, with much higher prices for a file if a user did not subscribe to the print indexes. ISI (Institute for Scientific Information) and Derwent were famous for this. I think they’re still that way.

Summit: I know that Derwent did that, and two or three others.

Ardito: Do you recall your negotiations with ISI?

Summit: I do. I remember Mel Weinstock. Oftentimes in an organization where there’s kind of a tough negotiation, you have an intermediary, somebody you can talk to about problems, who can also communicate with management. Mel Weinstock was that person at ISI when we were putting the contracts together.

Ardito: Did you have any direct negotiations with Dick Harris?

Summit: I did, yes. And I don’t recall those as being overly enjoyable. Dick is a persuasive person in an argument, and he could be intimidating.

Ardito: We wanted to ask you about negotiations with Sam Wolpert at Predicasts. Predicasts was an important supplier.

Summit: It was tough. Sam was always tough, and he was unreliable. “Yeah, sure, you can have an exclusive,” he said. “Well, Sam, why did you sign up with SDC?” “Well, they’re in a different market.” “No, Sam, they’re not in a different market.” He was wily in a negotiation.

Ardito: Was it always with Sam, at Predicasts?

Summit: It was always with Sam when I was doing it, yes, and then Jeff Sharp took over in the business area of Dialog, and did the negotiations.

Ardito: We also want to ask you about negotiations concerning ABI/INFORM, because we expect to talk with Dennis Auld as a pioneer.

Summit: Greg Payne was the originator of the database and the only one that I dealt with there. He was reasonable. He was very easy to talk with. In fact, Jeff Sharp reminded me that we could have bought the whole database for about $20,000 or $30,000 at one point, and should have done it, which I agree.

Bjørner: Now, why was Greg Payne so easy?

Summit: He was a constructive problem solver. “Let’s figure out how we can get this out and get people using it.” The question is whether you’re in it for the long run, to grow the pie, or whether you’re in it for the short run, to get as big a bonus as you can this year. I think that those are the two basic attitudes. Greg Payne was in for the long run, to grow the pie.

Ardito: I was really surprised to hear how “dirty” the ABI/INFORM database was.

Summit: Yes, I vaguely remember that there were a lot of typos and that there wasn’t good control on it. But that was essentially true, to a greater or lesser extent, of all the databases in the early days. Some of the suppliers — wonder who it was, it could have been Wolpert — really didn’t like our EXPAND command because it caught all the errors. He wanted to pull out!

Bjørner: When did your friendship with Jeffery Pemberton of Online Inc. start?

Summit: I don’t know the timing exactly. We were trying to get The New York Times as a database. Jeff worked for The New York Times as a reporter, but I was dealing with his boss, whose name was John Rothman. John was a squirrelly kind of guy. He had a system that he had designed that was little more than an automated card catalog. It had very, very broad classification categories such as you might expect to find in a card catalog. Anyway, regarding the negotiation, I thought we were doing quite well, and then all of a sudden I read about a “forever” exclusive contract Rothman made with LexisNexis.

Bjørner: That was forever, indeed! [LexisNexis did not lose exclusivity of the full text of The New York Times until 1995, when the NYT allowed other services a 90-day rolling current file. To this day, LexisNexis is the only pay-as-you-go system with rights to the deep, multi-year archive.]

Summit: And it was “forever,” yes. I thought, “Well, that’s the dumbest thing I ever heard of — to do a “forever” thing. Jeff had left The New York Times by then to become independent, and we retained him to work for us to try to get that database.

Bjørner: One of my first introductions to ONLINE magazine in the early 1980s was reading Jeff’s editorial about the sale of The New York Times; it was the most scathing editorial I think I’ve ever read.

Summit: It was such a surprise. That was the thing that got Nexis on the map, I presume, just like Predicasts got us into the business area.

Ardito: What about The Wall Street Journal? Did you ever go after that?

Summit: Oh, you bet. Let’s see, what was the guy’s name? I believe it was Bill Clabby. A very brash person. He couldn’t say a sentence without several expletives in it — about one expletive per noun, as I recall. We had discussions and negotiations, but there was no way we could come to an understanding on royalties.

Newspapers at that very early time didn’t really understand what we were doing or what kind of a relationship we should have. He tried to write the contract by analogy, like a newswire service. He had some terms and conditions that, I think, had we met, he would have been willing to give it to us, but it was a totally different model than the way we were dealing with it.

