December 18, 1995
Electronic science journals
Paperless papers: cyber-publishing
IT IS hard to recall in these days of cyber-everything, but the Internet was originally created as a place for scientists to do science. This accident of history is so easily overlooked because researchers, who just three years ago were the net's main inhabitants, seem to have disappeared from it.
An illusion. Scientists are still among the most wired people on the planet. It is just that, like shy woodland creatures faced with galumphing ramblers, they have gone underground_to private discussion groups and hideaways shielded from the mindless chatter that now fills the Internet's glades. But gone is not forgotten. The net's begetters are busy, and they have a project. They are using their creation to shake up their most sacred activity: publishing.
Publish or perish. So goes the maxim that strikes dread into the heart of every untenured academic. And publishers have grown rich on it. Around half a dozen of the biggest have made millions by slicing scientific disciplines ever thinner. They print volumes of incomprehensible papers for a few hundred readers and charge libraries thousands of dollars a year for the privilege of receiving them. And, to add insult to injury, the authors, too, often have to cough up, paying by the page for their pearls of wisdom to be printed, lest they be lost to the world_and to the appointments committees.
The fact that scientists are willing to lie back and enjoy this process does not make it any less rapacious. That is why they have taken so eagerly to the Internet as an alternative to the printed page. The first to break away were the physicists. Many of the original builders of the Internet came from among their ranks. They also have access to the whizziest Internet kit.
They used it first to reduce the delays in scientific publishing. From the day an article is submitted to a traditional journal to the day it reaches a subscriber's hands can amount to months of peer-review, editing, page approval and simple queueing for space. So, by the time an article is published, it is usually old news. ``Preprints'', which circulate months ahead of publication, are the real currency of physics.
Researchers at the Los Alamos National Laboratory in New Mexico took this a step further. In 1991 they set up a service to store and distribute electronic copies of preprints (or ``e-prints'') over the Internet_not only because it was faster, but also because it was more democratic. No more missing the latest breakthrough because you were not deemed worthy of being on the distribution list. Since then, their e-print service has grown to handle 40,000 requests a day from 20,000 researchers. As a result, the physics preprint is dying and the paper journals themselves have become mere archives.
The obvious next step is to bypass paper altogether. Instead, researchers submit their articles by e-mail, editors e-mail them to referees, and the final, revised versions are ``published'' in an electronic journal. The result: turnaround times measured in weeks rather than months, and subscription prices a fraction of paper_usually less than $100 per year and sometimes free. A study earlier this year by Philip McEldowney, at the University of Virginia in Charlottesville, found 306 electronic journals, 70% more than the number available a year before (see chart). And along with dozens in mathematics, physics, chemistry, biology and medicine, there are now more than 80 in the social sciences.
Most of these journals are run by volunteers, are government-funded, or are spin-offs from paper-based commercial ventures. But at least one traditional publisher is getting ready to dive into electronic publishing as a full-fledged business. In February next year, Current Science, a publishing group based in London and Cambridge, Massachusetts, will officially launch BioMedNet, a E5m-6m ($7m-9m) effort to create the Internet equivalent of a library, conference centre, shop and general scientific hang-out.
BioMedNet, which already has about 6,000 scientists using an early version of the service, bills itself as a ``club'', for lack of a better term. Its heart is a library of journals_mostly electronic versions of well-respected traditional publications. Today there are nearly 40 of them. By February there could be 200, thanks to agreements with 20 publishers.
The value of this, besides the ability to search all the journals and retrieve any article at the click of a mouse, comes in the way they will be linked. The company has several dozen programmers beavering away at the necessary software. Their aim is to ensure that when a paper refers to another paper, a click on the reference will bring up the second paper, along with a list of all other papers that cite it. In this way the interdependent web of the scientific world reveals itself: each discovery built on the bricks of hundreds before it, all now connected by a trail of mouse clickings.
Where BioMedNet hopes to break new ground is in building a ``virtual community'' outside this library. One model for this is an experiment at the Weizmann Institute in Israel. Known as BioMOO, for obscure historical reasons, it appears as a ``place'' on the Internet. It is somewhere where biologists can (through their PCs) meet, share data, work on papers together, attend lectures, and display their research. All this takes a good bit of imagination, as such socialising consists mostly of typed conversation and text descriptions. But the idea of a global scientific common room, where one can expect to meet like-minded researchers from anywhere in the world, is compelling enough to lure hundreds of scientists to their keyboards each day.
BioMedNet aims to adopt much of this, adding better graphics and more shared ``tools'' such as data-analysis and statistics programs. Because it will be a commercial service, with various ways to charge users, it could attract companies to invest more on-line, bringing professional qualities to what has so far been a mostly amateur effort. It plans to host on-line scientific meetings in shared ``rooms'', where lecturers can display data and answer typed questions live. Eventually, it sees itself as a data archive too. It is building a database on genetically altered mice. Similar services for chemistry, physics and perhaps even non-scientific subjects such as travel are in the pipeline.
Whether or not BioMedNet turns out to have the winning formula, no one doubts that science will increasingly be conducted on-line. E-mail has been the primary form of collaboration for several years now. In some disciplines e-journals are fast becoming the favoured means of communication. Traditional scientific publishers are already feeling the heat. Earlier this month Forbes, a business magazine, argued that Reed-Elsevier, an Anglo-Dutch publisher with a lot of expensive and obscure paper journals, could become ``the Internet's first victim''. Its shares dropped nearly 10% overnight.