[OAI-eprints] Introducing the Subject Categorization discussion

[Stevan Harnad]

> On Thu, 16 Jan 2003, [utf-8] Hugo Fjelsted Alr=C3=B8e wrote:
>=20
> Provided that there are good abstracts, the existing Eprint search
> facilities work fine because they can search the abstract. And full
> text search engines, such as Google, can complement this. The problem of
> relying on Google for archive searches (Google searches can be restricted
> to a specific domain, such as www.orgprints.org) is that the timescale
> for updates is something like 6-8 weeks.

I think the google harvesting cycle is faster for well-connected sites
(days rather than weeks), and fast enough for the open-access literature
such as it is. As it grows, it might be decided to do custom
inverted-indexing for the entire interoperable research corpus, possibly
even by a customized google.

Stevan Harnad