On Tue, 11 Feb 2003, D M Sergeant raises some frequently-raised points
that I think it is important to confront head-on:

The digital library community is very much concerned with preservation,
which is both commendable and a traditional responsibility of the
library community.

But there are two things about the rationale for the self-archiving
of refereed research that the library community keeps overlooking or
failing to understand, and as a result, the well-intended preservation
concerns of librarians are proving to be (unintentional) retardants to
self-archiving, instead of helping to speed it on its way.

The two things that librarians keep forgetting or overlooking are these:

(1) In the first instance, and for the time being, the self-archiving of
refereed research publications is not a *substitute* for existing forms
of publication and preservation, it is merely a *supplement* to them.

To put it more explicitly, the papers that researchers need to
self-archive (in order to maximise their research impact *now*) are all
still appearing, in parallel, in the traditional print journals and their
associated online editions. The librarians' preservation concerns and
initiatives should be focused on *those* continuing, primary, persistent
channels of publication. *That* is virtally where all the literature --
both in analog and digital form -- is. Their preservation concerns
should not be directed at the efforts to supplement those continuing,
primary, persistent channels of publication, through institutional
self-archiving.

The primary purpose of research self-archiving today is to remedy the
needless daily, cumulative research-impact loss that is occurring
because of toll-barriers that block access to this research for potential
researcher/users whose institutions cannot afford to pay the tolls to
access it: http://www.ecs.soton.ac.uk/~harnad/Temp/unto-others.html

Call that "filling the current access-gap." I hope this now makes it
more obvious that it is not the already-overdue supplementary measures,
intended to fill the current access-gap, that should be waiting for
preservation-problems to be solved, with self-archiving continuing to
be held back while we shop for future-proof self-archiving software!

Which brings us to the second point:

(2) If self-archiving had been held back -- pending digital
future-proofing -- by the physicists in 1991, then physics would have
lost the 12 years of rich access and impact provided by ArXiv during
that time: http://arxiv.org/show_monthly_submissions

For the 200,000 papers in ArXiv are still here today in 2003, still
being widely and openly read and used. ArXiv has since successfully
upgraded to OAI-compliance, and will no doubt continue upgrading its
contents to further usability standards as time goes on. Yet it is all
the other disciplines -- the ones that have *not* been self-archiving for
over a decade as the physicists have, the ones that have needlessly lost
another decade of potential research impact -- that are now being enjoined
by the well-meaning library community to pause [still longer] and
consider that:

> the wrong [software] choice may lead to a failure in the preservation.
> Other material is ergo being needlessly lost while ever it is not being
> preserved.

The library community is worrying about the "needless loss" of nonexistent
content -- content that (if only it had been self-archived!) would have
been but a supplement to its persistent primary incarnation, which is
today still in its publishers' proprietary analog and digital form and
not the object of any of this discussion -- while the research community
is still needlessly losing more years of potential research impact.

I would say that there was a certain incompatibility here between the
desiderata of the library community and the research community! Yet it
is all so simply resolved, if we simply remind ourselves that we are
talking here about immediate *supplements* to publication and existing
forms of preservation, not *substitutes* for them.

Note that the emphasis is on "immediate" rather than "delay" -- including
delays for the sake of future-proofing.

> How much do either [EPrints or DSpace -- or http://cdsware.cern.ch/]
> conform to the OAIS reference model?

How much do they *need* to (and why?), in order to provide many years
of enhanced access and impact to otherwise unaffordable research, *now*?

> It is unlikely that either [EPrints or DSpace] will be able to provide
> the full solution.

The full solution for what? The library community's *possible* long-term
concerns, or the research community's *certain* (and long overdue)
immediate ones?

Stevan Harnad