Later we got abstracts of The Wall Street Journal and The New York Times and had to make do for a long time, but we had years of negotiations with both companies for the full text, just as we had with Chemical Abstracts, to get the abstracts. That’s a long and bitter story.

I can tell you a story about ISI, too. When we put up DIALORDER, Gene Garfield was livid. He had heard that we were talking with BLLD (British Library Lending Division) about becoming a DIALORDER supplier, and he went berserk. I heard from him on that, how it was illegal and he was going to bring suit, etc. He didn’t like it at all. But we were just trying to make it easier for the user. ISI could become a Dialog supplier as well. We weren’t excluding anybody. We liked competition among our suppliers.

Competition

Ardito: In the group conversation [Searcher, July/August 2003] you had a very strong reaction to BRS.

Summit: You got it from both me and Carlos Cuadra, too, did you not?

Ardito: Yes.

Summit: I don’t know. Carlos was comfortable with us in the marketplace and we were comfortable with Carlos in the marketplace, because we didn’t cut each other’s systems or each other down in a serious way. Our unstated but mutual objective was to build the community using online, without denigrating the other service, to capture some of the few customers using online services at that time.

The first thing BRS did was to come in and start cutting existing systems.

Bjørner: Cutting them how?

Summit: Price. By saying “Oh, these people are exploiting you, they’re taking you to the cleaners. We’re a nice, cooperative bunch of friends; we’re going to work with you and help you out.”

Bjørner: But isn’t it true that the reason BRS formed as a commercial company was because they lost their funding from the SUNY system?

Summit: I doubt that was the only reason. Ron Quake is an entrepreneur. Jan Egeland is an astute businessperson. She’s very articulate; she presents well.

Ardito: She was on Carlos’ list as one of the people he thought was key.

Summit: Oh, absolutely. They did several things right and were effective in competing with two established services.

I think what bothered me most was the misleading and seemingly underhanded way they did their business. They advertised a very low, fixed, base rate (maybe $10 or $15 per hour), but they didn’t indicate in large print that they tacked royalties on top of the base rate. They weren’t in the same groove, and that was probably threatening to a paranoid like me.

Telebase and EasyNet

Ardito: Was it a tough decision to go with the EasyNet gateway?

Summit: I don’t recall it being a major decision. Dialog was not reaching end users. Dick Kollin had an idea of how to reach end users through a simpler interface. It was worth trying. We did a lot of things to broaden the user base — we tried Knowledge Index and a classroom search program. We had EasyNet set up in a pretty good way. We may have even given Dick some start-up money for Telebase. We’d do this often with an advance of royalties, as publishers sometimes do with authors to help them get going.

Bjørner: You weren’t worried that EasyNet was a gateway onto other systems, as well?

Summit: We had an understanding with Dick: Dialog was on the top of the list. We took Martha Williams’ report on overall online usage and agreed that Dialog usage on EasyNet would always be in proportion to the usage represented in Martha Williams’ report.

Bjørner: In the early version, I know they didn’t let the customer chose which system they used. Later on, the customer could specify, but when they were using the real end-user version, “the EasyNet system” decided which online service a query was sent to.

Summit: Yes. Dick managed it, so I said, “I don’t want you disadvantaging us beyond what the marketplace represents,” and he agreed. So those proportions would always keep us in a favorable position.

LexisNexis

Ardito: LexisNexis, how was that relationship?

Summit: It’s a little fuzzy in my mind. I remember the early days, with Dick Giering. We were competing against them for databases, and we won every time. I think in those days we had the government databases, like NTIS; Psych Abstracts was something they got and we took it away from them.

Bjørner: You don’t have the vehemence towards LexisNexis that you do towards BRS.

Summit: No. I think they were ethical.

Ardito: Who do you think really shaped LexisNexis?

Summit: Dick Giering designed and maintained the system itself and deserves credit for inventing proximity searching of inverted files. It was clearly Jerome Rubin, however, that is usually associated with the business development. There is another name, as well. This other guy would come up to me at meetings and conferences and say, “Roger, I’m really the one who was behind LexisNexis. It wasn’t Jerry Rubin.” And I’m sure I would have checked that out to some degree at the time, and had a sense of its truth.

Ardito: You said Giering invented proximity searching, right?

