Joe Halpern, and I have no serious disagreement at all. These points are really just about the fine-tuning:
On Wed, 3 Nov 2004, Joseph Halpern wrote: > For what it's worth, in CS, my anecdotal impression is that almost > all papers that I want to get are freely available on the web > (typically in citeseer or on author's home pages or both; occasionally > on CoRR; the CS part of the arXiv; hardly ever on departmental > archives; never, as far as I can recall, on university archives). So > it seems that, at least in CS, the vast, sluggish majority are > self-archiving somehow. This is not to say that it's not worth > encouraging similar behavior in other fields though! Joe may well be right about this, and it is very good news from the discipline that invented the Internet itself (CS: Computer Science)! > - Stevan says: > > > (8) What is certain is that if OAI-compliant self-archiving is to be > > mandated, it is institutions that are in the natural position to implement > > the mandate and monitor compliance (probably at the departmental level), > > for it is institutions (and not disciplines) that share with their own > > researchers the benefits of maximising research impact, and the costs of > > losing research impact. > > I agree that if there is going to be mandate then it will have to come > from either the universities or from funding agencies. My guess is > that it will be even more effective coming from funding agencies, but > that is not an argument against having universities mandate it as > well. (I have actually been trying to convince the NSF to impose > just such a mandate -- unsuccesfully so far.) If there is to be a > mandate at all, my distinct preference would be that it be to archive > on *some* OAI-compliant server, and not necessarily to archive on the > university server. I agree! The mandate should be OAI-compliant self-archiving. But because Joe is in a happy field, CS, where most authors already self-archive, he will perhaps not be as aware that for the "sluggish majority" of disciplines, both institutional OAI-compliant servers and central disciplinary OAI-compliant servers are still few. Joe will, however, appreciate that it is far cheaper and easier to create and maintain a local institutional OAI-compliant server, for local institutional self-archiving, than to create and maintain a central disciplinary OAI-compliant server, for self-archiving discipline-wide and worldwide. So that is the first reason for mandating self-archiving in any OAI-compliant server -- but expressing a preference for an institutional/departmental server. The second reason for preferring institutional/departmental servers is that institutions host all of their own disciplines, can mandate and monitor compliance locally, and can encourage the propagation of the practice of self-archiving across all of its disciplines and departments. None of this is true of central disciplinary servers. Most important of all: It doesn't matter, functionally, as both local institutional/departmental OAI-compliant archives and central disciplinary OAI-compliant archives are completely interoperable. >sh > (4) Logically and practically, if there existed a central, OAI-compliant >sh > archive for each discipline (and some central entity to foot the costs >sh > and maintain the entire disciplinary archive in each case, as the >sh > Physics ArXiv does today), then it would make absolutely no difference >sh > whether authors self-archived in their disciplinary OAI archive or their >sh > institutional archive. > > I disagree with this, at least the way things stand currently. In the > case of many subfields in physcis, the real "publication" of a paper > (in the sense of "making public") happens when it is posted on the > arXiv. Posting a paper on an institutional archive has a very > different effect (in terms of the paper being noticed) than posting in > on the arXiv. Maybe at some point it won't make a difference (when > all archives are linked into a centralized virtual archive), but now > it does. Agreed that in Physics, which today has a successful, long-standing central archive (ArXiv), containing a substantial portion of the discipline, a paper is more visible there than in a local institutional archive. But this is not true of the rest of the disciplines -- the "sluggish majority." It is not even true of Joe's discipline, CS, where *priority* (not "publication": both physics and CS still publishes in refereed journals and refereed conference proceedings) and immediate access is provided by self-archiving both preprints and postprints locally, and relying on citeseer (and google) to harvest them and make them accessible globally. Joe will agree that google is not the solution for navigating and searching the research literature as a whole, and that just as central OAI archives are rare and costly to create and maintain (ArXiv, CERN, SPIRES and PhysDoc, all in Physics, are the only significant examples), so home-page trawlers and harvesters like citeseer are rare and costly to create and maintain (with CS's citeseer and economics's RepEc the only two so far). Again, for the "sluggish majority", local OAI-compliant self-archiving and OAI harvesters like OAIster and citebase are the best bet. Once we have 100% OA, though, it will no longer matter much whether the papers happen to be in local university OAI archives or in central disciplinary ones. It matters now only for the likelihood of *reaching* 100% OA as soon as possible. (For the sluggish majority, it is already long overdue!) >sh > (2) What functionality does Joe think an individual OAI archive can >sh > provide for users (I am not speaking about depositing authors) that > sh> an OAI harvester and service provider could not provide, and better? > > I'm perhaps not imaginative enough to come up with lots of examples, > but the type of thing that I had in mind was that an art history > archive might provide particularly good ways of relating reproductions > that would be important for art historians. Similarly, a > genomics/computational biology archive might include gene sequencing > data and ways of accessing it. Clearly, both examples involve going > beyond just a repository of papers, but an archive of papers in a field > might well evolve in the direction of providing more than just a > collection of papers. Sounds like the more elaborate tagging, data-provision and tender-loving care that the papers' authors and institutions are more likely to have the time, resources and motivation to do (probably at a departmental level) than some as yet non-existent central discipline-based entity... >sh > What [archives] accept? It is journals that accept, and the target of OA >is >sh > the postprints accepted by the journals. The preprints are another matter >sh > and not central to OA. > > For me as a CS researcher, I'm often interested in the preprints that > haven't yet been accepted by journals. (The situation is more complicated > in CS > because conference papers are often never published in journals.) And > the arXiv definitely does have policies on what is acceptable, and it > varies by discipline. For example, the policy on CoRR is to accept > any paper with CS content, even if it's blatantly incorrect. The > physics arXiv tries to be (a teeny bit) more selective. Institutions and departments are quite as capable of establishing self-archiving policy guidelines for their own research output (i.e., who may self-archive what), locally, as is a central disciplinary entity, globally -- and institutions even have more experience and expertise interest vested in doing so, especially for unrefereed output! Cheers, Stevan
