---------- Forwarded message ---------- Date: Wed, 31 Mar 2004 12:39:51 +0530 From: Subbiah Arunachalam Friends: Here are some comments on one of our earlier postings. Arun [Subbiah Arunachalam] --------- Just some clarification on the posting of Subbiah Arunachalam, Leslie Chan, and Barbara Kirsop. Open Access to scientific documents as the solution for making articles visible to other users and thus maximizing their reading and citation (scientific impact) may involve various different forms of document space, effort-levels by the authors or their institutions, and degrees of success in achieving the aims of OA. 1. AUTHOR SELF-ARCHIVED, UNTAGGED PAPERS. The largest space of documents is raw, untagged preprints and scientific reports self-archived by their authors on their institution's webserver (e.g. on their research group's website). For Physics alone, we have already collected more than 500,000 such documents. These are harvestable with almost zero effort by our web crawlers. With Physnet http://www.physnet.net we have achieved already have almost complete coverage in some areas of Physics. (Clearly, however, the crawler of such websites can only be as good as the native language assistance offered to us: check a Finnish author!) Since the authors often provide their publication lists on their local websites, information on journal articles is also harvested. But because such authors have made no effort to tag their documents with metadata, retrieval is somewhat fuzzy -- yet almost complete for field-specific search engines such as PhysNet (started in 1995; but not so with google, scirus, etc.). 2. AUTHOR SELF-ARCHIVED, LIGHTLY TAGGED PAPERS. More functionality can be gained if the author adds some metadata, e.g., by using a web template such as http://www.isn-oldenburg.de/services/mmm/ and adding that to his document source code or by using a shadow file that points to the document. (See http://www.physik.uni-oldenburg.de/hydro/main_public.html [year 2002 and earlier] In case an author finds this too tedious, we [the ISN] will do it for your lists!) At the moment, in physics the portion of the document space provided in this form is under 10%. Visibility is much increased, since most search engines understand these metadata, thus they rank them favourably. Costs are small. Let as call 1 and 2 "individual author self-archiving." 3. INSTITUTIONAL OAI-COMPLIANT SELF-ARCHIVING. A University, Research Institute, Department, or University Library may set up an OAI-compliant data provider. The local document space is small; visibility is excellent because of the systematically added metadata, the larger public visibility of the institution as such, and easy retrievability by OAI service providers. This is Distributed institutional self-archiving and it increasing exponentially currently: http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0043.gif http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0036.png http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0037.png http://www.ecs.soton.ac.uk/~harnad/Temp/self-archiving_files/Slide0023.gif Costs may be larger (because library employees' salaries are involved). 4. CENTRAL OAI-COMPLIANT SELF-ARCHIVING. Authors send their documents to central archives, such as http://www.arXiv.org or the new European HAL http://hal.ccsd.cnrs.fr operated by the CCSD, supported by the French central research ageny CNRS, and under the auspices of the EPS with regard to physics. Costs depend on services (e.g., time stamp, permanent archiving, retrieval) and these are paid by the server's institution. Coverage is even smaller (ArXiv has some 269,439 documents to date) -- about 15% of physics journal publications at to date, but this may increase exponentially as more and more content authors convince others. 5. PEER REVIEWED JOURNAL ARCHIVES. The costs depend on the services offered. The prices charged are known. ACP (Atmospheric Chemistry and Physics) http://www.copernicus.org/EGU/acp/ has pretty much automatized the processing and covers its costs with $20 per page paid by the author, serving the journal open-access, with features such as several levels of refereeing, annotations, communications added by readers, etc. The Journals doing this form of archiving may be: (A) OPEN-ACCESS JOURNALS ('gold' journals in the Harnad and Romeo color-code), sometimes recovering costs by charging authors. Physics has 66 gold journals by now; see http://de.physnet.net/PhysNet/journals.html (B) TOLL-ACCESS JOURNALS ('green' journals in the Harnad and Romeo color-code), charging the reader but endorsing individual-article self-archiving by the author in one form or other (see the ROMEO list). <http://www.lboro.ac.uk/departments/ls/disresearch/romeo/Romeo%20Publisher%2 0Policies.htm> Now to close with some speculations about possible future scenarios for journal publishing: Since the refereeing normally takes place (and so it should) *after* the preprint is self-archived on the web in one of the above ways, refereeing can be much improved and diversified and can take its time (as it does with ACP http://www.copernicus.org/EGU/acp/publication_process.html ). Subscribing to a journal is thereby decoupled from gaining access to the information content of the raw document, and money is spent only on the refereeing and polishing, as well as the archiving of the document. Such a process of publicly self-archiving a document first and getting it refereed afterward would save money with which institutional libraries could subscribe to journals and would allow those publishers to flourish who add real value. The expert would gain the desired document directly from the author's website or elsewhere using search engines. Small departments in remote countries would be able to get the unrefereed information without having to pay, but they would miss the real added-value services. Eberhard R. Hilf, Dr. Prof.; CEO (Geschaeftsfuehrer) Institute for Science Networking Oldenburg GmbH an der Carl von Ossietzky Universitaet Ammerlaender Heerstr.121; D-26129 Oldenburg ISN-home: http://www.isn-oldenburg.de/ homepage: http://isn-oldenburg.de/~hilf email : hilf@isn-oldenburg.de tel : +49-441-798-2884 fax : +49-441-798-5851