Again, I'll exercise the chairman's privilege of a parenthetical. I was really surprised, James, that you managed to leave "the bug" out of your biological metaphor, because computer bugs are something all of us have dealt with.

We now come to, I think, the most interesting time of this session, where you in the audience have a chance to ask questions or make comments about the subject we have been discussing. If you make a comment, I'd ask you to keep it quite brief. If you ask a question, I'd ask the same thing. Who would like to have the first question or comment? There are microphones that will be handed out to you, I understand, so everyone can hear.

Audience member: The microphones are right here.

Bill Simpson: There actually is a very nice example that we've been doing for fifteen years of international collaboration and publication, and it's the Internet. There was the big science model, which was the International Standards Organization with large governments paying billions of dollars to develop an international network, and that was a failure. It was the competition for the Internet, which was developed bottom up.

Our publication technique was, unlike the International Standards Organization, instead of charging for the publication was to copyright not the individual contributions, but to copyright the collection, the RFC series, and to specify in the copyright that anyone else could also use or reprint the material without permission, that is, it had to be widely distributed. The individual authors also retained their own copyright so that we could then publish our contributions as chapters of books.

There is then a peer-reviewed part, so you have this large archive of material of contributions that were submitted by individuals and then you have a peer-reviewed part called the standard series, of the things that were selected by peer review for standardization. But when I heard the librarians talking about where is their place in this, and where are the peer-reviewed journals' place in this, there is an excellent opportunity for alternative selections among the very large amount of individual contributions that are now on the network to have the prestige of the journal behind their selection for their peer review. That's certainly what happened with Internet Engineering Task Force.

But the big-science model did not work at all in building the Internet. The method of funding the Internet was largely by volunteer, individual contributions for the joy of making something that actually worked. That in turn was funded by the fact that some companies could make money off of building the Internet. Now, they couldn't make money if they didn't cooperate with each other. Their parts alone were not good enough. It was the synergy of all of the companies and individuals coming together that made the Internet itself capable of being built and making money off of it. This could be a, it worked in that particular scenario as a research-funding model. This might be an alternative funding model for funding very large projects as collaborations between universities and corporations.

There was the one comment I saw from the electrical engineering person. He said, well, we could have a lottery to fund bets on how well individual corporations could do. I thought that was stock. You buy some stock, and if the company does well you make money. I don't know how that fits into how you would make stock in research, in pure research, in the replacement for the big international collaborations and going to the individual collaborations. The stock in the Internet was its ultimate workability, that it interoperated and people were able to make something together that worked. But this cuts across all the disciplines that I heard there, is that we actually have a very good example of an international collaboration that that has solved at least some of the problems that you raised.

Thomas Everhart: Thanks very much. Would any member of the panel like to comment on this lengthy comment? Yes, John.

John Ashworth: Can I just make a point on part of what you said. One of the very interesting problems that will face any legal deposit library is whether the appropriate national authority will extend the legal obligation to deposit to electronic material. And certainly the UK government has announced it will introduce such legislation. And the British Library, which views this with some horror I have to say, will find itself in principle having to accept or at least negotiate access to everything published electronically in the English language. So we will have, God help us, the corpus of material which you accurately describe.

And if you think what the historic role of a library has been, it's been to collect or try to collect all that was called knowledge at the time -- it's a very Enlightenment Age concept - and then allow other authors to mine it to produce more knowledge. Well, the electronic analogy of that is exactly as you describe. This is why I'm very intrigued with the relationship between libraries on the one hand and publishers on the other, which has always been a bit fraught in the era of the book. But in the era of electronic publishing it's going to change profoundly, and since the academic profession has always intimately been involved with both the library and the publishing profession, they will be changed by any change in that nexus.

Thomas Everhart: OK, thank you.

John Holland: Could I make one quick comment?

Thomas Everhart: No. Oh, sorry. Go ahead.

John Holland: As the EECS person, there's a large difference between "stock" and the proposal I was making. Stock has not been notably good about predicting the future, especially in terms of long-horizon allocations, and the notion of an option is a good idea and a derivative on that, but the lottery would make that market much larger, much more flexible.

Thomas Everhart: OK. On the, our left, your right, microphone please.

