On 14 Aug 2010, at 20:16, Stevan Harnad wrote:

> On Fri, 13 Aug 2010, Richard Poynder wrote:
> 
>> [1] Should institutional repositories [IRs] be viewed as preservation tools?
> 
> Not primarily. IRs' primary function should be to provide open access [OA] to
> institutional research article output.

Yes. We may have witnessed a golden age of digital preservation tools, and some of these have been built into repository software interfaces. To explore the practical application for repositories, see our structured and fully documented KeepIt course on digital preservation tools for repository managers:

Source materials http://bit.ly/afof8g
Blog http://blogs.ecs.soton.ac.uk/keepit/tag/keepit-course/

The underlying philosophy of the course is to enable users to evaluate the appropriate degree of commitment, responsibility and resource for preservation that is consistent with the aims and objectives of the institution and repository at a given time and looking forward. It follows that answers can range from high to low, even to nothing, providing the analysis has been thorough, the results documented and the decisions and consequences are fully understood.

Without commenting on priorities here, IRs are much wider than OA papers. For IR preservation it's this broad scope that matters, then how policy deals with the specifics, rather than simply OA concerns.

> 
>> [2] Should self-archiving mandates always be accompanied by a “preservation
>> mandate”?
> 
> Definitely not. (But IRs can, will, should and do preserve their
> contents.) For journal articles, the real digital preservation problem
> concerns the publisher's version-of-record. Self-archiving mandates
> pertain to the author's-draft.

Not an additional mandate, agreed, and it's important that institutional and repository policy, such as OA mandates, precede preservation policy and provide the basis for it. But it's interesting to ask whether OA mandates, since at the moment these are the most prominent form of repository policy, should make some reference to preservation. It's notable that research funder OA policies are more likely to make some brief reference to preservation than institutional policies.

To Stevan the answer may seem obvious in the particular case of OA, but the question is whether such policies would benefit from such a reference. Or more broadly, whether repository policies need to demonstrate some degree of reciprocity, not just preservation, for the demands they appear to make of authors. Given the weight of an institution's repository policy, it will have to address this at some stage, and omission, even from an OA mandate, since IRs are wider than OA, could begin to look curious and raise questions. The wider context is what repositories can offer in terms of responsible content management for access now and longer-term access. It will do no harm to sprinkle policies with features that will appeal to authors, where repositories can take practical steps to implement these. Stevan says IRs should and do preserve their contents; in which case, IRs simply need to specify and demonstrate what this means in practical terms, where possible, and policy is one prominent place to do this.

In this case return to [1] above, but first see conditions in [3] below.

> 
>> [3] Should Gold OA funds be used to enable preservation in institutional
>> repositories?
> 
> Funds committed to Gold OA should be used any way the university or
> research funder that can afford them elects to use them (though does
> seem a bit random to spend money designated to pay for publishing in
> Gold OA journals instead to preserve articles published in
> subscription journals).
> 
> But on no account should commitment to fund either Gold OA or digital
> preservation of the version-of-record be a condition for mandating
> Green OA self-archiving.
> 
>> More, including an interview with digital preservation specialist Neal
>> Beagrie, here: http://bit.ly/dur5EP

Stevan has long been concerned about costs and distractions, including preservation, to the core OA aim. Economics are the primary driver here. As Neil Beagrie said in the interview: "digital preservation is "a means to an end": the benefit and goal of digital preservation is access for as long as we require it". This can work for open access too. My experience is that repositories are not wasting time and effort on preservation where it may be unnecessary, e.g. empty repositories. On this basis, it is too stark for Richard Poynder to say: "Nevertheless it is hard not to conclude that there is a potential conflict between OA and preservation." 

For others the problem may be the opposite, of turning concern into action. There is emerging evidence that repositories will take the necessary actions on preservation where the tools are available and when the circumstances support this, e.g. these repositories:

NECTAR and the Data Asset Framework – first thoughts http://blogs.ecs.soton.ac.uk/keepit/2010/02/07/nectar-and-the-data-asset-framework-first-thoughts/
Digital Preservation, Risk Management, and UAL Research Online http://blogs.ecs.soton.ac.uk/keepit/2010/06/11/digital-preservation-risk-management-and-ual-research-online/
Digital Collections Risk Assessment at LSE: Using DRAMBORA http://blogs.ecs.soton.ac.uk/keepit/2010/07/19/digital-collections-risk-assessment-at-lse-using-drambora/

To try and gauge what circumstances might convert concern over preservation into action by repositories I recently proposed this rough metric
http://blogs.ecs.soton.ac.uk/keepit/2010/07/22/conditions-for-digital-preservation/

When these conditions apply, again, return to [1] above.

