1. At this stage in the evolution of electronic publishing, serious experimentation is needed. Models should be developed that allow for the continued expansion and enhancement of scholarly communication. Governments and others involved must avoid setting unduly restrictive rules that make such experimentation more difficult.
2. The wide availability of electronic journals and ease of access for browsing and searching is essential.
3. Publishers and librarians should collaborate to use the new medium to obtain information that allows them to improve the management of scientific publications and facilities for scientific use.
4. ICSU and UNESCO should recognize the value of broad meetings of this type for enhancing the scientific information chain. In the light of rapid changes in technology, such meetings might take place at shorter intervals than the five years between the first and second Conferences.

I hope everyone will agree that we have had a very interesting meeting made possible by the excellent talks and the enthusiastic participation in the break-out sessions. At the beginning I reviewed the last Conference and the ICSU/UNESCO activities in the intervening period and noted that much had happened in the meantime. This is amply confirmed by our discussions here. The electronic publication of journals is now well established with an ever increasing range of added value and sophistication. This has been made possible by large investment from the STM Publishers, be they commercial or learned societies, by shared services such as HighWire, and by referencing and citation initiatives such as CrossRef. We have also seen how the special properties of electronic publishing allows the visibility of and use of very large sets of data as in meteorology or in images as in brain research in ways which are quite impossible in the print medium. We have also seen how other methods of communication such as preprint servers pioneered at Los Alamos have been extended and modified in other disciplines such as Bio Science, Engineering and Chemistry.

Some of the problems associated with electronic publishing in science remain although they are more widely recognised than previously. One of these is the availability of multiple versions which need clear identification if they are not to confuse the user. The need for an authenticated peer reviewed version remains paramount. There are also ethical issues which appear with different emphasis in the electronic environment. The issue of maintaining a proper archive of electronic materials remains although serious publishers are taking care to preserve a proper archive themselves.

Whether this is adequate in the longer term remains to be seen and it is one of the issues where the changing role of the librarian is emphasised. The concerns of the librarians and the evolution of the services they provide in the electronic medium was also seriously discussed.

There remains no consensus about the most appropriate economic model for electronic journals but many are the subject of experimentation. It seems unlikely that a single specific system will evolve, certainly in the short term and there may well be a case for continuing different methods of funding. What was made clear was the need for the academic and research community to take a greater interest and hold over the dissemination process of their work and their behavior, that of their employers and the funding agencies, will hold the key to the new paradigm as it is developed. It is important for them to understand the legal framework which is being developed as traditional copyright laws are modified for the electronic environment. Some of these new laws have been unfriendly to the scientific endeavour because they are framed with different commercial imperatives in mind. This is another area where the academic and research community needs to engage more effectively in the debate.

We were fortunate in having a broad representation of colleagues from developing and transition countries who were able to show us the effects on their work of the real digital divide which undoubtedly exists. But in these areas there has already been considerable progress, both with outside help and with self help and the Conference will certainly convince everyone that these must be enhanced.

The break-out sessions engendered lively discussion about a broad range of issues and brief reports of their deliberations are given elsewhere in the Proceedings. From these discussions each of them produced a series of Recommendations which were discussed and broadly endorsed in the last session of the Conference. It is the similarity to the Recommendations of the last Conference that is striking and at first sight disappointing. But on reflection it only emphasises the rapidly evolving nature of electronic publishing in science. We are perhaps nearer the middle of the Revolution than we were five years ago but we still cannot be certain where it will lead as technology develops. In their original form as presented to the Conference a number of these Recommendations overlapped because several groups came back in their discussions to a number of key issues. The co-chairs of the Conference were therefore empowered to reorganise and rationalise these Recommendations which are now published along with the Proceedings. They are addressed broadly to all the stakeholders in the scientific information chain with some specifically slanted towards ICSU and UNESCO who will continue their work in this field in support of the scientific endeavour.

Theoretical Physics - University of Oxford
1 Keble Road, Oxford OX1 3NP
e-mail : m.barnes1@physics.ox.ac.uk
Fax : (+44) (0)1865 273947 Telephone : (+44) (0)1865 273996

The term "the best interests of science" must be interpreted from the viewpoints both of the scientists inside the enterprise and of the supporters and "users" affected by society's using the results of scientific research. This presentation will try to interpret the material presented throughout the Conference in terms of both these viewpoints, but all within the context of this presenter's perspective, as a scientist, regarding "Why should we have science?" This presentation will be an overview and summary of many ideas that emerged in discussions in the Conference, with no attempt to attribute those ideas to individuals because so many of them were said by several participants, here and elsewhere. Furthermore, the perspective here is certainly that of the scientist, not of the publisher.

