SUMMARY: The most effective and natural way to ensure that all institutions — the universal providers of all research, funded and unfunded, in all fields — provide open access (OA) to all of their peer-reviewed research (funded and unfunded, in all fields) is for both funders and institutions to mandate cooperative, convergent deposit, by the author, in the author’s own institutional repository, rather than competitive, divergent institutional-and/or-institution-external deposit by authors-and/or-publishers.
1. It is important for OA advocates to understand that it is not PubMed Central (PMC) that is making biomedical articles open access (OA) — it is the depositors of those articles. In the case of PMC, those depositors are authors (and publishers). PMC is serving both as a locus of deposit (i.e., a central, subject-based repository) and as a locus of search and use (like google).
2. The reason PMC has as much OA content as it does is the NIH OA mandate, requiring deposit — not the fact that PMC is a central locus of deposit.
3. The NIH mandate is a very good thing, and if OA is desired in other funded fields, the funders should mandate OA, as NIH and other funders have done (and as FRPAA is proposing to do).
4. But the fact that NIH requires direct deposit in PMC is not a good thing, even though it is being mindlessly imitated by some other funders and fields.
5. The universal providers of all research, funded and unfunded, in all fields, are institutions (the world’s universities and research institutes).
6. Hence, in order to ensure that all research, funded and unfunded, in all fields is made OA, it is essential that the universal providers, the institutions, mandate providing OA to their own research output.
7. Most research universities already have their own institutional repositories (IRs), but most do not yet require their researchers to deposit their research output in them; for universal OA, the world’s universities and research institutes must all mandate OA.
8. But if funder mandates stipulate that deposit must be institution-external (e.g. in PMC, or other institution-external repositories) then funder mandates not only fail to encourage and reinforce the incentive of institutions (the universal providers) to mandate OA (for all the rest of their research output, funded and unfunded, in all fields), but they compete with and discourage it, by needlessly requiring direct deposit elsewhere (and hence the deterrent prospect of authors having to do duplicate or multiple deposit, should their institution consider mandating OA for all of its research output).
9. For funders to split the task of fulfilling their OA mandates arbitrarily between authors and publishers is even more counterproductive, making it ambiguous who must deposit what, and ceding responsibility and control over when the deposit is done to publishers, whose own interests are not aligned with doing but with delaying deposit.
10. For these reasons, funders should unequivocally assign full responsibility for complying with their OA mandates to their fundees (and their institutions); those are in any case the only parties actually bound by or beholden to the funder’s mandate — and their own interests are aligned with immediate deposit rather than delay.
11. The most reliable, effective and natural way to ensure that all institutions — the universal providers of all research, funded and unfunded, in all fields — provide OA to all of their peer-reviewed research articles (funded and unfunded, in all fields) is accordingly for both funders and institutions to mandate cooperative, convergent deposit, by the author — once — in the author’s own own institutional repository, rather than competitive, divergent institutional-and/or-institution-external deposit by authors-and/or-publishers.
12. Central collections like PMC can then be harvested (or automatically exported, e.g. via SWORD) from the distributed, mandated IRs, rather than being deposited (or double-deposited) into directly. (General web content providers do not deposit directly in Google either; they host their own content locally and then Google and others harvest it!)
13. Of course, if an author or institution does not yet have an IR, it is fine to deposit in a central IR like DEPOT that has been created specifically for that purpose.
14. It has been a historic (but correctable) strategic mistake (by NIH and its imitators) to cause funder and institutional mandates to diverge and compete, because funders needlessly insist upon direct institution-external deposit.
15. There is no need to imitate this initial counterproductive trend, by founding still more central repositories and encouraging funders to require direct deposit in them:
16. This gratuitously dysfunctional trend should be countered, by encouraging funders to mandate institutional deposit, and by ensuring that central collections are harvested from distributed IRs, rather than being designated as the loci of direct deposit.
Swan, A., Needham, P., Probets, S., Muir, A., Oppenheim, C., O?Brien, A., Hardy, R., Rowland, F. and Brown, S. (2005) Developing a model for e-prints and open access journal content in UK further and higher education. Learned Publishing 18 (1). pp. 25-40.