I've made the case before that the issue between support for green and gold OA, from an institutional perspective, is one of chronology, and it's the same for IRs and preservation.

Steve

> 
> Richard Poynder's Interview is, as always, well worth reading.
> Comments follow (linked version is at http://bit.ly/DigPreservVSoa ):
> 
> Commentary on Richard Poynder's
> "Preserving the Scholarly Record:
> Interview with digital preservation specialist Neil Beagrie"
> 
> The trouble with universities (or nations) treating digital
> preservation (which is a genuine problem, and a genuine
> responsibility) as a single generic problem -- covering all the
> university's (or nation's) "digital output," whether published or
> unpublished, OA or non-OA -- is not only that adding an additional
> preservation cost and burden where it is not yet needed (by conflating
> Green OA self-archiving mandates with "preservation mandates" and
> their funding demands) makes it even harder to get a Green OA
> self-archiving mandate adopted at all. But taking an indiscriminate,
> scattershot approach to the preservation problem also disserves the
> digital preservation agenda itself.
> 
> As usual, what is needed is to sort out and understand the actual
> contingencies, and then to implement the priorities, clearly and
> explicitly, in the requisite causal order. The priorities here are to
> focus university (or national) preservation efforts and funds on what
> needs to be preserved today. And -- as far as universities' own
> institutional repositories (IRs) are concerned -- that does not
> include the publisher's official version-of-record for that
> university's (or nation's) journal article output. Preserving those
> versions-of-record is a matter to be worked out among deposit
> libraries and the publishers and institutional subscribers of the
> journals in question. Each university's own IR is for providing OA to
> its own authors' final, refereed drafts of those articles, in order to
> make them accessible to those users worldwide who do not have
> subscription access to the version-of-record. The author's draft does
> indeed need preservation too, but that's not the same preservation
> problem as the problem of preserving the published version-of-record
> (nor is it the same document!).
> 
> Perhaps one day universal Green OA mandates will cause journal
> subscriptions to become unsustainable, because the worldwide users of
> journal articles will be fully satisfied with just the author's final
> drafts rather than needing the publisher's version-of-record, and
> hence journal subscriptions will be cancelled. If and when we ever
> reach that point, the version-of-record will no longer be produced by
> the publisher, because the authors' drafts will effectively become the
> version-of-record. Journal publishers will then convert to Gold OA
> publishing, with what remains of the cost of publication paid for by
> institutions, per individual article published, out of their windfall
> subscription cancellation savings. (Some of those savings can then
> also be devoted to digital preservation of the institutional
> version-of-record.)
> 
> But conflating the (nonexistent) need to pay for this hypothetical
> future contingency today (when we still have next to no OA or OA
> mandates, and subscriptions are still going strong) with either
> universities' (or nations') digital preservation agenda or their OA IR
> agenda is not only incoherent but counterproductive.
> 
> Let's keep the agendas distinct: IRs can archive many different kinds
> of content. Let's work to preserve all IR content, of course, but
> let's not mistake that IR preservation function for journal article
> preservation or OA.
> For journal articles, worry about preserving the version-of-record --
> and that has nothing to do with what is being deposited in IRs today.
> 
> For OA, worry about mandating deposit of the author's version -- and
> that has nothing to do with digital preservation of the
> version-of-record.
> Nor should the need to mandate depositing the author's version be in
> any way hamstrung with extra expenses that concern the publish's
> version-of-record, or the university's IR, or OA. (Exactly the same
> thing is true, mutatis mutandis, at the national preservation level,
> insofar as journal articles are concerned: A journal's contents do not
> all come from one institution, nor from one nation.)
> 
> And, while we're at it, let's also keep university (or national)
> funding of Gold OA publishing costs distinct from the Green OA
> mandating agenda too. First things first. Needlessly over-reaching
> (for Gold OA funds or preservation funds) simply delays getting what
> is already fully within universities' (and nations') grasps -- which
> is the newfound (but mostly unused) potential to provide OA to the
> authors' drafts of all their refereed journal articles by requiring
> them to be deposited in their OA IRs (not by reforming journal
> publishing, nor by solving the digital preservation problem).
> 
> Stevan Harnad