As an administrator for an institutional repository with both full-text and metadata only records, there are three questions that I return to again and again:
- How to increase deposits of full-text?
- Is linking to an open access copy as good as hosting a copy?
- How careful should I be about copyright?
I haven’t found definitive answers to these questions. Here are my thoughts at the moment:
1. How to increase deposits of full-text?
Ask, and keep asking! For over two years I’ve been monitoring new papers published by University of Bath authors and requesting copies for our repository. Slowly but steadily the response rate is increasing. But overall, the rate of deposit is still disappointingly low. This is where I pin my hopes on funder mandates.
2. Is linking to an open access copy as good as hosting a copy?
Where there is a reputable and reliable open access copy of a publication elsewhere, I don’t think it’s worth spending time tracking down a copy and appropriate permissions to host it. Instead, I add a link flagged ‘free full-text’ to our metadata only record and consider my job done. We take these links into account when providing statistics on full-text access. But what constitutes a reputable and reliable open access copy? If the DOI links to an OA copy, that’s fine. PubMed, ArXiv: good enough for me. But what about research reports on charity or government websites? Often the copyright is held by the commissioning body, so linking is much easier than getting permission to host a copy. But how long before their website gets rearranged or simply disappears and my links break? Which leads me on to question 3:
3. How careful should I be about copyright?
I expect most of the copyright holders for these various reports and papers would be quite happy for the author to make a copy available from our repository. But getting that permission explicitly stated is time consuming. Should I just take the risk and upload a copy to our repository without asking for permission? Rely on our notice and take-down policy? I don’t because I know that, whether or not the rights holder minds, it breaches copyright. But this rigour works against my efforts to increase full-text deposit rates. I’m forever turning away proof copies, online first copies, and author’s personal use eprints explaining that although they’re not the final published version, they’re not the permissible author’s accepted version either. Is this admirable protection of my institution’s reputation? Or is it unhelpful hair-splitting? Copyright decisions are often a case of risk management: how much risk should I be willing to take to further open access?
7 thoughts on “Three Perennial Questions”
Let me start with an input on your final remark on Sword non-compliant commercial CRISs: the RepNet is already in conversations with both Atira (PURE) and Avedas (Converis) as to ensure both systems will shortly become Sword-compliant in order to make them available for Repository Junction Broker (RJB) submissions. Developers from both companies said providing Sword endpoints to PURE and Converis is rather straightforward and has been sitting in their to-do list for system enhancement for some time now, only they hadn’t seen the point for doing it so far since no one had specifically asked for it. This feature needs of course a complementary simple CERIF/DC mapping that will make the RJB agnostic to target systems and formats – we are also at it.
As for your comment on the false CRIS-IR dichotomy, I take note of your point, although it is also a fact that commercial CRISs seem to be overtaking the RIM arena for REF2014 compliance purposes, making CERIF the default “language” RIM speaks along the way. In order to find out whether we face a true or false dichotomy (bearing in mind for instance that the UoE ERA repository has now been downgraded and stripped of the research output mgt in favour of the brand new PURE), I think we first need to know what the actual RIM system landscape looks like across HEIs, in order to check how many institutions follow the Leeds Met pattern you describe. Once we know that we shall be in a good position to start promoting new repository services being sure that they will be useful for HEIs that rely on their IRs for research information mgt. In this regard I’m curious to know whether that fine-looking CRIS/IR data collection the RePosit googlegroup members started to put together some time ago was actually filed somewhere and could be retrieved for update.
Thanks for update on RJB…rather self-centred of me but already wondering what the implications might be for a Symplectic + IR set-up whereby we could take SWORD deposit into the IR but I’d then need to somehow link the output back to the Symplectic record, probably something to bring up with Symplectic…
We weren’t actually part of RePosit of course (though Leeds were) and I never did actually see the final report…think there was a delay with it being signed off by Jisc and I’m not sure if it is actually available? Regarding the CRIS/IR data collection I started a Google doc (which is probably out of date now but is editable) at https://docs.google.com/document/d/1BQadMoMXbKHucuzlBpJajQVuJTrxWzJ-gR_K2nhyniQ/edit
Little mention has been made along the heated Gold vs Green debate to the fact that while Finch/RCUK/BIS may be favouring Gold OA, the JISC is at the same time funding the UK RepositoryNet+ (aka RepNet) project for delivering new services to institutional repositories that may bring them back into the main picture at HEIs. While it is true that CRISs are very tough competitors for IRs and, as Nick says, they will change (and have already) the rules of research information management, it is also a fact that not every university runs a CRIS or has plans to implement one. Another scarcely discussed fact is that the current Open Access policy shift towards Gold OA may actually represent an opportunity for IRs to become extremely valuable institutional assets for tracking Open Access outputs and payments HEIs will face an increased need to report on. Part of the new services the RepNet project is designing for repositories -to be shortly presented and discussed at the next UKCoRR members meeting at Teesside U- aim to provide IR managers the tools to tackle some of the issues mentioned in this post – with an emphasis on the first one. “How to increase deposits of full-text?” (i) By implementing enhanced automated deposit procedures and (ii) by reporting on what is available and what is missing despite funder policies requiring it. More on this (and a number of additional things) on Nov 9th.
