Jim Till's experience with Journal of Cancer Integrative Medicine, that he among other things wanted to
> "test whether or not the editors and the publisher would permit me > to retain copyright(and, the right to self-archive a postprint)," is one we should all follow. I suggest that when we do publish in a non-OA journal we should always try to negotiate that at least our article be OA as a experiment by the publisher (preferably in the true sense of being accessible directly via the journal); it will often meet with a positive response--and at the least, will educate the publisher. (I have such a request pending now for a section I was asked to guest-edit in a journal). It would seem as a long range strategy that, apart from the experiment he mentions, Jim's desire that it be "the perceived merit of the *article* itself that should (if possible) be evaluated, as directly as possible, rather than via more indirect proxy-indicators (such as the perceived merit of the *journal*) would be best met by a general abandonment of the practice of publishing in journals at all, and replacing them by publication in an appropriate repository alone. If that is his meaning, I urge further experimentation on its real effects, as it is surely the least expensive of all solutions. I call attention to the implications of his experiment on the related topic of the effectiveness of A&I services. The journal in which he published is not even listed in Medline; it did not seem to hurt his readership. The journal cited in ref.1 of his paper (CA: a journal for clinicians) -- with one of the highest impact factors in the world -- is certainly in Medline, but that particular article for some incomprehensible reason is not. My understanding of the extensive relevant experimentation is that a very good A&I service will provide slightly higher performance than Google, but it seems that the practical performance of such A&I services in the real world is so greatly below the theory for even the best ones, such as Medline has always been reputed to be, that the theoretical advantage may not be relevant. Perhaps Stevan will consider some of these points irrelevant to his goal of achieving open access by his preferred means as soon as possible, but perhaps others will think it more prudent and more effective to advance in several mutually supportative ways. Dr. David Goodman Associate Professor Palmer School of Library and Information Science Long Island University dgood...@liu.edu and, formerly, Princeton University Library