EVERY day thousands of scientists visit the office of Paul Ginsparg, a theoretical physicist at the Los Alamos National Laboratory in New Mexico. Fortunately for Ginsparg, most of his visitors don鈥檛 make personal calls. Instead they talk to a small computer that sits under his desk, eager to find out about the latest developments in areas as diverse as economics and high-energy particle physics.
But these electronic dialogues are not just the idle banter typical of many exchanges on the Internet. They are symptomatic of what may become one of the biggest shake-ups science has ever seen.
Since the time of Isaac Newton, scientific research has not been taken seriously until it has been through a process known as peer review. This involves submission of new research findings, in the form of a written document, to a specialist journal. The paper is then sent to two or more experts in the field who advise the editor of the journal on the research鈥檚 fitness for publication. Results can take months, even years, to negotiate the obstacles of this refereeing process and reach a wider audience. Run into a stubborn referee, and your brilliant breakthrough may never see the light of day.
Advertisement
Post your results on Ginsparg鈥檚 computer, however, and you have a potential audience of millions within an hour. Anyone can contribute, and anyone can call up the end result. Ginsparg鈥檚 computer can let you see science advancing almost before your eyes. You simply call up the World Wide Web home page on Ginsparg鈥檚 computer, and point the cursor at the subject of your choice. The hypertext links built into the system will then transfer you to another part of the computer鈥檚 directory where that subject is stored. A couple more clicks of the mouse, and you are looking at the text of a pre-print 鈥� final pre-submission drafts of papers 鈥� submitted perhaps just hours before.
Ginsparg鈥檚 electronic publication service started out in 1991 as merely an electronic clearing house for pre-prints in high-energy physics. Now Ginsparg鈥檚 computer provides access to the latest findings in 20 disciplines. By March this year, the idea had generated so much interest that the National Science foundation in the US awarded Ginsparg $1 million to turn his cottage industry into an official service with full-time staff. Ginsparg himself is far from surprised. 鈥淢y contention has long been that the way referencing is handled is entirely inept,鈥� he says. 鈥淚t only results in pointless delays benefiting no one.鈥�
To some, abandonment of the conventional peer review process, with its stuffy anonymous referees settling old scores or burying rival research, may seem long overdue. Science would surely benefit from shaking off this intellectual strait-jacket and opening up the research and refereeing process.
But according to Stevan Harnad, professor of psychology at Southampton University and editor of the electronic journal Psycoloquy, such talk is not only naive, but potentially fatal for science. Without formal refereeing, he warns, scientists could find themselves drowning in a sea of unreliable research.
So far, Ginsparg鈥檚 clearing house has maintained a high standard. But this is misleading, says Harnad. 鈥淭he reason is that those pre-prints are all destined to appear in rigorously refereed paper journals 鈥� it鈥檚 the invisible hand of anonymous peer review you鈥檙e seeing in action.鈥�
Should the traditional paper journals backing up this system disappear, Harnad warns, the quality of postings to the electronic pre-print service would plummet to the abysmal level of the average Internet bulletin board. 鈥淪ome people just presume that others will simply make sacrifices [of time] and do refereeing as they do now, but you cannot get the Net to do refereeing for you,鈥� says Harnad.
It is this coexistence of the old and new that makes the current success of Ginsparg鈥檚 innovation so misleading 鈥� and what happens next so crucial. For scientists wanting to acquire credibility today, there is no other outlet but respectable paper journals. To some, this is one of the worst aspects of the present system. Brilliant but unorthodox papers can be rejected for years, a fate that famously befell the work of American biologist Lynn Margulis on the origin of eucaryotic cells in the late 1960s, and the American theorist Mitchell Feigenbaum鈥檚 seminal work on chaos theory ten years later. Meanwhile ho-hum research goes through on the nod: most scientific papers are never cited again.
Yet academics need to contribute quality research to respectable journals as often as possible to win funding.
Electronic publishing is not so cut and dried. Advocates say it offers the possibility of a 鈥減ublishing continuum鈥�, encompassing everything from sparky ideas in pre-print format, all the way up to full papers in rigorously refereed online journals. All these could make (or break) a scientific reputation.
The key question is how to combine the benefits of electronic publishing with the high standards of conventional peer review. One leading advocate of electronic journals, mathematician Andrew Odlyzko of AT&T Bell Laboratories, has a vision on how this might be achieved. In the future, an electronic journal would consist of, say, a World Wide Web address to which researchers would post their findings. Each journal would have a permanent editor with a broad knowledge of the field who would vet the postings. That would prevent the system becoming clogged with wacky proofs that the Earth is flat. In addition, there would then be a strict 鈥渘o withdrawal鈥� policy: once posted on the system, everything remains there 鈥� no matter how embarrassing. That, says Odlyzko, should also make serious contributors think twice before posting off-the-cuff ideas.
