Editorials

Electronic publishing in science

The revolution is only just beginning

The five years between two Paris conferences on electronic publishing in science have seen dramatic changes. Almost all scientists now use the internet routinely. Most journals have produced an electronic version, and many are moving beyond simply making the paper version available through the web. Disciplines in addition to physics have created eprint servers (where authors can make their research open to everybody as soon as it is completed), and many new beastslike PubMed Central and E-Bioscihave begun to stir in the information jungle. Nevertheless, most of those at last month's conference organised by Unesco and the International Council of Scientific Unions thought that we are still at the beginning of the electronic revolution in scientific publishing. The next five years will see greater change.

Electronic publishing opens up the prospect of all scientists having access to all research from their desktops. Access has improved for some, but not for alland for those in the developing world it has got worse. Everybody at the conference, including commercial publishers, agreed that it is good for science, scientists, governments, and the public for access to the results of scientific research to be unfettered. Unlike most physical commodities, the value of information increases as more people have accessbecause they can use the information in their own environments and add new insights, increasing the value for everybody.

Journals, which have since the 17th century been the main means of disseminating research, have been important because they provided some quality control, drew together research papers on particular subjects, and organised the distribution of research. But they also Balkanised the research, dividing it up into many different journals, most of which charge for access. Thus the research that matters to gastroenterologists, for example, might be scattered across 50 journals. It is difficult, time consuming, and expensive to access that research, and most gastroenterologists don't have the time, skills, or resources necessary. They hope that somehow what matters will be brought to their attention, but this Balkanisation of the evidence may explain some of the gap between the evidence and practice.

The price of access to scientific research has been increasing for years. Since the 1970s libraries have been cancelling subscriptions to journals (often to pay for subscriptions to new journals), and publishers have responded by raising prices by well above inflation. The Association of Research Libraries reported a 207% increase in the price of journal subscriptions between 1986 and 1999, during which time the number of journals increased by 55%. The result is a 6% reduction in the number of journals to which American research libraries (the richest in the world) subscribe and an even greater reduction in the proportion of journals that libraries provide for their users. In short, libraries are paying more for less.

Where has the money gone? One answer is to the big commercial publishers, who have such market power because some of their journals are prestigious journals in which academics have to publish to gain credit and which libraries have to buy (an annual subscription to Brain Research is famously $15 000). In 1997 the average net profit margin of the top four commercial publishers was nearly 19%a high margin. This exploitation has caused resentment in the academic community, particularly in the United States. But developing countries have been hardest hit by these price increases, and the library shelves of many developing countries are now bare.

There are other problems with journals. Increasingly research cannot be easily presented in a traditional printed article. Research in physics, chemistry, astronomy, and biology is increasingly concerned with huge databases, and it'll happen soon in medicine. The results of such research cannot be published on paper, and something more than the traditional journal article is needed. Many scientists are also keen to use the multimedia possibilities of electronic publishing.

William Shulenberger, provost of the University of Kansas, summarised the "solutions" to the problems of access to scientific information but pointed out that there is as yet no new system. Increasing library budgets is doomed to failure: they can never keep up. Preprint servers have not so far reduced the demand for journals. Open archives (for example, on university websites) have yet to have much impact, but they might in the longer run, particularly if combined with search engines that point readers to the most visited sites. What Shulenberger called "minimal refereeing services"like PubMed Central or BioMed Centralare just beginning but may not get far because some scientific communities, particularly medicine, are nervous of minimal refereeing.

Attempts by academic communities to create alternatives to journals owned by publishers have not so far been successful. But Hector Rubinstein, professor of theoretical physics in Stockholm, described how high energy physicists have switched from publishing in an expensive journal to publishing in the electronic Journal of High Energy Physics, which provides free access. Another "solution" is antitrust actions against publishers (the takeover of Harcourt General by Reed-Elsevier has been referred by the UK government to its Monopolies and Mergers Commission), but no publisher owns enough of the market for this to be effective in reducing prices and increasing access. The academic community has also discussed uncoupling publication from academic credit and creating buying cooperatives but made little progress. An authors' boycott of publishers who will not make their material free within six months of publication is currently signing up 250 people a day, but its effectiveness remains to be seen.

Most of these initiatives are aimed at clawing back from publishers the value they extract from the system, but some, particularly E-Biosci, are also about trying to move beyond journals and articlesfor example, by using many media and providing access to databases. Publishers may yet secure a profitable future not by unimaginatively increasing the price of journals but by doing the job of electronic publishing more cost effectively than the academic community. The fact that the cost of providing electronic access to one more user is close to zero opens up new pricing models that might mean that publishers could allow greater access for the same income. Derk Haank, the chief executive officer of Elsevier Science, was confident that publishers could increase access and retain profits.

The market may solve the problem of access to scientific research, but for the past 30 years it has been making it worse. One problem has been that publishers have been able to charge not simply for the value they add in publishing research but also for the value of the research itself. The academic community has failed to mount a coordinated response, and researchers have carried on publishing in expensive journals, leaving librarians to meet the costs. Now academics are trying to work together to achieve affordable access to research. In the absence of a system, US academics have produced a set of principles for emerging systems of scholarly publishing (see box). In a time of great change it is wise to return to first principles, and publishers may still be able to make profits while observing the principles. More probably, however, new methods of disseminating research will appear and the number of journals will diminish dramatically, with those that survive looking more like Cosmopolitan and less like Brain.

Richard Smith, editor. 

BMJ

Acknowledgments

RS is also chief executive of the BMJ Publishing Group, a medium sized medical publisher owned by the British Medical Association.