Dear Jim, Thank you very much for your questions and interest in the CPS. The CPS should indeed satisfy the "inter-operability" criterion when we achieve compliance with the Open Archives Initiative. It is our intention that the CPS will be compliant at the start of October. I also agree that the "views" and "ranking" statistics could provide indicators for the impact of a particular preprint. By a "citation data" indicator, I understand that the impact would be ascertained by examining the number of other papers (both inside and outside the server) which cite that preprint. This is an interesting idea and I would certainly like to like to learn more about how this could be achieved. For the "sign-posting" criterion, I agree that it is important to provide authors with the ability to link to the published version. As you mention, this is currently possible by adding the relevant hyperlink into the discussion group of a particular preprint. I would like to add that authors of preprints submitted to the CPS also have the option to redirect the final version of the preprint to the final version on the publisher's website, if and when the preprint is published. After the first version of the preprint has been submitted, the author (and only the author) is presented with three hyperlinks when they access their article page: 1) Add more supplementary files; 2) Revise the full text of the preprint; 3) Redirect to the published article. This redirection is achieved using the LitLink technology of MDL Information Systems. The author would submit the ISSN of the journal, the year, volume and starting page number. When users then view the article page, they are presented with a "Published full text" link. When this link is accessed, LitLink resolves the citation and finds from where the article may be downloaded. Clearly, if this is from a publisher's website, users would typically have to pay for access. However, all of the other information - including the preprint meta-data and any other files uploaded to the server - do of course remain completely free to access on the CPS. In terms of this "sign-posting", I do think that it is equally important that other authors link back to references which appear on preprint servers. Every article that is uploaded to the CPS is given its own citation reference in the form CPS: category/YYMMNNN where 'category' is the chemistry classification to which the preprint was submitted, 'YY' is the year, 'MM' the month, and 'NNN' represents the number of preprints submitted to that category in the month. In a similar manner, the article is also given a "friendly URL" - http://preprint.chemweb.com/category/YYMMNNN. If a user accesses this URL they are taken directly to the article, without having to first browse through the server. In this way, it is easy for authors to reference the CPS preprints. I hope that this helps to clarify. Any suggestions or comments would of course be very welcome. Kind regards, James
________________________________________________________________________ James Weeks Chemistry Preprint Server Coordinator ChemWeb Inc. 84 Theobald's Road London WC1X 8RR United Kingdom Tel: +44 (0) 20 7611 4480 Fax: +44 (0) 20 7611 4301 Email: james.we...@chemweb.com http://preprint.chemweb.com http://www.chemweb.com _______________________________________________________________________ -----Original Message----- From: Jim Till [mailto:t...@uhnres.utoronto.ca] Sent: 24 August 2001 19:10 To: american-scientist-open-access-fo...@listserver.sigmaxi.org Subject: Re: Elsevier's ChemWeb Preprint Archive A comment in response to previous messages from James Weeks: In a message that I posted to this forum on 24 May 2001, on the subject "Re: ClinMed NetPrints", I tried to outline three criteria (or, 'design & usability guidelines'?) for an eprint archive: 1) an 'inter-operability' criterion; 2) an 'impact-ranking' criterion; 3) a 'sign-posting' criterion. James, in the message that you posted to this forum on August 17, you indicated that there are plans for the Chemistry Preprint Server (CPS) to be OAI-compliant with the next two months. If that goal is accomplished, the first ('inter-operability') criterion will (I assume) have been met. Perhaps one way to begin the meet the second ('impact-ranking') criterion is to provide the kind of "views" and "ranking" indicators that are commented upon in the message that you posted on August 23. Better, though, might be the suitability of the eprint server for yielding citation data? As Tim Brody pointed out (as part of a previous thread, in a message posted on May 24, on the subject: "Re: ClinMed NetPrints"): [tb]> Which, from my technical point of view, is the reference lists for [tb]> the articles. As far as I'm aware no archives currently do this (I [tb]> know cogprints provides the facility for authors to give this [tb]> information, but does not re-export yet). Watch developments from [tb]> OpCit! Re the third ('sign-posting') criterion, it's possible for the author of a preprint posted at the CPS website to add, as part of a discussion thread, a citation to the published version of the preprint. (There will, of course, usually be a time delay between the posting of the preprint and the appearance of the published version, unless the 'preprint' is, in fact, a post-print). The arXiv server, in contrast, provides (what seems to be) a quite convenient means for authors to add, to their previously-posted preprint, a citation (and a link) to the published version. James, might you be willing to comment on these three proposed criteria (or, guidelines) for eprint servers, and on their relevance to the CPS? Jim Till University of Toronto