Audience member: Yes, hello. My name is Greg Merriman and I have a half-interest in a metasearch engine called Profusion, which won an editor's choice magazine award for most advanced metasearch in December of last year. I say that by prefix to my question. In libraries there are systems like the Dewey decimal system and in government the SUDOC system, in chemistry the periodic table of elements, etc. I see the opportunity - and I rarely have the chance to broach this subject to such a distinguished group, so I am taking this opportunity - I see the opportunity perhaps with the two decommissioned supercomputing centers recently or other distributed network resources to create a kind of dynamic matrix of concepts, based on periodicity of word usage or concept usage in the Internet itself. And since we do now have an online repository, an active repository of information to search among, and to essentially do kind of a constant sweep of, it's possible to establish that sort of periodicity or such. I'm unfortunately rather unskilled at public speaking and hence rather nervous about broaching this concept. But if any of you have any comments about the legitimacy or liability of such, I'd appreciate hearing them.

Thomas Everhart: Any member of the panel like to take that on?

Daniel Atkins: I'm not sure I fully understood your concept, but to the extent that I did, I think you started talking about classification systems, and I believe, although you didn't use the term, you were merging into the domain of metadata, of how we describe objects so that we can retrieve them. I guess you're speculating about futures in that area. I'll just make two observations that may be responsive to what you're saying.

One is that there is a trend now toward what are called structured documents, documents which HTML is one form, XML, SGML are others, in which they carry not only the content but they carry the description of the document and the semantic parts of the document, both for formatting and also for finding purposes, the author, the title, all of these various parts of the document are referenced. And to the extent to which we move to the structured document which contain themselves a lot of metadata, we have now entirely new ways of finding these things because they carry their own description, so to speak, and don't have to be a priori abstracted out and put into catalogues. So I think that's one important trend.

The other that you may have been referring to is, there is quite a bit of work under way in taking large corpi, corpuses of work, and trying to find semantic similarities between them using supercomputer type technology. Among other things, that creates potential relationships across disciplinary works that previously haven't been observed, and now of course there's this trend towards what's called data mining to kind of retrospectively extract information and knowledge, new knowledge, from existing collections.

Thomas Everhart: OK, a new question on the right.

Audience member: Yes, my name is Vlad Wielbut. I would like to ask about collaboration. Several speakers today referred to the issue of collaboration being supported and enforced by the new networks. In our thinking and speaking about collaboration, I see that we often forget, we often focus on the technological aspects, and indeed the network really supports collaboration, naturally leads itself to collaboration. And yet, what happens if people do not collaborate, if they do not want or do not have incentives to collaborate. It kind of reminds me of the comment that Bertolt Brecht once made about socialism: This is a wonderful system. We just need new people to fill it. On a somewhat comical note, it's difficult to convince French to become collaborateurs, but more seriously when you think about the role of teams in Japan, for instance, in the United States, there are profound differences of how teams are created and treated and the role in institutions. And finally, perhaps most importantly, what about the idea of attribution? If we have a team collaborating on a project, who gets the credit? I see those as very profound challenges, those cultural challenges to collaboration, and I would very much like to hear your comments on that. Thank you.

Thomas Everhart: Members of the panel like to take that? James?

James Shapiro: Yes, I'd like to say a couple things about that. First of all, I don't think we should forget the value of the individual. Often in science, when you collaborate a lot, everybody thinks the same or has to think in some kind of integrated fashion. But in science progress is often made by coming up with a radically different way of thinking about things. On the other hand, however, the Internet makes that kind of creativity easier, because an individual can have access to what everybody else is doing, have access to the databases and the libraries and so forth, in a way that an individual could never do in the past. I expect what we'll see is both more intense and productive collaborations and also a heightened activity of individuals who have now been empowered by this technology in a way that was impossible before the Internet.

Thomas Everhart: Michael wanted to say something.

Michael Gibbons: I think you raised a very important point, which touches on the line I was trying to lay in front of you about the changes in the way knowledge is produced.

