On Tue, 12 Feb 2002, Steve Hitchcock wrote:
> The situation as far as author-archived eprints are concerned is that there
> is one large disciplinary archive (arXiv) and a large distributed archive
> (RePEc). There are other smaller archives, some listed on the OAI site, and
> we are beginning to see the emergence of fledgling institutional archives.
> Apart from the former two, few have demonstrated longer-term stability.
I'm not sure what Steve Hitchcock means by "stability" and by
"longer-term."
Arxiv has existed for over 10 years; I know of no older archives than
that.
The OAI standard
http://www.openarchives.org has only been in existence
since 1999, and Eprints
http://www.eprints.org has only been available
since 2000.
So it is not clear what time-base Steve Hitchcock has in mind when he
says "longer-term."
I also don't know what he means by "stability" (if it is not just
synonymous with "longer-term," which in turn simply means "how long the
archive's been around").
Please let us not confuse (OAI) ARCHIVE stability and durability with
the durability of a PAPER or a FILE or a PERSONAL WEBSITE URL on the
Web.
It seems to me the archives registered at
http://www.openarchives.org
and/or
http://www.eprints.org have proved as stable and durable as one
could hope over the short (2-year) time-span they have existed. I am
not aware of any of these archives collapsing or vanishing (is Steve?).
What would be welcome, though (and perhaps this is what Steve had in
mind?) would be if those archives were filling faster and fuller. But
that's another matter.
The technical stability of Eprint Archives is a technical matter, and
for that we have to turn to technical masters like Chris Gutteridge at
http://www.eprints.org
The durability of Eprint Archives is an institutional policy matter, and
for that we have to look to serious, sensible policies from the
universities that create them in order to provide access to their
refereed research output. As it is very much in the interest of these
institutions to provide reliable, stable, long-term access to their
research, I do not think there are grounds for much worry on that score
at the moment.
> There are also very large numbers of papers posted individually by authors.
> ResearchIndex (RI) http://citeseer.nj.nec.com/cs is predicated on this
> observation. In contrast we have to bear in mind the near collapse of
> another computer science archive service, NCSTRL (see Krichel and
> Warner http://openlib.org/home/krichel/sants.html).
NCSTRL was not an archive, as far as I understand it, but a portal to
archives. In particular, a big chunk of NCSTRL was itself Arxiv!
> Research Index (RI) harvests the texts not just the metadata and is
> working, in terms of rights, in a grey area (as is Internet Archive,
> for example). What this suggests is that we need flexibility to allow
> services such as RI and other collective services that might build on,
> or save, the smaller archives and individual author pages, to operate.
Reasearch Index
http://citeseer.nj.nec.com/cs is superb, but I have
the opposite sense of what is needed! It is not RI that is needed to
preserve the contents of distributed archives. (That needs to be done,
but there are much better ways to do it, building on their
OAI-compliance and interoperability.) Rather, what is needed is to
get the contents of RI into a registered OAI-compliant Archive (or
Archives)!
On the other hand, harvesters and collectors and services building on
the distributed OAI-compliant Archives are desirable and welcome too.
> Is there an archive, large or small, that could not benefit from a RI-like
> service?
It is the research papers scattered willy-nilly all over the Web that
benefit from RI-like harvesters. OAI-compliant Eprint Archives are
already in a much stronger position, and can be served by OAI-based
harvesters and search engines like Cite-Base
http://cite-base.ecs.soton.ac.uk/help/index.php3 rather than needing
the nuclear power of RI to trawl and process all arbitrary websites and
files on the Web.
> As the self-archiving movement beyond arXiv makes its tentative steps
> forward it needs to think about supporting rights that enable autonomous
> services to build on the potentially vast array of archives and standards,
> and to save archives in the case of disasters if necessary.
That is cerainly true, but for interoperable, registered, OAI-compliant
Archives, the technical preservation problem will be much more
tractable. I also think the "rights" issue is a red herring at this
point. What is needed is (much) more content archived, not more
a-priori hair-splitting about "rights" for this author give-away
content.
Stevan Harnad
Received on Tue Feb 12 2002 - 13:15:32 GMT