The computer would then time-stamp each contribution, to deal with disputes about who submitted what and when. After that, the contribution would be available to everyone else in the field for debate and comment.
Odlyzko believes that to maximise the number of people willing to comment, both attributed and anonymous comments should be allowed. The cloak of anonymity would protect those who might otherwise feel embarrassed about pointing out glaring errors in, say, a colleague鈥檚 submission.
Once a reasonable number of comments had come in, authors would then be allowed to submit revisions: 鈥淎 research paper would be a living document, evolving as new comments and revisions were added,鈥� says Odlyzko. 鈥淭his process could continue indefinitely, even a hundred years after the initial submission.鈥�
This two-way process, he believes, would go a long way toward ensuring the reliability of submissions. But, like Harnad, Odlyzko believes that there will still have to be a formal refereeing process to ensure uniform standards. The comments received would help in selecting referees, and pointing them towards flaws in a submission. Yet in the end, says Odlyzko, they are not enough: most researchers will still only feel happiest quoting contributions that have been formally refereed as well. It is quality not quantity that researchers want from the academic publishing media of the future.
The issue of maintaining high academic quality may give today鈥檚 paper publishers a way to survive in the electronic era. At present, traditional academic publishers think electronic journals are just paper journals stuck on, say, the World Wide Web with perhaps a hypertext link to the author鈥檚 biography.
But such 鈥渋nnovations鈥� are little more than tinkering, according to advocates of full-blown electronic publishing like Harnad. They argue that traditional publishers will have to be far more radical in their approach to keep up with the electronic revolution.
Why pay?
At present, Harnad points out, publishers of traditional journals typically pay neither their contributors nor their referees. Instead, they promise to look after the dissemination of research, and make their profit by levying huge subscription costs of hundreds of pounds for perhaps just 12 issues.
But, says Harnad, electronic publishing makes dissemination virtually free, universal and instantaneous. So why would anyone pay for something they can get free of charge elsewhere? The only benefit that paper publications still offer is of academic quality, underwritten by the refereeing process.
So traditional publishers, says Harnad, can only keep a niche in the electronic era by guarding the quality of research rather than merely disseminating it. He proposes that publishers would still employ editors to organise refereeing, and the referees would not be paid. But publishers would not make their profits from subscriptions, but by charging authors for the right to appear in refereed electronic journals.
鈥淎t present, academics generally get nothing for refereeing and pay nothing to appear in print, but the cost of publishing is huge, and so are the subscription charges,鈥� points out Harnad. 鈥淚n the future, authors will simply pay a little to be published in a prestigious refereed electronic journal, which will then be made available to everyone at no cost.鈥�
Whether this radical new approach to academic publishing will be economically viable depends on whose analysis of the cost of electronic publishing you believe. 鈥淧aper publishers usually say that there鈥檚 a saving of about 20 to 30 per cent to be gained by going electronic,鈥� says Harnad. 鈥淏ut my estimate is at least a 70 per cent saving. Publishers think that electronic publishing is just an extra perk added to paper and possibly saleable on the subscription model with the same old infrastructure.鈥�
With production and distribution costs slashed, says Harnad, subscriptions to electronic journals really could be free, the whole enterprise being funded primarily by the page charges levied on authors. The end result sounds positively utopian. Academics would have a greater range of ways to contribute to research, from provocative pre-prints to announcements of fundamental discoveries. Quality would be maintained by peer-group pressure, informal comments and formal refereeing. All levels of contribution could then count towards one鈥檚 academic reputation. The electronic format would also allow archives to be searched anytime, anywhere, and would take up a tiny fraction of the space of today鈥檚 paper literature.
No rubbish
Best of all, the results of research would be available far faster than today, and to virtually anyone who wants them, from an undergraduate in one of the developing world鈥檚 universities to a Nobel prizewinner in the US.
What happens next depends on the issue of academic quality. Some traditional publishers already sense this: 鈥淭he key concern for us and the research community is how to prevent ourselves being overwhelmed by rubbish,鈥� says Mark Ware, publishing director of Institute of Physics Publishing. 鈥淲e have to persuade people to insist on refereed formats, and not on pre-prints.鈥�
So far, however, most publishers seem to be adopting a 鈥渨ait and see鈥� attitude, toying with essentially irrelevant electronic publishing ventures like back issues on CD-ROM. Yet according to the electronic prophets, traditional publishers must recognise the importance of academic quality in the electronic era 鈥� or risk becoming mere camp followers to the March of Science.