Return to Nature Debates: e-access

Los Alamos loses physics archive as preprint pioneer heads east

Declan Butler
European correspondent, Nature
Low vantage point: the loss of the preprint server is a blow for the lab on the New Mexico mesa.

The Los Alamos preprint server, which has established itself as physicists' favourite place for early circulation of their research, is leaving the New Mexico laboratory to set up shop at Cornell University in New York state.

Paul Ginsparg, who founded the server — now known as arXiv — 10 years ago, is leaving the Los Alamos National Laboratory (LANL) to take up a faculty position at Cornell, and the server will move with him. Cornell plans to expand arXiv's reach into other disciplines, including computer science, and to use it as a test bed for research into digital libraries.

Ginsparg says growing dissatisfaction with LANL is a major reason for his departure, citing a lack of enthusiasm for the archive among senior staff. Only his former group leader Geoffrey West and library director Rick Luce gave the archive strong support, he says. He adds that the nuclear-weapons laboratory has been shifting its support towards large groups at the expense of individual investigators, and is suffering from declining morale in the wake of recent security scandals.

Los Alamos experienced a painful security clamp-down after Wen Ho Lee, a Taiwanese-born engineer at the laboratory, was arrested two years ago on espionage charges and then released. The loss of the prestigious server delivers another blow to the laboratory's standing in the scientific community.

William Press, deputy director of the laboratory, says: "We're sorry to see Paul go, but Cornell has created a very unique opportunity for him. We are very proud to have been the incubator of this revolution in scientific publishing." He adds that senior laboratory staff have strongly supported the archive activity, but admits that it was sometimes "a struggle to see where it would fit in" with the laboratory's other activities.

The archive currently receives around US$300,000 in annual funding from the National Science Foundation, the Department of Energy, which runs the lab, and LANL itself.

Paul Ginsparg: set to move to Cornell.

Ginsparg says that consultation with the archive's advisory board, funding agencies and the American Physical Society, produced a consensus that the operation would enjoy more secure funding and stronger intellectual support at a university than at LANL.

But for Ginsparg, the last straw was his recent salary review, which, he says, described him as "a strictly average performer by overall lab standards; with no particular computer skills contributing to lab programs; easily replaced, and moreover overpaid, according to an external market survey".

LANL officials declined to comment on Ginsparg's case, but said that some recent salary increases at the laboratory have been available only to certain combinations of programmes and individual skills.

Peter Lepage, chair of Cornell's physics department, notes wryly of the LANL assessment: "Evidently their form didn't have a box for: 'completely transformed the nature and reach of scientific information in physics and other fields'."

Ginsparg has been hired jointly by the physics department and the recently-created Faculty of Computing and Information Science (FCI). "Paul brings vast first hand knowledge about one of the most significant events in the growth of the field of information science," says Robert Constable, head of the faculty. In his new role, Ginsparg will be answering questions such as "How will libraries evolve in the sciences? How can we automate more of the intellectual tasks provided by librarians?" adds Constable; "we are thrilled that he will continue this dialogue from a base at Cornell."

"The Physics Department was pleased to support the efforts of Bob Constable and the FCI to bring Paul here, because he is an excellent and versatile theoretical physicist," adds Lepage. "Also Paul's archive is transforming the way physicists interact with each other, and will ultimately redefine the nature of journals and scientific publication within physics."

Cornell University Library and the computing faculty will jointly fund both the operation of the archive and an associated research programme in digital libraries and information science. Adding a peer review layer to the archives is one area that will be explored, says Constable, as well as developing new metrics for evaluating scientific research. The archive will also be used as a testbed to develop better search, retrieval and data-mining mechanisms.

Sarah Thomas, director of the library, says that the move will guarantee long-term preservation of the archive. She estimates that maintaining the basic archive service would cost Cornell as little as US$150,000 annually, and require two staff, but that the library intends to invest in functionality, such as developing better submission and viewing facilities for users. Thomas points out that the archive will enhance Cornell's existing digital library initiatives, and will be used to explore and deploy new digital scholarly publishing solutions, aimed, in particular, at strengthening the role of research institutions themselves in the publishing chain.

(This is a slightly longer version of the original article published in Nature 412, 3-4; 5 July 2001)