On Sat, Jul 26, 2008 at 5:47 AM, Thomas Krichel <kric...@openlib.org> wrote:
>  TK: >>  [If NIH and other funders were to stipulate that institutional IRs >>  are to be the preferred locus of deposit then] >  [T]he funder  [1] would still have to be aware of all institutional >  repositories, [2] harvest the metadata from all, [3] police IRs >  so that the funder information is actually correct, [4] find themselves >  in the funder field, [5] adapt procedures to find the full-text from the >  metadata, [6] harvest the full text. [7] It's very difficult to do at present, >  because no appropriate registries exist. Please weigh your hypothetical concerns about [1] - [7] in the light of NIH's actual current implementation options for the means of deposit: Authors may submit a paper to the journal of their choice for publication. There are four methods to ensure that a manuscript is submitted to PubMed Central in compliance with the NIH Public Access Policy. Method A: Publish in a journal that deposits all NIH-funded final published articles in PubMed Central (PMC) without author involvement. Method B: Make arrangements to have a publisher deposit a specific final published article in PubMed Central. Method C: Deposit the final peer-reviewed manuscript in PMC yourself via the NIH Manuscript Submission System (NIHMS). Method D: Complete the submission process for a final peer-reviewed manuscript that the publisher has deposited in the NIH Manuscript Submission System (NIHMS). These methods vary in the version of the paper submitted, and the actions undertaken by the author and publisher. Please see Submission Process for more information. >  [T]he funder  [1] would still have to be aware of all institutional repositories, Not if automatic IR export to PMC were stipulated as NIH's first and preferred deposit option. The author and institution would have to be aware of the need to comply (and that is normal), and would be systematically set up to do so efficiently. The institution would also be the one monitoring compliance as part of the grant fulfillment condition, which is also normal. NIH actually would have less monitoring work rather than more, as now, when deposit might come from author or publisher. >  [2] harvest the metadata from all, That is part of the automatic export for the IR to PMC. >  [3] police IRs so that the funder information is actually correct, IRs would be "policing" their own deposits, and the institution would be monitoring NIH transfer as part of its grant fulfillment condition (as authors and universities already do with their grant fulfillment conditions). >  [4] find themselves in the funder field, Automatic export of metadata from IR to PMC, along with the URL of the deposit. >  [5] adapt procedures to find the full-text from the metadata, The URL in the IR is the link to the full-text. The full-text can also be part of the export, if desired. >  [6] harvest the full text. It can be automatically exported, if desired. >  [7] It's very difficult to do at present, because no appropriate registries exist. No registries are needed. (Registries are always welcome, but for this purpose, they are not necessary.) >  If a universal registry of authors and institutions can be >  be built, then it becomes reasonably easy to gather within an >  IR the OA material authored by all authored of the institution >  irrespectively of initial locus of deposit. I am working >  precisely on these registries at this time. Your universal registry would certainly be useful for many purposes, but it is not necessary for this purpose. NIH already knows who is funded. It is the responsibility of the author and the author's institution to ensure that all peer-reviewed articles resulting from that funding are deposited. The IR deposit and export would provide the articles as well as the record. > I am opposed to institutional mandates as the way to >  populate an IR. Institutions should encourage deposit but >  respect the freedom of academics to publish their work as >  they see fit. Alma Swan's international, interdisciplinary author surveys show that most authors disagree with you: Few of them deposit spontaneously, but 95% of them report that they would deposit if their employers and/or their funders mandated it, and over 80% of them say they would do so willingly. In addition, Arthur Sale's outcome analyses comparing mandated and unmandated IRs have shown that authors do indeed comply, as the surveys indicate, when deposit is actually mandated. And apparently the faculty of Harvard University's faculty of Arts and Sciences and the Faculty of Law, Stanford University's Faculty of Education, and two dozen other institutions and departments worldwide differ with your view that deposit should only be encouraged rather than mandated. Arthur Sale's analyses also show that encouraging alone is ineffective. And of course NIH's initial policy of merely encouraging rather than mandating has famously failed and been replaced now by a mandate. It seems rather arbitrary to suggest that funders should mandate but institutions should only encourage -- not only because encouragement fails, but because both funders and employers are in effect funding their research output. My guess is that your confidence in unmandated deposits comes from the practices in your field of economics: >  No RePEc archive that I know of (there are now >  over 900 contributing archives, so I can't be sure) has been >  populated with a mandate. But economists (in RePEc) and physicists (in Arxiv), along with computer scientists (on their websites, as harvested by Citeseerx) are long known to be the only three, long-standing exceptions to the status quo, which is that spontaneous (unmandated) self-archiving across all disciplines is 15% or less. (Economics, Physics and Computer Science probably raises the overall average.) These practices have not changed in a decade and a half: Economists, physicists and computer scientists continue to do it spontaneously, but their practice does not transfer to other disciplines, which, when surveyed, state exactly what they stated in Alma Swan's surveys: We will only deposit if/when it is mandated (over 80% of us willingly). So mandates are clearly the way. The issue here concerns how to leverage funder mandates, like NIH's, so they can facilitate rather than complicate the adoption of institutional mandates -- by converging on one-stop IR deposit and export to PMC instead of divergent and multiple deposit that might discourage the adoption of institutional deposit mandates. >  Experts >  use RePEc not an example for central deposit but to emphasise >  the need to get community involvement. Some of the elements >  that made RePEc a success can be exported to an interdisciplary >  level. Two of these are the author and institutional registration. >  Again this is precisely what I am working on. Thus is the same >  way that I have been battling for years to set up RePEc, against >  all odds since no such system had been set up, I am now battling to >  on these registries. Your hypothesis would seem to be that the reason why, for 15 years now, no other discipline than physics, economics or computer science has been spontaneously self-archiving above the 15% baseline is because, unlike economics, they lack a registry: Registries seem fine, but they are certainly not the reason computer scientists deposit unmandated, and I doubt registries are the reason physicists or economists deposit unmandated. For their own cultural reasons (different in each case), these three disciplines have taken to self-archiving of their own accord, and no other discipline has done so, across the same decade and a half, despite the availability of FTP, the web, the OAI-PMS protocol, IRs and CRs, and OA impact statistics. They only deposit if mandated -- and then most do so willingly. The evidence seems clear. Hence what is needed is something that will facilitate the adoption of deposit mandates, especially by institutions, which are the universal providers of all research, funded and unfunded, across all disciplines. There is good reason to believe that convergent institutional and funder mandates will facilitate the adoption of deposit mandates,  especially by institutions, whereas divergent institutional and funder mandates, requiring differing, multiple loci of deposit will not. The simple solution is for both funders and institutions to mandate IR deposit, with IRs automatically exporting deposit also to funder-mandated CRs. Stevan Harnad