I’m looking forwards to hearing more about RepNet from the UKCoRR members meeting on the 9th Nov. Complying with, and showing that we’re complying with, the RCUK’s new policy is going to be an important new part of repository work.
I just wanted to add that personally I’m not necessarily keen on the terminology “CRIS” which (I think) tends to refer to one of the 3 main commercial systems and perhaps reveals a (false) dichotomy between CRIS and repositories with the latter ideally being a component of an integrated research management infrastructure, whether that comprises commercial CRIS software or, say, a fully plugged-in EPrints repository in the Glasgow Enlighten model. At my own institution, for example, we have Symplectic and I’m working on implementing that software’s API to feed publication lists to researcher web-pages including links to DOI and full-text in our repository.
I certainly agree that, post Finch, IRs have the potential to become extremely valuable institutional assets for tracking Open Access outputs which is one of the reasons I’m so excited about the possibility of COUNTER compliant usage data from IRUS-UK; also the potential of the Repository Junction Broker to automate deposit (though had an interesting breakout with Ian Stewart at OR2012 about possible limitations for CRIS systems which, afaik, aren’t currently SWORD compliant.)
These 3 points are absolutely fundamental to repository administrators though I do wonder how the changing landscape of Open Access (Finch / new RCUK policy) and new software and tools (I’m thinking particularly of CRIS in an institutional context but also Mendeley as a potential disruptive technology of the “social” web) will impact on them.
It all ties in to some extent with recent discussions on both our main mailing list around the the new RCUK policy and alleged favouring of Gold over Green and on a separate discussion list we run for users of Symplectic Elements (which may be more or less relevant to users of Atira Pure / Converis) around lack of embargo support in that software and the balancing act still required to integrate with a repository workflow that publishes straight to live OR maintain a mediated deposit process. I’m not so sure myself I’m afraid, whether the current minefield of Open Access copyright (even mediated by the Godsend that is SHERPA RoMEO and its API increasingly integrated into CRIS) is suited to scalable self-deposit, even with a mandate. It’s just too confusing! Whether or not you agree with the potential shift to Gold as the best way to realise full OA (and many of us in UKCoRR and beyond have reservations, particularly around cost) might it mean that *institutional* repositories (perhaps as a component of CRIS) are actually in a stronger position to preserve (in the DCC sense) and disseminate the full-text (and associated outputs such as data-sets) of their scholarly communities?
The new RCUK policy, for example, informed though it may be by Finch and controversial as it has become in some quarters, states that in order to be compliant a journal must offer a suitable gold option i.e. immediate (unembargoed) OA to the version of record from its own web site, under a CC-BY license, AND must allow immediate deposit of the version of record in an OA repository, also under a CC-BY license OR a suitable green option i.e. allow deposit of the peer-reviewed manuscript in an OA repository not operated by the publisher subject to an optional embargo of no more than 6 or 12 months depending on the specific Council.
As one of our colleagues from the RSP has pointed out, although the RCUK gives some support for green open-access it does not necessarily mean that it will happen and if academics wanted or knew how to self-archive they would be doing so by now, however, there is no doubt that awareness of OA has increased which along with government/RCUK policy (caveats not withstanding) and improved tools / related infrastructure can help make repositories remain, or even become increasingly valuable to preserve and disseminate the “version of record” for our institutions…and may help to mitigate all three of the points you raise here?
The increased awareness of OA is definietly helping deposit rates. I like to think that awareness and understanding of the copyright issues and options will grow enough that there’s no longer a need for the safety net of repository staff checking all the material deposited in the repoistory. But I don’t think we’ve reached that point yet.
Andrew Adams’ is much bolder, his advice is:
“Do not put any librarian or other administrators in the way of the deposit for “quality checking”, “publisher policy checking” or anything else. … No publisher has ever sued a university over making their academics’ papers available. At most you need to respond to “take-down” requests by setting access to closed instead of open …” http://www.a-cubed.info/OA/
I wouldn’t be comfortable with that yet, but maybe one day?