We are so close to the adoption of Open Access Self-Archiving Mandates worldwide (with four of eight RCUK Research Councils plus the Wellcome Trust having already adopted them in the UK, the FRPAA proposing their adoption in the US, the EC Recommendation A1 proposing their adoption in Europe, at least 125 US university provosts expressing their support, and a number of individual universities and research institutions already adopting institutional self-archiving mandates of their own). This is the opportune time to think of optimizing the formulation of these mandates, so that they systematically interdigitate with one another to generate all of OA's target content, across institutions, disciplines, and nations worldwide, to confer the maximum of benefit in a minimum of time. A seemingly small parametric or verbal variant can make a vast difference in terms of the amount of OA a self-archiving mandate produces, and how quickly and reliably.
Other contents are more than welcome too -- pre-refereeing preprints, research data, theses, book-chapters, etc. -- but let us not forget that peer-reviewed research is the primary target and raison d'être of the OA movement.
(Only the final peer-reviewed draft need be deposited, not the publisher's proprietary PDF or XML, which should instead be linked, via a direct pointer to its URL or DOI on the publisher's website.)
WHERE: The optimal locus for self-archiving is the author's own OAI-compliant Institutional Repository (IR).
That is the locus which, once mandated, will systematically scale up to cover all of research output space, worldwide. It is highly inadvisable to mandate direct deposit in a Central Repository (CR) -- whether discipline-based, funder-based, multidisciplinary or national. The right way to get OA content into CRs is to harvest it, via the OAI metadata-harvesting protocol, from the distributed OAI-compliant IRs. Not only should research institutions -- the primary research-providers -- mandate the self-archiving of their own researchers' output in their own institutional IRs, but research funders too should mandate that their fundees self-archive in their own institutional IRs. That is the most natural, universal and systematic way to reach 100% OA worldwide, and also the fastest and surest.
WHEN: The author's final, peer-reviewed draft (postprint) should be deposited in the author's IR immediately upon acceptance for publication.
Most journals now endorse immediate OA self-archiving by their authors. But for the minority of journals that do not, the deposit should be mandated to be immediate anyway, and any allowable delay or embargo should apply only to the access-setting (i.e., whether access to the deposited article is immediately set to Open Access or provisionally set to Closed Access, in which only the author can access the deposited text). This is called the "Immediate Deposit / Optional Access" (ID/OA) mandate and it is infinitely preferable to any delayed-deposit policy: In the ID/OA, the article's metadata (author, title, journal, date, etc.) are immediately accessible webwide in any case, and would-be users can request individual email copies from the author via the IR's semi-automated EMAIL EPRINT REQUEST button during any embargoed access period.
The case for immediate access is exactly the same as the case for Open Access itself: otherwise research uptake, usage, impact, productivity and progress are needlessly delayed or lost. And in many fast-moving fields the "growth tip" for research progress is within the first 6-12 months from the time new results are available.
WHY: The purpose of mandating OA self-archiving is to maximize research usage and impact by maximizing user access to it.
The motivation for the Open Access movement -- and hence for OA self-archiving by researchers and OA Self-Archiving Mandates by researchers' institutions and funders -- is to maximize research access in order to maximize research uptake, usage, impact, productivity and progress, for the benefit of research, researchers, their institutions and funders, and the tax-paying public that supports them and in whose interests the research is being conducted and published.
HOW: Depositing a postprint in an author's IR and keying in its metadata (author, title, journal, date, etc.) takes less than 10 minutes per paper.
However, surveys show that only 15% of authors will self-archive unless it is mandated. Just requesting or recommending deposit does not work. Deposit analyses comparing mandated and unmandated self-archiving rates have shown that mandates (and only mandates) work, with self-archiving approaching 100% of annual institutional research output within a few years. Without a mandate, IR content just hovers for years at the spontaneous 15% self-archiving rate.
American Scientist Open Access Forum