Physicist Paul Ginsparg, who created and maintains the archive -- known by scientists around the world as "arXiv.org" -- will join the Cornell faculty this fall, and he is bringing the archive with him.

It will become a service of Cornell University Library, which has developed several other digital academic resources. Both Ginsparg and library officials express hope that the archive will improve and expand in its new home.

The archive currently is receiving about 2 million visits a week, more than two-thirds of them from outside the United States.

"There should be many advantages to being at a private educational institution," Ginsparg said. However, he noted, the LANL environment was essential for launching the archive in 1991. "It probably wouldn't have been possible had I been a university faculty member with too many other obligations. But now it has achieved a level of maturity which makes it possible to institutionalize in a new and more appropriate academic setting," he said.

The arXiv has operated with about $300,000 in annual funding from the National Science Foundation, the Department of Energy and LANL. For the time being, Cornell and LANL will share the costs and services previously provided by LANL.

The arXiv will remain a cooperative effort between LANL and Cornell, since much key expertise will remain at the LANL library, Ginsparg said. The existing LANL server will become a primary backup.

Ginsparg already has been collaborating with the Digital Library Group in Cornell's computer science department from a distance. He will become a member of Cornell's Faculty of Computing and Information (FCI), a universitywide, interdisciplinary unit, separate from but related to the computer science department.

The FCI was formed last year in recognition of the fact that computing has become an integral part of almost every academic discipline. Ginsparg, who earned his Ph.D. in physics at Cornell in 1981, expects to divide his time equally between work on the archive and physics research.

His field is string theory, the so-called "theory of everything" that aims to unify all of the forces of nature. "I am eagerly looking forward to having all the routine aspects of the arXiv handled by information professionals so that I can focus again on cutting-edge areas in research," he said.

Ginsparg was named a fellow of the American Physical Society in November 2000 "for his work relating to chiral symmetry on the lattice, for fundamental contributions to string theory and for establishment and development of the revolutionary Los Alamos E-Print Archive."

"I think ultimately it will be seen as a coup for the university to have attracted him and the archive," said Sarah Thomas, Cornell University librarian. "It's a captivating example of how technology has interacted with the advancement of knowledge." But she added that she understands some people are apprehensive about the impact online publishing of scientific information may have on traditional journals.

The arXiv contains some 170,000 brief papers in physics, mathematics and computer science, with almost 3,000 new submissions coming in each month. Unlike articles submitted to professional journals, papers submitted to the archive are immediately available online, at no cost to the user. Also unlike articles submitted to professional journals, postings to arXiv.org are not peer-reviewed. Except for some rudimentary screening for inappropriate off-topic submissions, almost anyone can post almost anything. It's up to the reader to decide what is worthwhile.

The result, Ginsparg has said, is to "level the playing field." Researchers in Third World countries, where paper copies of journals may arrive months after publication if at all, have the same access to research reports as do researchers in industrialized nations.

On the other side of the coin, researchers in small, obscure places have just as much chance to make their voices heard as those in Ivy League halls. In one recent incident, Lubos Motl, an undergraduate physics student at Charles University in Prague, Czechoslovakia, scooped the Ph.D.s with an elegant solution to a major problem. On the Internet, it seems, no one knows you're an undergraduate.

Ginsparg believes that all scientific publishing eventually will move to the Internet, doing away with paper journals. That move will streamline a system in which, as Ginsparg puts it, scholars give their material to publishers for free and their institutions then pay thousands of dollars in subscription fees to read it in the journals. The compensation, up to now at least, has been that the leading journals provide "peer review," where respected members of a field of study read submitted articles and report to the journal on whether or not they represent good, original research.

The prestige of passing peer review and publishing in an established journal is still important to the careers of academic researchers, as is the quality control provided to the archival literature.

The papers that appear on arXiv.org are technically "preprints," the electronic equivalent of paper reports that researchers circulate among themselves in advance of formal publication. But more and more, at least in the physical sciences, researchers are communicating new results via their online postings, with journal publication a later formality.

Cornell librarians hope to explore the extension of this idea into other disciplines.

"There are a number of initiatives to look at how that would work in the biological sciences," Thomas said. "I would want to position Cornell so that we could be a very active contributor to the reconception of scholarly communication."

Cornell currently is engaged in a project to facilitate the electronic publication of mathematics journals, so far with strict controls on access. But, Thomas said, a movement is under way to persuade publishers to allow open access beginning several months after publication. "There are some models that suggest that the economic value of information [to publishers] declines sharply as it ages," she explained.

Ginsparg will begin transferring content to Cornell servers shortly, he said, with the expectation that the archive will officially move at the end of summer. For users, the transition should be seamless. The URL will remain the same: http://arXiv.org. (The old URL, http://xxx.lanl.gov still works and will continue to be a primary backup site.)

The move, Ginsparg said, coincides with the 10th anniversary of the archive, as well as the 20th anniversary of his Cornell Ph.D. and the first birthday of his daughter. - By Bill Steele

Related websites:

Ginsparg's online article on the archive, with thoughts on the future of scholarly communication

Cornell Faculty of Computing and Information

Cornell Digital Library Research Group

[Contact: Bill Steele]

18-Jul-2001