Introduction

Publishing through formal channels has long had a productive symbiosis with scientific research. Publishing in established, printed journals and monographs was the most efficient, reliable way for scientists to distribute and archive the results of their work. It was a reliable, profitable way for publishers to add essential value to the products of research, so it was also profitable for professional societies and commercial publishers.

Electronic publishing changed that situation. Distribution of their results by electronic means has already become an important alternative to distribution of printed paper, for working scientists. We are in a time of drastic transition, of exploration and unease, as established methods and institutions confront challenges from new, competitive alternatives. This meeting is itself part of our way of figuring out how to make our adaptation to the opportunities and perhaps threats of new technology.

The players in this world seem, perhaps, to have very distinct roles: the researchers, the supporters of the research, the publishers and the archivists. In fact, however, there are many overlaps among these roles that muddle how we can address the problem of adapting. Scientists belong to professional organizations, some of which survive financially by earning money from their publications. Librarians used to be the archivists and distributors of scientific information; with the advent of electronic means, that role can also be taken by journals themselves and by stand-alone archives such as that at Los Alamos that Paul Ginsparg created.

To understand how we might adapt to using electronic publishing, we must examine the motivations and responsibilities of these players. First, we need to recognize the motivations of the research scientists. This is something obvious to scientists but not to almost anyone outside that community. Scientists are not, for the most part, motivated to do research in order to make money. If they were, most would be in different fields. The primary motivation for most research scientists is the desire for influence and impact on the thinking of others about the natural world, unless the desire for their own personal understanding is even stronger. But in any case, earning an adequate income from doing science is taken to be an inevitable consequence of doing at least a competent job of research, a very indirect product of doing that research. The currency of the researcher is the extent to which her or his ideas influence the thinking of others. When one recognizes this, many aspects of scientific activity can be put into a coherent economic context and interpreted therefrom. But starting with a traditional business model, based on profit-making, one is sure to get an inappropriate picture of how the scientific world functions. What this implies is that distribution of the results of research has an extremely high priority for any working scientist, apart from those whose work is behind proprietary walls. Those are the situations in which the motivation for doing the research may be superseded by the motivations of those who support the research.

This brings us to the second group of players, the supporters of research. These fall into two categories, the private, for-profit firms and the governments and not-for-profit foundations. We shall consider the latter here primarily, because the former use traditional publication for only a selected part of their work, and a part considered off the main track that motivates their work. We must ask, "Why should or does a government or not-for-profit foundation support research?" The answer, in terms useful to us here, is easy to find: it is because the research so supported produces public goods in quantities large enough to justify the investment. We recall the economists' definition of a public good: a good whose value does not diminish with its use. The results of scientific research constitute a special (but not necessarily unique) kind of public good; the more those results are used, the greater is their value. This makes especially strong the importance of assuring that the results of publicly-supported research are distributed as widely as possible. Any activity that interferes with or hinders that distribution acts against the intent and implied interests of the government or other institution that supports the research. Any activity that enhances the distribution is acting to fulfill the goals of the research supporter.

The situation is not always altogether simple, of course. The research enterprise has many components and manifestations. Some government-supported research moves into the private, proprietary sector under provisions of the Bayh-Dole act. Some research is supported by a combination of private, industrial funds and government funds, as under the Advanced Technology Program and Small Business Industrial Research program. But here, since we are examining the way electronic publishing affects science, we focus on research that scientists would, traditionally, expect to appear in the professional literature. This means research that carries some risk of failure, and an unpredictable fraction of the successful work which will yield public goods, sooner or later.

Most scientists would probably say that professional organizations are important for the maintenance of a successful research establishment. Professional scientific societies have shown a wide range of ways of dealing with and adapting to electronic publication. Most scientists would also probably say that private publishing firms have largely played a positive role in sustaining the research enterprise. How the societies and the commercial publishers - and the research communities themselves - will distribute the results of research in the future is the challenge we face in this Conference. How can these continue to work to the mutual advantage of the publishers and the researchers, both as producers and consumers?

What next? How do we adapt?

Many of the opportunities offered by electronic publication are apparent. Low-cost distribution to all parts of the world, search capabilities impossible with paper, linking, presentation of many kinds of information impossible in the fixed format of paper, all are now in use. We can probably expect people to discover and invent other new ways to use electronic publishing that we have not yet seen.

