Convergent IR Deposit Mandates vs. Divergent CR Deposit Mandates

From: Stevan Harnad <amsciforum_at_GMAIL.COM>
Date: Fri, 25 Jul 2008 09:03:59 -0400

Charles, if you don't mind, I've refocused the subject thread on the
actual topic under discussion rather than on my faulty thinking

On Fri, Jul 25, 2008 at 7:32 AM, Charles Oppenheim
<> wrote:

> As long as the deposited material is searchable using
> OAI-PMH, I, as a user, am indifferent what repository it
> sits in.

As user, you are absolutely right to be indifferent (except inasmuch
as you may be concerned about material that is not as yet being

Alas, the problem is not with the users of OA content, but with the
providers of it, namely, the authors.

And the problem is not with what *is* already being deposited, but
with what is *not* yet being deposited.

Mandates are to get material that is not OA to be made OA. Mandates
are needed, because otherwise only about 15% of OA's target material
is being spontaneously made OA by its authors.

You are absolutely right that OAI-PMH-compliant IRs and CRs are
completely equivalent and interoperable. Hence it makes absolutely no
difference to the user what repository the OA material was deposited
in, or sits in.

What matters is *that* it is indeed deposited.

And my point about why funders should mandate direct deposit in IRs
(and then harvest, if they wish, to CRs like PMC) instead of
unnecessarily mandating direct deposit in CRs is that CR mandates
reduce the probability that institutions (the source of all OA
content, funded and unfunded) will adopt institutional deposit
mandates, because (in a nutshell) they would entail divergent double
(or multiple) deposit for funded articles, instead of convergent
single deposit in the IR (and optional harvesting to or by CRs --
which is what the OAI-PMH was created for).
And the problem is getting authors to deposit even just once!

> Funders are entitled to insist where the outputs go.

Of course they are. And you know how gratified I and many others who
have worked hard for it are that RCUK, NIH, ERC and other funders are
mandating deposit. But the point is that OAI-PMH has freed the locus
of direct deposit to be any OAI-PMH-compliant repository.

So whereas mandating deposit at all is a great benefit for OA,
mandating convergent deposit in IRs (and harvesting to CRs), rather
than direct deposit in CRs, would provide an even greater benefit --
by facilitating instead of complicating institutional mandates, both
their implementation and their likelihood of adoption and compliance.

>They will often want it in their repository so they can keep
> track both of everything they have funded and also how
> many downloads such items receive.  

Harvesting it from the IRs not only tracks funding just as well as
direct deposit in CRs (instead of depositing in the CR, the author
merely provides the URL for its deposit locus in the IR), but the
monitoring of compliance is reinforced by the institution, and its
own record-keeping. (The Wellcome Trust has reported problems
monitoring compliance with its CR mandate, especially with divergent
deposit, sometimes being done by authors, sometimes done by

Monitoring download metrics is an important but complex technical
issue, currently in flux, with its outcome depending on how well OA
content-provision grows. I am not an expert on webstats, but here too
I suspect that download monitoring will be an OAI service, harvesting
from IRs and CRs alike. Again, convergence facilitates this, and
divergence complicates it.

> Stevan may disagree,
> but I don't think requiring deposit  for funded outputs in
> one particular repository reduces the appeal for IRs for
> research outputs.  Does Stevan have any evidence, e.g.,
> survey results, to support his views?  To deposit in both
> is anyway easy.

I have no formal survey evidence that authors don't mind depositing
twice or multiply instead of just once (but I have definitely heard a
number of complaints, particularly from long-standing Arxiv authors,
against having to do a second deposit in their own IRs; this has been
sufficient in some cases to discourage some universities from
pursuing institutional deposit mandates further, or has at least made
it harder for them).

There is abundat evidence that most authors will not deposit at all
unless mandated, and that many authors have the (wrong) idea that
depositing is a lot of work.

From this it seems reasonable to infer that getting authors to agree
Status: O
Message-ID: <>

to deposit even once is a hurdle, requiring a mandate. Bets are off
if it comes to the prospect of requiring them to deposit twice, or
multiple times, depending on funding source.

Authors have reported  (in Alma Swan's surveys) that they will comply
(95%) with a mandate (over 80% willingly), and when mandated they
actually do comply (Arthur Sale's studies), but no one has asked
authors about how they would feel about having to deposit twice, or
multiple times. Nor about how it alters the probability that an
institution will successfully seek or adopt a mandate if there is the
extra prospect of double or multiple deposit for funded research

(By the way, as Peter Suber has noted, OAI-PMH-based harvesting works
both ways: institutional IRs could back-harvest their own funded
output from funder CRs. But not only does that make multiple
institutions and authors do what one funder could do, far more
simply, but it is precisely contrary to the OAI-PMH concept, which is
that central service-providers harvest from distributed
content-providers, and not vice versa.)

> What is important is that OA marches on, not where the OA
> material appears.

Quite. But the probability that authors' fingers will march their
papers into an OA repository depends on their getting their marching
orders from deposit mandates; and both the likelihood of adoption and
the likelihood of compliance with deposit mandates is higher if the
deposit mandates are institutional and convergent rather than
multiple and divergent -- especially for the sleeping giant, which is
the institutions themselves, the providers of all the research,
across all disciplines, funded and unfunded.

Stevan Harnad
Received on Fri Jul 25 2008 - 15:19:40 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:49:23 GMT