Librarians and scientists both fear that if a commercial journal exists only in electronic form, the unprofitable archives of back files would not be a high priority and might get dumped. Not only that, but when your on-line subscription expires, so does your access. In contrast, if you don't renew a print journal, the back issues you've already paid for are still yours to keep.

But e-journals have a lot of advantages over printed publications: papers can be disseminated almost instantaneously; all sorts of search options are available; papers on similar topics can be retrieved through links; references and an author's previous work can be linked to the current paper; comments and comments on comments can be hooked onto a paper; video, sound, 3-D graphics, and data sets can be incorporated into the text. Even something as simple as color images can be added without the extra charges that publishers currently impose. So it isn't only the economics of journal pricing that is prompting the revolution; the technology is already there and waiting for it.

Caltech's first action was to convene a conference on "The Future of Scholarly Communication" in March 1997. Attended by 55 representatives of 29 universities, the conference featured four speakers who were prominent proponents of electronic publishing, as well as two panels-one of university provosts and one made up of representatives of professional societies.

"We brought together people who had the power to make decisions," says Flagan, an avid advocate of the e-revolution, "some librarians, but also people who oversee the library budget and who are motivated to see something happen. The questions that were addressed at that conference were basically: What is it that universities need to communicate for the future? What would we do with a clean slate? Suppose there were no journals today, and suddenly this thing called the Internet came along and we wanted to do something to communicate the results of our research? How would we do it if we were starting from scratch?"Journals do, of course, contribute some value for the money: they provide mechanisms for editing, for distribution, for easy access to information, for preservation of the scholarly record, and for certi-fication. Certification, in the form of peer review, is critical to the functioning of research universities, and it's what gives the journals their enormous clout. This stamp of approval on someone's work determines who gets hired, who gets ten-ured, who gets promoted. But the main insight to come out of Caltech's 1997 conference is that peer review is not inherently tied to a print jour-nal. Academics are doing this work for free any-way; they could just as well do it in another kind of distribution system, say an electronic one, if universities agreed to stand together to accept this stamp of approval. Koonin is credited with advancing this notion, henceforth referred to as "decoupling" the refereeing from the journals. And it goes further: you can also decouple the editorial function and the archiving.

The librarians (who have to pay the journal bills) and the provosts (who have to come up with the money for the librarians) left the conference inspired by the prospect of decoupling. The group most dependent on the journal system, however-the faculty-was not so easy to convince. If tenure and promotion are tied up with the old system, who would want to take a risk on something new and unknown? Many believe that what the journals provide-vetted and edited papers aggregated into neat little packages as the traditional ticket to tenure-is worth the cost to libraries, as well as the price of giving up copyright.

But in a faculty meeting in the spring of 1998, Koonin issued a challenge. He noted that Caltech already requires faculty members to sign an agreement that all patents and copyrights that result from their Institute- or grant-supported work belong to Caltech. No one had ever thought much about copyright, and that part of the agree-ment was never enforced. But suppose we started enforcing it? mused Koonin and then suggested that Caltech authors withhold copyright from the publishers. No groundswell emerged of professors eager to attempt this, and Caltech, of course, never followed through on the implied threat. Koonin, at least, has practiced what he preached, and one journal to which he contributed subsequently changed its copyright policy in response to his raising the issue.

The copyright challenge did rouse a few ad-herents at both poles-those enthusiastic about taking on the journals and those who fervently believed that the present system worked just fine-with the vast majority of the faculty indifferent to the entire issue. Professor of Planetary Science Bruce Murray set up an on-line, "threaded" discussion system-a "hyperforum"-to discuss copyright and the question: "Will the accelerating trends toward electronic publishing and Internet commerce overturn traditional relationships between university researchers, publishers, and the scientific endeavor?" Over the three months that it was up in the fall of 1998, the hyperforum attracted log-ons from only 40 members of the faculty, 16 grad students, 1 postdoc, and 40 members of the staff. Of these, only 16 posted comments.