The controversies over how to adapt have focused heavily, thus far, on ownership of the intellectual property, on the rights and rules of its distribution, and on the role of certification or "refereeing". The issue of paying the costs arises as well, but, as we shall see, is properly classed as a secondary issue. The spectrum of opinion regarding ownership of scientific information has tended (but not exclusively) to follow the tradition of science publishing, in which the publisher takes ownership of copyright. This is in sharp contrast, for example, to how novels are handled; their authors typically retain the copyrights to their writing. Many people in the scientific world would now say that the most important question here, at least for the scientist, is not "Who owns the copyright?" but "Who has rights to distribute the results of research and what are those rights?" The American Physical Society has addressed this question by taking copyright but allowing the author virtually free rights to distribute the results of the research. Others, such as the American Chemical Society and the American Association for the Advancement of Science, have insisted on retaining both the copyright and the right to distribute the results in the form of publication, apart, of course, from the authors' rights to give away reprints they have purchases and to make fair-use copies of their publications.

The appearance of e-print archives such as xxx.lanl and PubMed Central has changed the situation because these are open, no-charge repositories. Each has its own rules for depositing material. The Los Alamos archive accepts papers without scientific review; PubMed Central requires refereeing. The Los Alamos archive accepts papers directly from their authors; PubMed Central takes papers from journals. Both, however, make themselves available without charge to anyone who can use the Internet. This is because they both are paid for by government funding. This is logically entirely plausible because, for the government that supports the research to achieve its goals, that government must see that the results of the research are indeed distributed as widely as possible. Furthermore the cost of distributing the results are never more than a tiny fraction of the total costs of the research; the added value from publishing the results is enormous.

Another change related to the e-print archives is the "virtual journal," that has no separate existence but is an access list of papers. The American Physical Society has begun some of these for specialized fields. Still closer to conventional journals are the all-electronic but refereed and edited journals such as the Internet Journal of Chemistry. These are examples of the new modes of publication that are still part of our experiments in adaptation.

PubMed Central is an intermediate between traditional print media translated into electronic form, as some commercial publishers now do, and open acceptance and distribution, as the Los Alamos archive does. PubMed Central distributes without charge, but its content is that of its member journals, such as Proceedings of the National Academy of Sciences of the U.S.A. This is a kind of compromise, that can assure that only reviewed works are in this archive. Of course each member journal has its own standards and procedures; there is no intent to maintain any uniform standard or threshold for admission of a paper. Journals participating in PubMedCentral may determine the length of delay between their publication date and the release of articles to PubMedCentral; this is typically a few months, at most.

The logical consequence of this reasoning, in a capitalist economy, is this: so long as there is an economically efficient private means of distributing the results of research, i.e. of "Publishing," then that can serve the communities that can use those results. If those modes of "Publishing" become inefficient, then "publishing," i.e. some other, non-traditional mode of distributing the results should come into play (to use the terminology of Martin Blume). This can take the form of distribution via an e-print archive, or of individual postings on a personal Web page, or of active transmission of manuscripts to a chosen audience. There will surely be other modes of "publishing" invented, too.

How do all these serve the ends of science? The most important characteristics that contribute positively are the wide distribution, the searchability, the linking and the negligible direct cost to the immediate user. These obviously bring a widening population into the scientific community, and enable research to be done in a more scholarly, responsible way. Unacknowledged rediscoveries will certainly happen, but their likelihood is reduced simply because the earlier literature can be searched. Even when they happen and are eventually recognized, it is likely that the merging of two or more independent recognitions of an idea will produce something more than the sum of the individual contributions. In other words, the increased transmission of existing information will make science advance faster and more effectively. Graduate students will learn of prior work that even their research directors did not know. The availability of links is a convenience that is fast becoming something we take for granted. It will become even more so as electronic publication grows.

What are the negatives for science, real or potential, associated with electronic publishing? One, in the belief of several professional societies, is that the professional societies will lose revenue and consequently lose effectiveness. Another is that commercial publishers will find scientific publication unprofitable and cease publishing their scientific journals and perhaps books. Still a third is that unrefereed articles may be irresponsible and outright wrong, and will mislead unwary readers. There of course have been substantive responses to each of these.

