Shared publicly  - 
 
Comments on the RCUK's draft new OA policy

Here are my comments on the draft new OA policy <http://goo.gl/O8Plt> for the Research Councils UK. (The RCUK has called <http://goo.gl/BaXbR> for public comments on the draft. To submit your own, email them to <communications@rcuk.ac.uk> and use "Open Access Feedback" in the subject line.)

First, I support all the major elements in the new draft:

1. I support its demand for libre OA, not merely gratis OA, and its demand for the least restrictive open license (CC-BY or equivalent). This will maximize the usefulness and usability of RCUK-funded research.

I should add that in the past I've hesitated to call for libre OA mandates. If they're adopted too early, then some publishers may refuse to publish authors bound by them. But whenever I expressed this concern, I also called on funders and universities to watch for the moment when it would be safe to adopt libre OA mandates, and to expect that to a large extent this safety would a consequence rather than a precondition of adopting a libre mandate. I believe the time has come for libre OA mandates, and that a group of large and important funders, like the RCUK, can compel publisher accommodation more readily than a small funder.

2. I support the requirement that RCUK-based OA include explicit permission for text and data mining. Works that are OA under CC-BY licenses may already carry all the permissions needed for text and data mining. But they may not; I've seen some disagreement about that. Hence the new explicitness can only help. One of the chief benefits of libre OA is to remove doubt about permissions so that users needn't slow down to ask permission, take the risk of proceeding without permission, or err on the side of non-use.

3. I support the way the policy allows gold or green OA to satisfy the OA requirement.

4. I support capping embargoes at six months for the natural sciences. I understand the case for variable embargo periods, and for longer embargoes in the social sciences and humanities than in the natural sciences. I regard twelve-month embargoes in the SSH fields (indeed six-month embargoes in the STM fields) as a compromise with the public interest. But I accept that some compromise may necessary to muster political support for policies that are strong in other respects. For the same reason, I support attempts to minimize compromise and return as far as possible to the public interest in immediate OA. Hence, I support the passage announcing that the AHRC and ESRC will work towards reducing their maximum embargo periods from twelve to six months.

5. I support the way the new draft defines OA conditions for the publishers of RCUK-funded research. As I read it, this would essentially bring the RCUK into the line of precedents created by the Wellcome Trust and NIH. If a given publisher will not allow OA on the funder's terms, then grantee must look for another publisher. Because the NIH has taken this strong position, 100% of surveyed publishers now accommodate the NIH policy.
http://oad.simmons.edu/oadwiki/Publisher_policies_on_NIH-funded_authors

6. I support the requirement for a data access plan, and I understand why the data access policy cannot yet be made specific and uniform across agencies in different fields.

In addition to supporting these aspects of the draft policy, I'd like to offer three recommendations:

1. For grantees choosing green rather than gold OA, the policy should require deposit at the time of acceptance for publication, even if deposits are not made OA until the embargo period runs. The NIH has this kind of immediate deposit requirement, but FRPAA <http://bit.ly/hoap-frpaa> does not, and its omission is one of the only weaknesses in FRPAA. All modern repositories can host "dark" or non-OA deposits, which allows policy-makers to require deposit at one time and OA release at a later time. Modern repositories can also be programmed to open up access for a certain article on a certain date (when the embargo runs), so that no human needs to remember to do so.

The chief advantage of immediate deposit is to increase the compliance rate. Authors deposit when their minds are still on the new publication, not several months later when they're deep into a new project and don't want to be bothered with the loose ends of a previous project. A side benefit is that while an article is on dark deposit, its metadata can be OA, making the work visible to search engines. This helps other researchers in the field learn that the work exists; and if they wish, it enables those without institutional access to write to the author for a copy. Some repositories facilitate this process with an "email request button" accompanying each dark deposit.

2. For grantees choosing green rather than gold OA, some Research Councils will have reason to require deposit in UKPMC. But unless a funder has a good reason to require deposit in a particular repository, I recommend that the new policy allow flexibility on this front. For example, FRPAA would allow US federal agencies to require deposit in any repository meeting certain conditions of OA, interoperability, and long-term preservation. This opens the door for a FRPAA-covered agency to require deposit in the author's institutional repository rather than a central disciplinary repository. (BTW, I support the way Section 3.b of the draft policy requires that the repository be independent of the publisher.)

3. I can tell that the policy allows RCUK grantees to use grant funds to pay publication fees at fee-based OA journals. And I can tell that when the RCUK pays any part of the cost of publishing an article, then the article must become OA immediately (without embargo). But the policy does not seem to take the next step, taken by the Wellcome Trust and UK Funders Group, requiring that when the funder pays any part of the cost of publishing an article, then the article must become libe OA under an open license (preferably CC-BY). I recommend that the RCUK policy take this additional step. The Wellcome Trust and UK Funders Group are absolutely right to insist on it, and the fact that the NIH has not yet followed suit is one of the only failings of the NIH policy. Those who pay a publication fee to cover a journal's expenses should get libre OA for their money. By covering the journal's expenses, they free the journal from the need to protect a revenue stream, and this freedom should be reflected in enhanced user rights under an open license.

Peter Suber
Director, Harvard Open Access Project
Fellow, Berkman Center for Internet & Society
Senior Researcher, SPARC
gplus.to/petersuber

#oa #openaccess #rcuk
16
10
Vladimir Blagoderov's profile photoFabiana Kubke's profile photoOpen Science's profile photoNicolás Caitán's profile photo
9 comments
 
Timothy: The RCUK call for comments didn't mention a deadline. So to be safe, I'd say the sooner, the better.
 
