Thursday, October 18, 2001

40 Computer Scientists Abandon a Print Journal, Preferring Its Online Competitor By ANDREA L. FOSTER In another sign of scientists' discontent with costly traditional journals, at least half the editorial board of a journal on artificial intelligence has resigned to join a competing journal that is distributed free online. Forty computer scientists at the traditional journal, Machine Learning, signed a resignation letter this month and distributed it widely. It said the journal's subscription fee was so high that scientists' articles were not reaching a wide-enough readership, and also complained about the publisher's restrictions on circulating articles online. Kluwer Academic Publishers, which produces the monthly journal, responded by saying that it would increase the number of journal pages published each year, from 1,100 to about 1,300, and reduce the annual subscription rate for individuals, from as much as $460 to $120. The company also agreed to disseminate articles on its Web site and to allow authors to post journal articles on their Web sites. The subscription rate for universities and research centers remains at $1,050. Many of the scientists who resigned from Machine Learning said they believe that traditional publishers, such as Kluwer, have become superfluous to the circulation of scientific articles. And many predicted that the new online publication they are backing, the Journal of Machine Learning Research, would become the established journal on artificial intelligence, perhaps threatening the viability of Machine Learning. The scientists' letter said that "journals should principally serve the needs of the intellectual community, in particular, by providing the immediate and universal access to journal articles that modern technology supports, and doing so at a cost that excludes no one." It also said that the Journal of Machine Learning Research embraced the scientists' vision of a contemporary scholarly journal. The editorial board and authors "retain significant control over the journal's content and distribution," and articles are available freely on the journal's Web site, the letter said. Articles are later available in print through the MIT Press. Another reason the computer scientists said they supported the online journal is that it can take more than a year for authors to see their articles in print in Machine Learning, whereas it takes much less time for articles to appear in its online competitor. Both publications require articles to be peer-reviewed. The letter, dated October 8, was drafted by Michael I. Jordan, a professor of computer science and statistics at the University of California at Berkeley. He said Machine Learning is not affordable for many institutions in developing nations, such as China and India. "It doesn't help any of us in our research efforts," he said, to lock out a large part of the scholarly community. He said he had circulated the letter of resignation, in large part, to facilitate the name recognition of the Journal of Machine Learning Research among scientists who may be reluctant to submit articles to a journal established only 18 months ago. Its founder is Leslie P. Kaebling, a computer scientist at the Massachusetts Institute of Technology who was formerly on the editorial board of Machine Learning. The executive editor of Machine Learning, Robert C. Holte, said that his publication was not threatened by the resignations, and that he believed the market could sustain two journals on artificial intelligence. But he also questioned the upstart publication's ability to survive. "You can't instantly have a reputation of 15 years," he said, referring to Machine Learning's age. "It takes time for the community to catch on, to be recognized by tenure committees and research-assessment boards and outside communities who might be interested in aspects" of artificial intelligence. Mr. Holte, a computer-science professor at the University of Alberta, in Canada, said his colleagues had pressed him to resign from Machine Learning and become the executive editor of the Journal of Machine Learning Research, but he declined. "Only one journal has a proven delivery system that meets the community's needs," he said. Nonetheless, he acknowledged that Kluwer had reduced its subscription fee and eased restrictions on the distribution of journal articles only after the establishment of the Journal of Machine Learning Research. The editorial board had asked Kluwer to institute those changes five or six years ago, he added. Representatives of Kluwer, which is based in the Netherlands, were unavailable for comment. The resignations at Machine Learning reflect widespread dissatisfaction among scientists with the traditional model for publishing scholarly papers. Many editors of the journal Logic Programming resigned in 1999 to join Theory and Practice of Logic Programming, a computer-science journal published by Cambridge University Press that encourages the electronic dissemination of articles. And prominent scientists who are members of a group called Public Library of Science started a boycott on September 1 of scholarly journals that refuse to make articles freely accessible online within six months of their publication in print. Background articles from The Chronicle: Opinion: No Free Lunches: We Should Resist the Push to Rush Research Online (10/12/2001) Scholars Urge Boycott of Journals That Won't Join Free Archives (4/6/2001) 'Open Archives' Project Promises Alternative to Costly Journals (12/3/1999) Opinion: The New Scientific Publishing Wars (9/17/1999) Easy-to-print version E-mail this article

Colleges have cut the percentage of full-time faculty members, study finds NASA administrator, praised as an advocate of space studies, will step down New degrees, majors, or programs are announced by 12 institutions 40 computer scientists abandon a print journal, preferring its online competitor E-book provider netLibrary puts itself up for sale, worrying librarians Indiana University upgrades its supercomputer, said to be academe's most powerful Higher-education council backs a new virtual university for Virginia