An open letter to Jim Stasheff, re the proposed unified electronic preprint archive for mathematics Dear Jim, Last week you published a proposal for a unified electronic preprint archive in mathematics, at http://xxx.lanl.gov/list/math/info making use of the software at Los Alamos National Laboratory which was written by Paul Ginsparg, originally for a certain branch of Physics. Following our extensive semi-private correspondence on this subject, I think it is time for me to write something to the "categories" list. First, I want to commend the efforts of yourself and your committee in taking the initiative to call for a unified archive. Mathematics is a notoriously parochial subject, and the existence of such a central resource may in the long term help us to form The Big Picture. Having said this about the mathematical issues, I am unhappy about the way in which this proposal has been made in terms of management and technical issues. Such a unified archive will immediately acquire considerable de facto authority, and it creates a monopoly over the means of publication in mathematics which is potentially far more significant than the contraints of the existing journal system from which we are trying to free ourselves. You have also accepted wholesale Paul Ginsparg's design decisions, some of which may be accidents of implementation. In several respects there are or could be other paradigms of electronic publishing, which ought to be encouraged to co-exist with his ways of doing things, so that the community at large can decide over a period of years which method is to be preferred. I am worried that you may be making assumptions that things have to be done in one way when in fact alternatives are possible. In short, the proposal should have been "put out to tender" for other software designs and implementations to compete on fair terms on the basis of the facilities that they provide to authors and readers. In these respects this issue is analogous to the debate about TeX macros for commutative diagrams, on which a certain amount of blood was spilt some years ago. In that case someone who had used package X proposed that there be an electronic journal in category theory in TeX source, standarding on package X for commutative diagrams, without being aware of the existence of packages Y, Z, ... Actually this is quite different: in using TeX, or whatever technology they use for their papers, authors are making a considerable commitment, and some of us have made a far greater commitment by writing and maintaining macro packages. In the case of electronic dissemination (of existing files), papers can easily be copied to and linked from many different places without any further effort by the author. Nevertheless, bad decisions at this stage about the database which says where everything (and everybody) is could be significant handicaps in the development of the technology at a later date. I therefore beg you at least to stay the widespread advertising of the new archive until its potential competitors have had a chance to discuss the way in which they should co-operate and compete. I intend to contact Paul Ginsparg myself to discuss some technical issues. In particular I have a number of suggestions to make about the process of registering authors and submitting papers. His collected bibliographical data should be provided in BibTeX format for processing by other services. It would also be possible to integrate his author registration with Hypatia's, which would greatly enhance the usefulness of both systems. In fact it is remarkable how closely the strengths of one system coincide with the shortcomings of the other. Yours, Paul Taylor PS The technical discussion which I have been having with Jim Stasheff, Bob Rosebrugh, Mike Barr and others has now moved to a new email list run from Queen Mary and Westfield College. Details of how to join this, together with the archive of the discussion and other information, can be found at http://hypatia.dcs.qmw.ac.uk/html/eprint.html This message is followed by a separate one on the issue of submission in TeX source form.