First, the professional societies are clearly institutions that the scientific communities want. Whether they are supported by dues, contributions, subscriptions or some other means is only a secondary question. The primary question, "Do the members of the societies and the scientific communities that use them want the professional societies enough to support them?" has an obvious answer -- "Yes!" If subscriptions cease to be an effective way for societies to get their income, then they simply will have to use other means. If they want to continue to derive their income from publication, then they might move to modes such as publication charges. Or they could increase dues or fees for conferences. The communities simply won't let their professional organizations disappear. They might have to restructure, to "downsize" and trim budgets, but the commercial world has shown that this can be done without destroying the institutions.

Second, regarding commercial publishers, each of them has a responsibility to engage in those activities that they find are economically sound. There is no reason to assume that publishing The Journal of Extraterrestrial Biomedicine can remain profitable. If commercial publication of scientific journals is made obsolete by electronic publication, then so be it. If the community of scientists wants to continue the existence of The Journal of Extraterrestrial Biomedicine and its commercial publisher decides to drop that journal, then a professional society or an informal group of scientists must act on the motivation of the relevant community and produce an electronic journal, at very least, that will carry on publication. Moreover, to the extent that the research published in that journal is supported by government funding, the supporting governments have the responsibility to see that funds are available in some form to cover the costs of publication. Journals of this kind, with very small circulations, of order a few hundred subscribers, that serve relatively small communities of scientists who, by and large, know one another, are the ideal candidates for all-electronic, unrefereed, even unedited publication. The "papers" in such communities can be made available to everyone in the community for roughly $1 per new publication, in terms of marginal cost, in contrast to $2-5,000 for a fully reviewed and edited paper in a conventional journal.

We should not expect any single publishing model to be optimal for all the sciences. We should probably not even expect a single model to be the sole mode chosen by any single field with any sizeable number of researchers. Within a field as large as chemistry, we can expect many kinds of publishing to occur, ranging from the most open e-archives to the most protective, paper-oriented and traditional journal. This will be part of the process of adaptation that we are now experiencing. We should experiment, as much as possible, see how different modes work and evaluate the consequences of the various experiments. Let the different modes compete, and let's see which stay in use and can survive.

Paying costs

Let us turn to paying the costs. The responsibility, as implied previously, for seeing that government-supported research results get published and distributed, falls ultimately to the government that supported the research. The publication costs are a tiny fraction of the costs of doing the research, so the marginal costs of publication are very small, but without publication, the results of the research cannot generate public goods, whose creation is the motivation for the support. (The question gets more complex if we examine public-private collaborations, but we avoid that issue today.) Not "Who should pay?" but "How should the payment be channeled?" is the appropriate way to address this issue. There are several possible modes, ranging from payment by author (via research grants) either as submission or page charges, through subscriptions to direct subsidies to journals. There are advantages and disadvantages to each; the glib answer is, "Choose the method that minimizes transaction costs." However there are secondary costs to some of these approaches that are difficult to evaluate, especially in the short term, so we can't expect to come to a simple answer, especially not to a single answer that will apply equally to all fields. We need to experiment with many modes, and to evaluate the consequences of each experiment.

This brings up a tricky point, particularly in the context of this UNESCO-ICSU meeting. Many journals, especially those of highest quality and esteem, now publish articles from many countries. Journals are no longer national productions. How can we expect a German author to help support a journal published in the United States? The most obvious mechanism for this is for the journal to require payment of a publication charge. This, in turn, would mean that there would have to be an agreement among nations to pay such charges. At present, scientists in several nations are not permitted to pay such fees, according to the rules of their national science agencies. In principle, the total funds required for this would be the same or less than are now required, but now much of this money goes via overhead to support library subscriptions. This overhead should decrease if library subscription costs drop as commercial journals disappear, but we know that the libraries, pinched as they have been, are likely to substitute other purchases, rather than decrease their budgets, if some of their journals vanish. This means that a realistic way to face a transition from commercial paper journals to professional electronic journals may have to be some form of funding via research grants.

Some warnings

One thing we want to avoid is any inhibitions on experiments with electronic publication. Each journal and archive should be free to make its own rules and set its own standards, but no journal or set of journals -- or publisher or legislative body -- should restrict any group or journal from carrying out its own experiment, so long as it violates no laws of plagiarism or theft. Restrictions at this time would be extremely counterproductive.

