BY DAWN LEVY
Publish or perish. That's the credo by which scientists live. But in the world of high-energy physics (HEP), where the acceptance rate of papers submitted to journals is about 70 percent, that credo has gone the way of the dodo bird and eight-track tapes.
At least as far as publishing is concerned, no one is perishing. Instead of the old method of writing to a researcher to get a reprint of his or her journal article, HEP scientists are using the World Wide Web to access e-prints preprint articles that let physicists share results faster than they could through the journals in which they ultimately publish.
"The physics community had a really rapid adoption of this because in a sense it was just an evolutionary process rather than a revolutionary one," said Heath O'Connell, HEP database manager for the Stanford Linear Accelerator Center (SLAC). O'Connell spoke about the impact of self-publishing on information management during a Feb. 21 session on paperless publishing at the annual meeting of the American Association for the Advancement of Science in Washington, D.C.
Back in the 1960s, SLAC librarians started collecting submitted preprints, and each week they would create a list of all of them, which they would mail around the world. Subscribers could then write to authors of the papers that interested them and request reprints.
In 1991, physicist Paul Ginsparg set up an electronic system at Los Alamos National Laboratory in New Mexico that would allow authors from around the world to send papers on physics, mathematics, computer science and other topics to a central repository. In most cases the papers had not undergone peer review, but the archive had checks to ensure that most submissions were of professional quality. For example, the software checked to see if the author was affiliated with a university or other credible source. Scientists sent their preprints to Los Alamos at the same time they sent them to the journals.
Each day the Los Alamos archive system e-mailed information about these papers to about 2,000 subscribers. A subscriber interested in a particular paper could e-mail Los Alamos to have an electronic copy of the paper sent via e-mail.
"SLAC had already been doing something similar, but the great thing about the Los Alamos system was it was all done electronically," O'Connell said. "The beauty of the electronic part was it enabled the entire paper to be sent as an electronic file."
The submission rate to print journals is just as high as it ever was, but paperless publishing may be increasing the overall quality of journal articles, as scientists can incorporate peer feedback before articles are set in immutable ink.
And e-prints are timely. Says O'Connell: "You've done the work. You don't want to sit on it any longer. You want it out. You want it communicated as quickly as possible. And it could be six months to, in some extreme cases, two years before it appears in a journal."
The part of the SLAC system on which the Los Alamos electronic archive of journal papers was modeled was the SPIRES-HEP database (for Stanford Public Information Retrieval System High Energy Physics). Every day SLAC downloads information from Los Alamos's HEP section into its SPIRES database, and every week SPIRES sends publication notes for e-prints to Los Alamos.
"The SPIRES database offers a lot of value-added services, such as reference and citation linking, which are very important because people want to see how many times their work's being cited," O'Connell said. "It's a measure of professional accomplishment. Getting more citations is the research equivalent of earning more money. You know how important your work is, how many people have stood on your shoulders."
Users can reference e-prints just as they would print articles. Ever since the inception of the Los Alamos archives in 1991, each article has been assigned a number that is a unique specifier. To submit a paper to the journal Physical Review for consideration, the author need only inform the journal of the Los Alamos e-print number.
The SLAC database is to physicists what the National Library of Medicine's Medline database is to doctors only beefier. More than a virtual library, the database contains more than 400,000 records of preprints, journal articles, reports, conference papers and theses from 1974 to the present. It provides links to more than 475 HEP experiments worldwide. It also gives links to the full text of e-prints stored at Los Alamos, as well as print articles stored on file servers of physics journal publishers, such as the American Physical Society and Elsevier, and physics laboratories worldwide, such as KEK in Japan and DESY in Germany.
Duplicates of the archive exist on servers in Germany, England and Japan. "The more places you have exact copies of all your data, the less it matters if, say, a machine blows up and it all gets lost, because you've got somewhere else where it all gets stored," O'Connell said.
So far physicists have been the biggest benefactors of paperless publishing because they have a long history of sending out their preprints as paper. SLAC's management of the database saves scientists time. "Instead of having to send out your paper to every reader in the world, you could just send it to SLAC and then SLAC would send a list of all the preprints it received," O'Connell explained. "Now Los Alamos National Laboratory has made this effort of managing the papers themselves even more rapid and efficient."
The physicists use a special typesetting program called LaTeX (pronounced lah-teck) to convey equations without the loss of formatting that can occur when files are compressed and e-mailed. "This is a language which allows you to just type in text characters and then process the file, and it will come up on your computer screen as appropriately rendered mathematics," O'Connell said. "Everyone knows how to use this, and every department has its own LaTeX processor."
Although mathematicians have expressed interest in a similar database to openly share information, other scientific disciplines may not be ready for paperless publishing.
"In some of the other fields it's been quite a battle," O'Connell admitted. "It just hasn't caught on. I think in large part you can really thank LaTeX for giving us a good system to use that would enable this to all be done so simply. Also, maybe the extremely noncommercial nature of high-energy physics contributes it's not like biology and medicine where maybe your discoveries could earn you a lot of money and you might indeed want to keep them hush-hush for some period of time. In physics the openness is really what it's all about."
While other communities worry about the so-called "digital divide" between technology "haves" and "have nots," the physics community is using technology to make the discipline more inclusive. "It's useful to countries that perhaps aren't as rich as some of the others. Now all of a sudden they have access the same access that anybody else does to research done all over the world," O'Connell said. "So long as you have a computer, which is getting to be not such an expenditure, you can keep up with everyone else. I think it's really doing something to bring this community together all over the world. Even in the days of paper preprints this wasn't always possible." SR