We have a system in place which works extremely well for identifying and evaluating individual creativity. Everything I've heard this morning tells me, I think this is what the person who asked the question was trying to say. Everything I've heard this morning is talking about collaboration. It's not the collaboration which is based around a huge instrument. These collaborations we're talking about are essentially transient arrangements. Problem-solving complexes are things that bubble up like molasses on the stove. They bubble for a while and they disappear again, so the kind of collaboration that I think we're talking about is one that is an ongoing shift of intellectual resources, moving from area to area, problem to problem, grouping to grouping. Now, in that case, I believe the key thing to be developed by the universities is some index of a person's competence to work in a teamlike environment. I'll call that, just for the purposes of being provocative, a kind of group creativity. And, James, I think as universities we're actually appalling at developing that skill. The whole reward structure is on another line. But there is room for a reward structure to carry the rest when more of our academic is going to be spent in combinations of various kinds.

Thomas Everhart: If I could take again the chairman's prerogative, I don't think this has to be an either/or proposition. It can be a both/and proposition, and one would hate to throw away the individual creativity that springs out of a single human mind, just as one would hate to throw away the ability to collaborate with people across the world that can stimulate your thoughts and get you thinking in ways you might not think otherwise.

We'll go on to the last question.

Audience member: Yes, I'd like to just very briefly address the subject of the richness of human intellectual diversity, a richness which is I think not very well represented by the present all-male panel but at least -

(Applause)

Thomas Everhart: The chairman did not choose the panel.

Audience member: At least Michael Gibbons and John Ashworth have truly global connections. My name is Linn Hobbs (name uncertain) and I'm a material scientist, and I'm about to go next week on a workshop that the NSF is sponsoring in Rio on promoting collaborative materials research between North and South America. It's interesting that materials is a subject along with information technology and biotechnology, two things that have been addressed today, which is likely to have the largest impact on life in the twenty-first century.

The Web and the collaboratories that have been talked about I think have the potential to widely increase the access to the knowledge base, to instruments, to sources of research funding, to what the economists still euphemistically refer to as "the developing world." And I'd like to ask the panelists if they might want to think about or comment on what are the implications of the globalization of the academy and utilization of a far more diverse source, database, or a base of human potential than we have perhaps had before on what I think can be achieved in research and in the workings of the academy.

Thomas Everhart: Thank you. Who would like to take that?

Daniel Atkins: I'll make two brief comments. First, one of the most exciting aspects of the collaboratory potential I see is this issue of access. We already have instances of faculty who choose, for personal reasons, to go to predominantly undergraduate universities but now remain richly connected to first-tier research communities, and they're also using collaboratories to enrich undergraduate teaching.

The question of emerging countries is one of great personal interest to me, and we're working with both the Mellon and Kellogg foundations in leap-frogging strategies in South Africa and parts of Latin America. The good news is that many of these major foundations are now starting to see technology as of strategic importance to their agendas. They want to make major investments in leap-frogging, so for example, in digital libraries, to allow these places to have kind of instantaneous access to collections they otherwise would not have. And possible exploration of extending educational opportunities through distance-independent technology. So I think it all has perhaps quite profound implications for broader participation on a world scale.

Thomas Everhart: I'll let John Ashworth have next-to-the-last word.

John Ashworth: Thank you. Just one point of perhaps discordant comment. There is a very real problem in the kinds of information that we've been talking about. It is by definition explicit, and every research study that I'm aware of on the sociology of the exploitation of research results tells us that a key component in whether or not the explicit knowledge obtained is used effectively within any cultural or social environment is the tacit knowledge which that environment has. I think the truth is the Internet gives enormous advantages to English-speaking people who have the kind of tacit knowledge that goes with the acquisition of English at a very early age, and there is a serious problem for those who don't have the tacit knowledge that, dare I say it, even the English and the Americans share, and I know we are different cultures and I know there are problems, but at least we are closer and linguistically more comfortable with each other than say Chinese or Indonesian or other cultures. And I think one of the biggest cultural and social challenges for those who are developing the Internet is going to revolve around the conversion of tacit knowledge to explicit knowledge and the distribution of that through electronic media.

Thomas Everhart: I'd like to bring this panel to a close. I'd like to thank all of the panel members for their contributions, for the preparation they put into it, as well as what they said here, and I'd like to thank you, our audience, for being here. We're adjourned, just a few moments late until, I think, eleven o'clock when the next panel takes place. Thank you.