I n 2001, 40 members of the editorial board of Machine Learning-then published by Kluwer Academic Publishers, now part of Springer-tendered their resignation. The reason behind the mass resignation was Kluwer's policy of restricting online access to Machine Learning articles to only subscribers. "Our resignation... reflects our belief that journals should principally serve the needs of the intellectual community, in particular by providing the immediate and universal access to journal articles

more »

... hat modern technology supports, and doing so at a cost that excludes no one," the members wrote in an open letter. Instead, they decided to found the Journal of Machine Learning Research (JMLR), a peer-reviewed, open access publication whose articles are freely available at the journal's Web site. "Open access is a model that makes sense," says Lawrence K. Saul, JMLR's editor-in-chief and a professor of computer science and engineering at University of California, San Diego. "You want your research to be read by as many people as possible. You don't want it to be gated by artificial barriers." The sentiment is common among today's computer scientists. In a paper world, goes the logic, the need of the scientific community to see its research circulate widely was aligned with the business model of commercial publishers, which had a financial incentive to ensure broad distribution. In the Internet age, of course, distribution is as easy as connecting to the Internet. Many researchers have thus grown reluctant to entrust their papers to journals whose online archives are restricted to those who can pay for them. A variety of issues are bound up in the discussion, from peer review and proofreading to research and the scholarly record. At the core, for many scientists, is copyright, and the con-Further Reading