Re: Elsevier Still Solidly on the Side of the Angels on Open Access

From: Stevan Harnad <>
Date: Wed, 18 Jul 2007 21:03:32 +0100 (BST)

On Wed, 18 Jul 2007, Steve Hitchcock wrote:

> Until recently I would have thought that David
> was overstating the case about the limitations of
> discovering papers in IRs, but now I tend to
> agree. There is no conspiracy with journal
> publishers. Simply the services available have
> not kept up. Where is Web 2.0 for IRs?

Nope, the problem is not with the findability of the OA content (in IRs
or anywhere). It is with the absence of (85% of) it. Don't blame
services for failing to find absent content!

> There are problems with identifying full-text
> availability and versioning/duplication in IRs.

Nope again: The problem is not plenty of content yet too many
versions/duplicates of it. (If that were the problem, it would
quickly and easily be solved!) The problem is the absence of
*any* version of (85% of) the content!

> To overcome this we need better OAI and better
> services. There is work going on that will offer
> some opportunities but it needs to be focussed by
> recognising the problems we are dealing with.

Better OAI and services are always welcome, but they will not
solve the real problem, which is absent OA content. That is
the (*only*) real problem we are dealing with (and dancing
around, in all directions, instead of solving it.)

> One of the principal reasons for introducing IRs
> was the lack of takeup of the subject model
> beyond Arxiv, and the explicit link with the
> author's institution.

Lack of takeup up in the sense of failure to self-archive,
not in the sense of preferring to self-archive here, rather than
there. IRs were introduced so all institutions could provide OA
to their own research output (in all subjects).

> Now the situation may be
> different. Most recent research funder OA
> mandates are open about subject repositories vs
> IRs (although where they are less open about this
> mandates tend to favour subjects).

The fact that some funder mandates favour central repositories is a big
strategic error, and not one to be admired, encouraged or emulated,
but one to be steered firmly in a more sensible, thought-through
direction: IRs. (The same is true with the ill-considered mandates that allow
deposit to be embargoed, instead of requiring immediate deposit and only
allowing Open Access-setting to be embargoed: That too is not to be cited
as a rational law of nature but as a silly, short-sighted decision by
a few of the first funding mandators, to be corrected before it hardens
into common practice.)

> If subject
> repositories exploit the inherent advantage of
> visibility in a given field they could claim more
> content.

"Subject" repositories have no inherent advantage of visibility in a
given field if they are empty. And if they happen to have content
(beyond the 15% spontaneous baseline) as Arxiv did, then their
advantage is not because of their centrality but because of their
content! That very same content would have had the very same
visibility if it had been deposited in IRs and harvested by OAIster
or Google Scholar or Citebase or Citeseer -- or, for that matter,

If someone has the hypothesis that having subject content in
subject-based central repositories will help generate more OA
self-archiving -- or more OA self-archiving mandates -- then let them
harvest IR (and other) OA content metadata into subject-based
meta-repositories. Let them not confuse the issue by recommending
direct central self-archiving (again) after it has already failed.

> Of itself that isn't a problem for those
> fields covered by mandates, but what about the
> rest?

Isn't the answer obvious: Mandate the rest too! Not subject by subject
but institution by institution, for all of an institution's subjects.
And funders should mandate *institutional* self-archiving too. Then, if
you wish, harvest the IR content metadata into subject-based

> IRs remain important, and have to do more
> to be highly visible, or they risk becoming
> secondary sources, as David suggests, with the
> consequences that follow.

Both David and Steve are missing the crux of the matter: There is
little spontaneous self-archiving, either way, except in physics and
computer science (for cultural and historical reasons). In physics,
spontaneous self-archiving happens to have been much more substantial,
and central. In computer science spontaneous self-archiving happens to have
been much more substantial, and distributed (and harvested by Citeseer,
and now Google Scholar). Both cases are exceptional solely because they
are providing high volumes of content spontaneously (unmandated); not
for any other reason (one being central and the other distributed).

IRs do not need "to do more to be highly visible." Their problem is
not their invisibility, it is their emptiness. And Steve ought to
know this, because his own department's IR is anything but invisible
-- for the simple reason that it has content; and it has content
because self-archiving is mandated!

With too many of the (still few) funders foolishly mandating central
self-archiving instead of Institutional Self-Archiving, it remains
for the sleeping giant -- the universities, the primary providers of
*all* research, in *all* subjects, whether funded or unfunded -- to set
the right example, by mandating self-archiving in their own IRs. Then
funders will catch on and reinforce institutional self-archiving by
requiring their fundees to self-archive in their IRs too. And then,
if they wish it, central, subject-based repositories can harvest from
the IRs willy-nilly, as they see fit.

Stevan Harnad

If you have adopted or plan to adopt a policy of providing Open Access
to your own research article output, please describe your policy at:
    BOAI-1 ("Green"): Publish your article in a suitable toll-access journal
    BOAI-2 ("Gold"): Publish your article in an open-access journal if/when
    a suitable one exists.
    in BOTH cases self-archive a supplementary version of your article
    in your own institutional repository.
Received on Thu Jul 19 2007 - 21:22:40 BST

This archive was generated by hypermail 2.3.0 : Fri Dec 10 2010 - 19:49:00 GMT