Summit: Yes, and he developed the whole file organization of that early Data Corporation system. It was called Data Central and its development was supported by the Air Force.

Dialog didn’t introduce proximity searching until around 1970, and this was done at the insistence of a staff member, Mark Radwin.

The Maxwell Overture

Ardito: What about Robert Maxwell? Did you have any dealings with him?

Summit: Oh, that’s a whole chapter in another book. Yes. He was probably one of the most genial people I’ve ever met. Charming, and genial. I didn’t ever think he was other than capable and accomplished, and we spent quite a bit of time together.

Bjørner: So was it a big surprise? The downfall of his company and the suddenness with which it happened?

Summit: Yes, probably it was a surprise. He wanted to buy Dialog. After reading an unpublished and unauthorized biography of Robert Maxwell, and after talking with some ex-employees, I decided we didn’t want him in the bidding in 1988. I got the chairman of Lockheed, Larry Kitchen, to agree. I believe Larry felt a certain responsibility of stewardship for Dialog. Later, Kitchen showed me a letter from a U.S. senator, who was being considered for the position of Secretary of Defense at the time, requesting we at least speak with Mr. Maxwell.... Larry left it up to me and I agreed to a preliminary meeting. We met with Maxwell’s son, Kevin, an equally charming and astute businessman who convinced us to include Maxwell in the bidding. When the bids were opened, however, Knight-Ridder was on top.

The bidding process with Goldman Sachs and the final negotiations in 1988 were very cooperative processes between Dialog and Lockheed management. We were supportive of Lockheed’s objectives, and they treated us very fairly.

Current Activities

Bjørner: What is your current status now with Dialog, and what else are you doing?

Summit: There have been four major passions in my life: family, technology, music, and sailing, all of which I’m currently enjoying.

Ginger and I married in 1964. I had known her for a couple of years. The woman who rented me a little house in Menlo Park also rented two adjacent houses to Stanford women graduate students. What a great shopping opportunity that provided! To shorten what could be a novel-length story, I settled on Ginger, one of my next-door neighbors, and persuaded her to take a chance with me.

We have two kids, Jennifer and Scott. Jennifer is married to a Brit and is a tenured professor at Stanford in English Literature. Scott has his own industrial design business in San Francisco [http://summitid.com].

My sailing interest and love of the sea probably originated from my naval tour of duty. Ginger and I spent much of our honeymoon on my first sailboat, a 30-foot sloop sailing around the San Joaquin Delta in the Sacramento River. She has indicated that if she had realized that that trip was the prelude it turned out to be, she would have walked away, never to return.

Recently, hearing of a need for trombone players in a couple of local big band jazz orchestras, I took up my trombone after some 30 years and now enjoy playing with two of the local bands, the Daddios and the Leratones.

Bjørner: It sounds like a full life. Do you have current activities with Dialog?

Summit: I have a retainer from Dialog; it’s kind of an on-call arrangement. It isn’t a large retainer, but it keeps me in contact with them and them in contact with me. I have the official title of chairman emeritus. So that’s Dialog, and I have a non-compete clause, which is really broadly written in terms of ethics.

I’m on various boards of other organizations. I’m on the advisory board of a company called ebrary. The president and CEO of ebrary is a fellow named Chris Warnock. He’s the son of the founder of Adobe and worked on the PDF file format, and he’s just in love with literature and wants to figure out a way to make literature more available.

Ardito: How did he get to you, Roger?

Summit: Through my son. They went to school together; they’re the same age. Scott, my son, said, “Chris, you really ought to talk to my dad.” They knew about Dialog…his dad knew about Dialog. So we finally got together and they asked me to be on their board of advisors. I’m not on the board of directors.

I’m on the advisory board of Entelos, a company that does simulation modeling of the human organism, so you can test drugs on the system through the simulation model. That invitation came through Kevin Bowling, one of the Dialog employees, who went to work for them.

I was also on the financial committee for a School of Music and Arts, a not-for-profit organization encouraging music and art mainly for kids, but they have programs in schools and adult education, too. And finally, I served for 7 years as library commissioner for the town of Los Altos.

Bjørner: You are busy.

Summit: Yes, the months seem to slip by at an accelerating pace.

Ardito: Well, we didn’t talk as much as we should have about the Internet.

Summit: I think Ginger’s statement summed it up when she said to me, “You know, the Internet’s come along just in time for your retirement.”