This brings us to a related issue for which some regulation is justified. As in all capitalistic situations, the market system works only if there is true competition and entry into the market is not prohibitive. At present, in the realm of scientific publication of "articles," there seems to be no particular threat of monopolization. However in the realm of publication of scientific databases, there is a very serious danger of monopolization and the loss of ready access by scientists to the data that makes their efforts possible. Some of this threat comes from the inappropriate application to scientific databases of reasoning regarding protection of privately-created databases intended for commercial use. Some of it comes from the threat of privatizing generation of scientific databases that were heretofore considered government responsibilities. Were there protection against monopoly in the database field, ways could perhaps be found to pursue some of the protective courses that have been proposed and even adopted, e.g. in the European Union's Database Directive. However without such protection, there is a very serious threat to the health of the scientific and technological enterprise. Database laws that virtually hand monopoly power to the first venturer into a database subject are likely to be severely damaging by destroying motivations for improvement or even for verification. Governments should stay in the business of distributing databases at the marginal cost of delivery whenever they have generated the data, and if they privatize the data generation, they must create a competitive market for the production and sale of the data. Otherwise, the governments have obligations to stay in the market as competitors of the private firms that would otherwise be monopolistic.

Conclusions

The most important theme we have identified is recognizing how important it is to keep open all the possibilities for experimentation and competition among various experiments. These should be experiments in openness to unlimited distribution, to forms of refereeing, to avenues for payment and pricing, and anything else we can try in adapting to electronic publication. As the results appear, we will have to take on the responsibility of evaluating and comparing the outcomes of these experiments. Each of us has her or his own guesses about which will succeed and which will evaporate, but the important thing now is to try.

Note: the file of PowerPoint slides that accompanied the presentation of this paper are available on this site.

I must admit I’ve had a few difficult slots at presentations and seminars before but I’ve never acted after the concluding remarks of the Chairman so apparently I can have no impact whatsoever on any conclusions from this conference. Some people might find that very reassuring.

I have been asked to answer the question: Is electronic publishing being used in the best interests of scientists? I am happy to do so and I would also use this opportunity to give a few personal observations on where the industry is going and where we should be going. I hope it will not be too boring for you at the end of this conference.

The answer to the question is: Not yet, but it will be. Electronic publishing is being used but not to its fullest extent yet, but it is absolutely great. It's the best thing since sliced bread that could happen to the scientific community. I can get very excited and carried away by that because it is fascinating what is already possible and, even more fascinating is what will be possible in a few years’ time. I think we have only seen the beginning but once we are further down the road in a couple of years, I honestly think that electronic publishing will allow the literature to play a different role in the whole scientific research process and that should be in the interests of everybody - developed and developing countries. I don’t want to seem obnoxious - but I do think that there is room for improvement in the research process. There is a lot of "reinventing of the wheel" still going on within countries and also in different countries. Admittedly I was not a very good researcher - that’s probably the reason I landed in this job in the first place - but it is from my recollection of my own experience that we did not check the literature every day. We did a lot of literature research, wrote about it, and at the end of the day we looked for a few references from the official literature to beef up the results. You can say "that was you and that is why you were no longer allowed at university", but I would suspect there is more of that going on. If we can develop a system where literature is at the desktop of everybody, available 24 hours a day, that should greatly enhance the role literature plays in the research process. That is the way we are going and that is the way our company is committed and you should judge all our actions and policies in that light. I am happy to explain a little bit what we are trying to do, but let me first make a remark about commercial publishing, because sometimes we have been treated not as suspects but as already convicted people. I was pleased to notice in the concluding remarks of the Chairman that this conference has taken a much more balanced view than I occasionally encounter.

The issue "yes or no" for profit is not the big thing - it is about what kind of service do we deliver to the end-user, to the customer. If we get very hung up on an almost religious tone like "Thou shalt not make a profit from science publishing", I don’t think that is very productive. I am personally not ashamed of making a profit and I would like to make more profit instead of less profit to be honest, but I would be ashamed if people were to say "you are not delivering a good service to the community" and that is what we are trying to do. To be fair, I am now in this position for two-and-a-half years, I do think there was a disconnection between the profits of some commercial publishers and the service we delivered to the community. Let me start on that note. But there is no use crying over spilled milk - just milk new milk. That is what we are doing at the moment. We are trying to enter into this bright new future and meanwhile correcting some of the issues from the past. Because the main problem from the past, in my opinion, is not the high price of a single journal subscription. That is a symptom but not the basic problem.

