Jay,
Combined responses to several subthreads:
I think the database is a great idea. I'll join the new list to
discuss it when its created.
There are a number of existing data formats which are at least related
to the concept. MARC was already mentioned. It's capable of doing
this, but is indeed very heavy-weight. It's also a transfer format;
relatively little software stores pure MARC in database form for
routine processing. OTOH, as a data format, it's clearly on topic,
originating over 30 years ago on mainframe systems. :)
Another one we should probably look at is Encoded Archival Description
(EAD), for ideas if nothing else. EAD is XML-encoded, which probably
makes it unsuitable in the minds of many, but it was designed by
professional archivists, who do this sort of thing for a living.
Whatever data format is selected, it'd be nice to be able to identify
duplicate holdings and provide multiple links to the searcher, so if
one source is down or gone, another is available.
The OAI Harvesting Protocol is lightweight and should probably get a
serious look as the mechanism for gathering this data into the central
database. It was designed for exactly this kind of application.
We might consider replicating the central database as well; perhaps
not live for the public, but having the software and data ready to
go in a few other places.
De