[This is for SEPTEMBER98-FORUM] I would like to underline David Goodman's comment on Jim Till's list of reasons that the literature could and should be free. David is exactly right that the arXiv is faster and more convenient than the journal system. This is not so much a "could" or a "should" as an "is". It simply is the raison d'etre of the arXiv; consequently it is the main reason that the recent literature in many areas of physics and a few areas of mathematics is free. The beautiful citation age graph at
http://opcit.eprints.org/tdb198/opcit/citationage/ all but proves it. The citation cycle in high-energy theory is now shorter than the time to publication. Therefore formal publication is too slow to insert into the cycle; there is no putting it back. The "shoulds" that Jim Till lists - the information gap, the library crisis, public availability, and academic freedom - are all noble causes. But none of them apply to the biggest users of the arXiv, which as Stevan likes to emphasize is a user-driven system. Prestigious researchers have always been the first in the door in successful sections of the arXiv. Surveys of the math arXiv reveal that there are relatively more submissions from the Ivy League and other prestigious universities, and at the other end that the most prestigious journals are overrepresented as venues of publication of math arXiv articles. Harvard University has the largest and most expensive library system of any university in the United States, possibly of any university in the world. Yet it is also the source of 262 articles in the math arXiv as of today, which is many multiples of the per university average: http://front.math.ucdavis.edu/search/from:harvard Most of these submissions come from postdoctoral members who, while they have full access to the Harvard libraries, couldn't care less about publishing "issues" or "policy". They read and feed the arXiv strictly for their own convenience. -- /\ Greg Kuperberg (UC Davis) / \ \ / Visit the Math ArXiv Front at http://front.math.ucdavis.edu/ \/ * All the math that's fit to e-print *