The basic problem is - and it is also the reason why these single items were so high - that fewer and fewer people took a subscription to the journal leaving the remaining customers to fund the whole bill of the system. That has gradually developed over many, many, many years and if you look back you would then wonder how is it possible, why was it not intervened before, but it is true. No journal started as a big journal. All journals start as small journals with thousands of subscribers. But with the explosion of scientific research, journals became bigger and bigger and the best journals became bigger more quickly than the poor journals. What happened is that prices went up, and I’ll leave aside whether the price should have gone up as much as it did, but with every price increase you lose a few people on the edges who are marginally interested. I believe that first we lost all the students, then we lost the faculty staff, then we lost the marginally interested library, so we are now down to the hard core of big libraries normally very interested libraries but there are only a few of those in every field who have to fund the cost of the whole system and that leads to take prices of thousands of dollars per journal. Now, I think that for too long commercial publishers have accepted that if fewer people pay but they pay more, then the end result is the same. To me that is a false feeling of security because the main problem is that the accessibility to literature became less and less visibility means that in general the role literature plays diminishes, the journal becomes less important, is seen less and cited less, and ultimately has to be closed down. So, in my view the main problem is the visibility of the journals with only a few people paying the costs. To take the famous example of "Brain Research" which apparently costs $15,000 per annum, there are only a few hundred big institutes paying. Beyond that there is a very inefficient system of inter-library loan, document delivery, legal and illegal photocopying, with thousands of people using it without paying anything. This is not a very acceptable situation and we should be correcting that. I would not have known how to correct that in the paper world and I probably would not have taken this job because I don’t like to start on an impossible mission: but I’m just lucky because the day I arrived electronic publishing broke through, and in my opinion electronic publishing will help us to solve all these problems in one go and do a lot more.

Electronic publishing does many things but first let me tell you what it doesn’t do. It does not lower the total cost of the system - the infrastructure costs. The fact that articles need to be written, need to be reviewed, need to be typeset and need to be archived, made accessible, open link, whatever - those costs, if anything, will increase and they will continue to increase. But what it does do is to dramatically lower the marginal costs of allowing access to, instead of 600 people, to 601 people. The extra cost of that is virtually nil and that means that we should be more creative in the business model in the future. I understand there has been some debate about differential pricing and is even one of the recommendations. In my opinion that is the way to go and we have started with it already.

In developing the way we migrate our current paper customers we find that more and more customers get together through consortia. The result normally is that the existing customers keep on paying roughly the same as they did in the paper environment, but in return they get both paper and electronic and quite often more access to more journals only electronically. But they haven’t lost out; they have gained more access. Typically, the consortium also has a few smaller players on board - small libraries that were previously very small customers to us or not customers at all - and they get the same access but at a lower price, which I think is the way to go. Also in our company we had lengthy debates - and I know that within consortia there are quite often debates - as to whether it is fair that University X has access to, say, a thousand journals and pays half a million dollars, whereas another university only pays $100,000 for exactly the same access? In my opinion that is completely justified. You could also say that the paper-pricing model was flawed. Why is it fair that a paper journal has a fixed price so that Los Alamos pays exactly the same as a university in Zimbabwe? I am not saying that we are deliberately changing one unfairness for another but let’s not be too moral about these issues. Let’s try to be practical because the problem with the paper model is that as prices went up the only people who took it, as I explained before, were the people who were very interested and normally also very big, leaving developing countries and small institutes always short-changed because although they might be interested in Brain Research they could not justify $15,000. If, in the electronic environment, we could for instance lower the price for existing customers from $15,000 to, say, $10,000 and have hundreds of small departments paying $1,000 each for unlimited access to Brain Research, I think we would all win. That is, in our opinion, the way to go. What we are basically doing is to say that you pay depending on how useful the publication is for you - estimated by how often you use it. That is something very different from paying by the drink. I do not believe that paying by the drink is a suitable model; it is cumbersome for us, but it is also cumbersome for our contract partners - the libraries and universities - and because, in science, the people who drink are not the people who are paying, it is a disastrous model as any bartender can tell you. So, we should have models where we make a deal with the university, the consortia or the whole country, where we say for this amount we will allow all your people to use our material, unlimited, 24 hours per day. And, basically the price then depends on a rough estimate of how useful is that product for you; and we can adjust it over time. It is a principle, which, in my view, is not immoral. We want to distinguish between big universities vs. small universities, corporate vs. universities, and maybe rich countries vs. developing countries. There is nothing wrong in that and any combination of the three, as long as people pay something for it, because I don’t believe in giving it away for free. There is no such thing as a free lunch, as I was always told, and it only appears free for the end-user. This is appropriate but there are costs involved in the system, which somebody has to pay that I think are better done at a certain level per university, per country, etc. because then we have the best of all worlds. We have a very efficient system, where everybody who is even vaguely interested in the material has unlimited use. What more is required? We need organizations to run such a system which could be learned societies or groups of scientists or commercial publishers. It is up to us - the commercial publishers - to prove that we can provide a good service that justifies the price we want to charge in the future. And if we can’t prove that, we will be out of business. It is as easy as that and we don’t want to be out of business too s` use I still have a mortgage to pay. So that is a good incentive to keep on working on it.

