Three academic groups have jointly floated a draft proposal in response to the US Government’s OSTP Public Access mandate memo. These groups are the Association of American Universities (AAU), the Association of Public and Land-grant Universities (APLU), and the Association of Research Libraries (ARL). Their proposal can be found here.
The public access system these groups propose to build is called SHARE which stands for SHared Access Research Ecosystem. It is not my intention to go over the proposal in detail. At this point I just want to make some simple comments on it. It is worth noting at the outset that it is not entirely clear from the proposal just who pays for SHARE, and a key part of the OSTP memo is that any new policy must be paid for from existing budgets.
The idea behind SHARE is basically that of scaling up the existing institutional repositories to contain all research articles based on federal funding, as a service to the government. This includes integrating these diverse repositories into a single searchable system. The underlying premise seems to be that since the principle investigators (PIs) of the federal research grants and contracts are at the universities, then the universities can handle the article ingest process via their grants management systems.
The driving motivation for SHARE, as the name itself suggests, seems to be that of discoverability and use of the articles. The proposal features what I would call the standard wish list of Open Access (OA) functionality, up to and including data mining. Granting these wishes is a tall order for the government, as can be seen from this paragraph in the SHARE proposal, which may give the government pause:
”Federal funding agencies need to receive sufficient copyright licenses to peer-reviewed scholarly publications (either final accepted manuscripts or preferably final published articles) resulting from their grants to enable them to carry out their roles in the national public access scheme. Such licenses would enable the placement of peer-reviewed content in publicly accessible repositories capable of preservation, discovery, sharing, and machine-based services such as text mining, once an embargo has expired.” (Page 10)
Also this:
“Federal agencies should require the use of persistent, unique identifiers for awards, publications, data, and authors to foster reuse of content and enable better grant tracking and the development of new services by individuals and machines.” (Page 10)
These are both very complex requests, requiring serious technological development, not to mention likely opening up a legal can of worms. Without either, this proposal may have great difficulty fulfilling its promise. Passing the buck, asking someone else to worry about the difficult parts of the policy, has not proven a successful strategy, at least for the RCUK where academic institutions are still up in arms over the abdication of responsibility for distributing funds to researchers.
On the other hand discoverability and use are serious policy issues which have yet to be addressed. Keep in mind that a multi-agency rulemaking lays ahead, where hopefully issues like this will be deeply discussed. With luck we might even see some cost-benefit analysis, which is required for significant new rules. In any case we will see some burden estimates as these are required for all new rules.
The primary shortcomings of SHARE seem to be these three: First it takes readers away from the publishers’ version of the article, reducing traffic and revenues, which will likely result in increased subscription and author charge rates. Second it imposes significant new burdens on the authors. Third it requires the government to assert a new set of rights. The benefits appear to lie in discoverability and general use, which are vague at best.
I also wonder about the apparent assumption that the lead author always works for a SHARE university. The federal public access claim seems to be on all articles reporting research funded at least in part by a federal agency. How big that part must be remains to be determined. Given the prevalence of multiple co-authors and especially international collaborations, this assumption is likely wrong in many cases, that is the PI need not be a SHARE university employee. Then too there are the national laboratories which produce a great deal of federally funded research under contract outside of the university system. In these cases the SHARE process gets decidedly more complex.
In my taxonomy of confusions the prominent features of SHARE seem to be overly complex procedures and weak factual assumptions.
Thus the complexities and burdens of SHARE seem to be significantly greater than those with the simple tagging and linking model proposed by the CHORUS group. There may indeed be a significant role for universities and university repositories to play in the overall scheme of things. But SHARE requires reinventing the wheel, recreating technologies and systems that publishers already have in place.
How this will play out in the final rulemaking remains to be seen. At this point there are basically four options on the federal table. There are CHORUS and SHARE, both of which involve public-private partnerships. Then there are two internal solutions, namely extending PubMed Central or cloning it at the agency level.
Discovery versus simplicity may turn out to be the key issue. Discovery is an important concept, but a separate one from access. Whether distrust (or dislike) of publishers is strong enough to drive a decision toward a more complicated, more expensive system remains to be seen. Perhaps the ideal solution lies in taking the best of each proposal—the straightforward nature of CHORUS as the front end, backed by federal and university systems to provide archiving and assurance for any areas where CHORUS falls short.
Discussion
7 Thoughts on "Universities Propose to SHARE Federal Funding Based Articles"
I think one of the biggest hurdles share will have to clear is to get all potential research institutions which might get funding to participate in share. I suppose participation could be mandated by the federal government, whether by statute or by requirements of the grant, but I don’t see this as easier to accomplish than getting all publishers to participate in CHORUS.
Anyone who thinks that integrating all these repositories is an easy and inexpensive tasks is not well versed in database management. SHARE refers to metadata needing to be tagged the same way, which is a nightmare. I am also concerned that the respositories as they exist now don’t do a sufficient job of tracking a paper. Errata, corrections and retractions are hardly ever recorded in the institutional repositories leaving the public/readers accessing faulty research. CrossMark tries to address that but publishers are not going to give SHARE final PDFs.
I will bet this proposal is handed over to a committee at each University. The committees will then have to create a SHARE committee to oversee the individual committees. The committees being spread accross the nation will require regional committees.
In short, the proposal will die in committee!
The guidelines take a holistic and integrated systems approach to digital academic research institutional repository management. The focus during development of the system has been extremely long term sustainability and visibility, on the internet, using open standards and open source software.
Indeed SHARE does take a holistic approach. The question is how this fits into the US Federal public access initiative, if at all? Quite frankly the proposal seems to be oblivious to Federal needs and constraints. My first thought when I read it was that it was a wish list written by repository managers. While grand, that may not be what the feds need at this point. They need a simple, cheap and quick solution. Budgets are being cut.