On Wed, Nov 10, 2004 at 09:01:07PM +0100, Marcello Perathoner wrote:
Greg Newby wrote:
I don't want to over-specify how I think the workflow should happen. I think that's still to be determined. But the overall flow needs to be somewhat circular: librarians need to import existing PG catalog records, preferably in MARC format, to existing software. (Alev has a couple of programs for this; PGLAF can probably acquire software for other folks who'd like to work a lot on this activity.) Then, updated records would need to be shipped back into the catalog.
I think an easier solution would be to build an ASCII list containing the etext-number and the LoC Call Number for all etexts we have.
We would then import the LoC Call Number into a field in the database.
The catalog software could then update a number of fields (Subject, LoC Class, Unified Title) automatically from the LoC database (TODO Check copyright status of LoC database !!!)
Then we could do a manual pass over the database with the MARC record at hand and fix the author / coauthor attributions, link into wikipedia if an article exists, add summaries etc.
I like this idea, but am concerned that there will still need to be human oversight. Just importing records will only work if there are unambiguous matches, and it seems that matching is often ambiguous.
From doing lots of copyright clearances, I know that many items are not in the LoC database (most of our non-English is not in there). But this would be a good start, and there are other national library catalogs that offer Z39.50 access to their records. -- Greg