Now, enough about differential pricing. Let me say a few things on the way we see the products going. And that is what is really exciting. Taking one step back: two-and-a-half years ago when I came back to Elsevier Science, I found a company that was completely different from when I had left it 10 years before. Then it had a business model for journals - we launched 30 to 40 new journals a year and off we go. When I came back the company was in complete disarray. There was a combination of fear and excitement that we still have, but fortunately there is more excitement and less fear nowadays. There was fear because people did realize that things are changing and changing for good and for ever. That might mean that somebody could steal away our business - different players like the Microsofts of this world or the preprint servers or whatever. So there was fear among some people. But there was even more excitement because although we are a commercial company, don’t be mistaken about that, we employ thousands of scientists who are very, very committed to scientific publishing and are very committed to the scientific world. They feel themselves to be part of that community, and they saw very quickly what electronic publishing could do for their role. They saw the prospect of getting the information to the desktop of people, of adding moving images in all kinds of colours, three dimensional pictures, multimedia, etc. So there was a lot more excitement than fear but the problem was that everybody was doing their own thing. I don’t want to bore you with all organizational issues but it was a lot of hard work, I must admit - not for me but for my colleagues - to sort it all out and it provides a good raison d’être for the existence of commercial publishers. Because, if there ever was a role for commercial publishers it is now more than, say, 20 years ago. Although some people feel it has become easier because of electronic publishing, it has in fact become an awful lot more complicated and an awful lot more capital-intensive. At Elsevier Science last year we invested $30 million in electronic publishing activities. That is more than we invested between 1950 and 1980 combined. Even allowing for the fact that we are a big company, so we have probably wasted half of the money, it is still a lot of money to spend on electronic publishing. This is not a time for amateurs to get involved, with all due respect. It takes people who are committed, who are well funded and are in it for the long term. I have seen it in my own company: small publishing groups getting excited about Web servers and after three months they got bored because it is rather tedious. In electronic publishing, because you have to adhere to very strict protocols, it is not just exciting; it’s a lot of hard work. It is tedious and at the same time intellectually challenging. It is nice but it is not something on the side. It is not something that I find is done effectively by the scientists themselves. I understand their feelings that because of the high journal prices they wanted to have a go at the commercial publishers. If that is your sole motivation, I think that you are in for a shock because it is a lot of hard work and is that really what you want to do? If you really want to do that, then come and work for us. But if you want to be a scientist stay away from it because you underestimate the amount of work, the funds and also the perseverance that is needed to make it work in the long term. But I see this as a cry for attention, by people irritated by the fact that the old system was not working as well as it should. Now it is up to us - the commercial publishers and learned societies - to prove that we can develop a service that is well run and delivers what it should be delivering. Then there is no need for these initiatives from the scientific community. It is ironic that the whole world is talking about out-sourcing and the academic community would in-source a tedious job like publishing. It is much easier and more appropriate to leave it to people who do it full time and who are not clever enough for academic research and end up in publishing.

