SPARC's greatest rhetorical achievement has been to establish the equivalence of support for FRPAA with support for Open Access. If you're on the right side of the issue and believe that the public must have free access to at least the author's final version of peer reviewed published papers, then you must be in support of FRPAA. If you question FRPAA then you must be in the pocket of the evil publishers, deviously trying to lock away the results of federally funded research from the deserving public.
But I'm troubled by FRPAA's willingness to make do with the author's final manuscript and with the scant lip service that it pays to interoperability, data-mining, and preservation. I think its one size fits all approach doesn't reflect the actual diversity of approaches to the literature among disciplines, and that its unwillingness to engage in some level of compromise with the publishing community sets up an us/them framework that has already done considerable damage to the scholarly community and that makes enemies of people who, in fact, have much more aligned interests than is generally recognized. Most worrisome of all, perhaps, the fact that FRPAA is presented as THE solution to public access gets in the way of productive discussion about how we might achieve a better path to open access.
The report from the Scholarly Publishing Roundtable, now two years old, still reflects the complexity of issues pretty well, and I'm encouraged by the fact that OSTP's RFI process clearly reflects those concerns. Although the report doesn't specify a particular policy approach it does outline pretty well the issues that ought to be balanced in the development of such policies.
There have been a variety of ideas floated over the years that are worth looking at in developing a policy approach that improves on FRPAA's flaws, in particular the linking proposal that was made by a coalition of 57 not-for-profit publishers to NIH back in 2005/6, as well as Stuart Shieber's 2009 paper "Equity for Open-Access Publishing".
Elements that might be incorporated into such policies would include requiring that grantees publish in OA journals (whether fully OA or hybrid), that metadata be submitted to granting agencies to link back to the Version of Record, that standards for data-mining and interoperability be developed across all of the federal granting agencies, and that publishers either adopt robust preservation/archiving strategies on their own, or participate in "trusted" endeavors (however those turn out to be defined).
Policies along these lines would privilege the Version of Record, eliminate the embargo period, and emphasize the development of standards for interoperability. They would provide a mechanism to encourage subscription publishers to develop fully OA business models, while injecting a level of market competition that could help to hold publication fees down.
They would also bump up against some issues of academic freedom by limiting where scientists could choose to publish, they would result in providing subsidies to commercial publishers for moving to OA, and they would not necessarily result in government controlled repositories like PubMed Central.
So there's still plenty to argue about in the details. But now is the time to have those arguments. With the publishing industry already experimenting in numerous ways with OA business models, with many publishers on record opposing RWA, while stopping short of endorsing FRPAA, and with the OSTP folks sifting through all of the comments in response to the RFI, now is the time to move past the limitations of FRPAA and try to engage fully with all of the stakeholders to achieve immediate open access to the Version of Record. Why in the world would we want to spend the next year engaging in a legislative fight that, even if it was won, wouldn't get us close to that?