Thanks for this, Peter. I drafted my own comments (but have not yet sent them) before reading yours, so as to avoid inadvertent influence. What you have written seems very good to me, except for one point which I find mystifying:

"Some Research Councils will have reason to require deposit in UKPMC. But unless a funder has a good reason to require deposit in a particular repository, I recommend that the new policy allow flexibility on this front ... This opens the door for [an] agency to require deposit in the author's institutional repository."

Why would this be a good thing? I see the proliferation of thousands of separate, disconnected and incompatible institutional repositories as a bit of a disaster. Aren't Green-OA articles much easier to discover when they're in a big, well-known repository? (Obvious example, though it's not quite the same thing, is arXiv.)

Am I missing something?
 
Hi Mike. All good OA repositories today are interoperable by virtue of complying with the OAI-PMH (Open Archives Initiative Protocol for Metadata Harvesting). A user can search all compliant repositories without knowing which ones exist, where they exist, or what they contain. They are not separate silos requiring separate searches, and that is the point. In addition to OAI-PMH interoperability, these separate repositories are indexed by all major search engines, academic and non-academic, which is another reason why there's no disaster. In the benefit column, if OA mandates from funding agencies require deposit in the author's institutional repository, then they contribute to a culture of self-archiving at that institution, which will carry over to research not funded by that agency. Hence, there will be a net gain in the number of articles becoming OA. Moreover distributing the literature among nodes in a network of repositories makes the whole system more robust in the face of the usual threats to funding and long-term preservation. (Think: risks of monoculture.) I'd be happy to say more if you like, but that's the short version of the argument. On the other hand, I've often said in public that the stakes are fairly low on this point. I'd be happy to see these works deposited in disciplinary repositories like PMC and arXiv, even if I also think there's an extra benefit in requiring deposit in institutional repositories.
 
Thanks, Peter. I know that in principle all repositiories can be harvested by OAI-PMH (I am a metadata programmer in my day-job). In practice, even when OAI-PMH is supported and even when a good enough subset of metadata is provided that a reasonable description of the reposited papers can be created, the promise that this facility makes doesn't seem to be fulfilled. You say "A user can search all compliant repositories without knowing which ones exist, where they exist, or what they contain" -- is there such a single search point for the world's repositories? If there is, then my objection largely evaporates, but I've not seen it.

"if OA mandates from funding agencies require deposit in the author's institutional repository, then they contribute to a culture of self-archiving at that institution, which will carry over to research not funded by that agency" -- why wouldn't that statement be equally true if we substituted "in a subject repository"?

As for robustness: a distributed network is more robust if you want to ensure that SOME papers survive, but a single (replicated) meganode is more likely to ensure that ALL papers survive.

I'd be happy if I knew that somewhere we have a meganode that is aggregating metadata (and ideally mirroring content) from ALL the various institutional repositories.

(I hope it's apparent that I'm not looking for a fight here -- just wanting to understand what the state of the art is before I submit my comments to RCUK.)
 
Hi Mike. G+ comments don't make it easy to respond to many points at once, but here goes.

(1) OAIster used the be the most comprehensive OAI-PMH-based cross-archive search engine. But today most cross-archive searches use full-text indexing from Google, Bing, et al. rather than OAI-PMH. In my view this is a step up. Full-text searching is more useful than metadata searching, and Google (and company) are more familiar to users than OAIster. Full-text searches also bypass the (fairly serious) problem of incomplete and inconsistent metadata.

(2) It's true that depositing in a subject repository can nurture a culture of self-archiving. But it will have less institutional involvement and support. It's a matter of degree. At least to date, more librarians, administrators, and student workers mobilize to support faculty deposits in an institutional repository than to support faculty deposits in a subject repository. In general, institutions are in a position to foster self-archiving (through education, assistance, and incentives), and disciplines are not. OA policies that steer work toward institutional repositories take advantage of that network of support.

(3) A central repository could be well-preserved, but there is still the monoculture problem. Earlier this year the National Biological Information Infrastructure (NBII) was defunded by Congress and shut down. It had many pieces, all of them OA. I believe that all the pieces were preserved, but not all of them are still OA and it may be years before they are all OA again. You write, "I'd be happy if I knew that somewhere we have a meganode that is aggregating metadata (and ideally mirroring content) from ALL the various institutional repositories." Me too. In 2004 I got permission from Brewster Kahle to build that kind of meganode at the Internet Archive. But I didn't have the technical competence to build it myself and people with the competence generally didn't have the time. In principle, this project could be revived.
 
Thanks, Peter.

On OAIster -- my brief experiments have left me unimpressed. For example, a search for "dinosauria" gets zero hits -- this for the name of maybe the most-studied fossil group. Either no-one anywhere in the world has ever reposited a dinosaur paper with their institution, or (more likely) OAIster is not very good. http://www.oclc.org/search/search.pl?site=new_worldwide&client=new_worldwide&q=dinosauria&submit=Go

So we are down to the Kill-'em-all-and-let-Google-sort-it-out approach. In general I am a big fan of that -- I like the idea that the group that does accessioning doesn't have to be the group that does indexing, UI, etc. But in this case I'm not convinced it's working. I use Google a lot in searching for papers, and it's vanishingly rare that it's found for me one that it's in an IR.

Still, there's enough in your response that I now plan to not make a recommendation either way as to institutional vs. subject repos in my submission. (I think there are still plenty of problems with institutional, but they more or less balance out.)

Regarding building a meganode: I'd like to talk to you off-list but I (perhaps stupidly) can't find the send-an-email button. (I am pretty new to Google+, having resisted FaceBook-alikes for some time.) Could you please contact me on Sauropoda@gmail.com? Thanks.
 
For others reading this thread, I just responded to Mike offlist. My public email address is <peter.suber@gmail.com>.
Add a comment...