So, where do we see it going? What we have done to date is to complete the migration from the paper journals to an electronic database which contains all our journals with all articles as of 1995. That is 1.2 million articles in the database covering all the 1,200 journals under the umbrella of Science Direct. What you find is, the more you have the more extras you want. There is no limit to what people want and researchers and librarians are almost human in that respect: they also want more and more. Whereas two years ago people said that there is no need for back files, now they scream for back files because they wish they could link to the old literature. Yes, that would be nice and it will be possible. We have embarked on a big investment that at Elsevier Science within five years we want all our stuff going back to Volume I, Issue 1 of all 1,200 journals. We want also a seamless link with the non-refereed material because those distinctions are becoming blurred: preprints, refereed material, there will be one seamless flow of material from what I would like to call the "academic workbench". I would like to see a situation where, the moment you start writing an article you do it in such a format that it is constantly linked with the official and unofficial literature and where you are pointed towards stuff that is of importance to you. So when you start writing your article, after a few sentences our product should direct towards an article that might be relevant. That is making more effective use of literature instead of first finishing your article and then going on your bike to the library. The publisher should be doing that work and providing pointers to that material. So we see preprints and official journals merging into one journal and we also see the refereed material being completely merged with everything that is out there on the Web because - and I have also picked it up from this conference - everybody believes there is a future for refereed material and certainly we do that as well, but you shall not live by refereed material alone. There is more out there and what we will do in Science Direct, in the new release that will be available in March, is to allow a search through our own database of all refereed material while at the same time searching for everything that is out there on the Web through a search engine we have developed that is specifically aimed for scientific purposes, because we all know that search engines do not search the whole Web, they only search what you ask them to do. So we have asked the search engine to search the Web for scientific material and whether that is universities, author websites, etc. it will all be made available so you can then compare the research results from the refereed stuff and the unrefereed stuff. It is up to you to then choose, but we do believe that it is important that we make a distinction between those two. We are not trying to hide anything from the scientists. Let everybody have access to everything as long as it is made clear what is refereed and what is not refereed. I think that if we achieve that, we should be able to enhance the role literature plays and that is the main goal. As we said, we want the stuff to be available 24 hours a day at the desktop of everybody. We imagine a situation where everybody, the moment they get into the university or at their desk at home and switch on their computer, are immediately alerted to relevant material - derived from what they have been doing in the past - through a very sophisticated search. But it could also include mundane matters such as switching on your computer to find a first message "You might like to know that last night you were cited in this paper". The first thing you do is click on that button and look for what paper cited you. It’s fun, it’s pleasant, but at the same time you use literature more than you otherwise would have done because you look at the article and find it is interesting; so you look at a few references, and you are bound to click on one of these references and before you know it you are working with literature. So I am very optimistic about these possibilities.

You don’t have to believe everything I say - hardly anybody ever does - but I would say we believe that the future electronic world is bright. We want to be part of it and it is up to us to prove that we play a leading role. There is more choice for the community and that is good because a bit of competition keeps us on our toes. We don’t mind that at all. That is one concluding remark I would like to make. It is very, very important for you to understand that we are convinced that we can play a role but only in an open technology environment. We will make all our material linkable with anybody else who is there. We are not trying to corner the market and say: "We have a box that is only accessible to us". That is in nobody’s interest - not of us and not of the end-user. I find it ironic, and also laughable to be honest if I may say so, that I often have the impression that we are more liberal now as commercial publishers than some of these society publishers who seem to be defending not just the interests of their members but also their own interests. I find it unbelievable and beyond words that the American Chemical Society refuses to publish anything that has appeared on a preprint server that is not run by the American Chemical Society. That kind of monopolistic behaviour we wouldn’t even dare to think of. That cannot be the way to go. There is a role for everybody, and there is much more to be done. Big investment is needed but there are also bigger rewards to be gained because of the extra use that will drive it. The use of our database has gone up within a y`0 than 400%; that is incredible and there is no slowing down. If we have another year of 400% increase, I think people will use it so much that the debate on prices will go away. As a concluding remark then, prices are high per journal but the prices are not high for the whole system. Every university, as you have also learned this week, spends about only 1% of the total budget on all literature - not just ours - all literature, books and journals combined. But we also know that every library spends only a quarter of their budget on literature, and the rest is infrastructure. I am actually convinced that if we all deliver using the models that I have described, the infrastructure costs will come down dramatically. As you might have heard we have signed a deal with all Dutch universities that they have unlimited access to all the material Elsevier Science publishes - we have got a complaint from one librarian that we were destroying his document delivery business. I mean, what kind of comment is that? I am pleased to hear that the British Library last year for the first year ever, saw a decrease in the documents delivered because more and more publishers have these special arrangements (not just Elsevier Science but also Wiley Academic Press and the like) where people have more access to our material and, therefore, less need for document delivery which is sub-optimal. One last remark to illustrate this: two years ago we had about 500,000 subscriptions at Elsevier Science and we were used to the fact those numbers went down every year by a few per cent as librarians had to cancel. Now, with new contracts combining electronic and paper delivery, we have, at the end of December, more than 700,000 subscribers. This means there is no longer attrition. There are now 200,000 extra institutes who have access to our material than there was two years ago. As far as I am concerned we should not stop at 700,000; we aim for 7 million users, and this is not because I have made a big investment but because it is good for all of us. That is the way to go. Let’s grow this business. We are so used to attrition in the number of subscribers when we are in a growth business. There is still more research, there are still more researchers every day. This is a very thriving business. The only thing that is not thriving is paper publishing but electronic publishing will do that and I hope to live to see it.