Quoting "Jacobs, Jane W" <Jane.W.Jacobs_at_QUEENSLIBRARY.ORG>:
>
> I'm guessing that the problem may have come from trying to automate the
> match of the (electronic) item to the data. If you didn't grab both
> item and metadata at once and tie them together with an ISBN or some
> sort of control number, linking the electronic item to ANY metadata
> would require that you re-extract enough data from the item to search
> for a matching metadata record.
This has all been worked out. Google receives files of MARC records
from the participating libraries that include the barcode numbers on
the library books. As the book is scanned, the barcode number on the
physical piece is matched to the record in the file submitted by the
library. Believe me, Google has the library's MARC record for the item
in hand.
Rather I
> expect that Google programmers were not knowledgeable about data quality
> to grab the best available or tie it to the item where they ought to
> have done so.
Although I suppose this is possible, it does not seem plausible to me.
1) with all of its billions, Google could easily hire a librarian, and
we do not know if there is anyone on the GBS staff with library
cataloging experience 2) Google is working closely with its partner
libraries, who do have the knowledge of the library data, and have an
interest in Google's use of metadata because they expect to serve
their (the libraries') users through the subscription service that
will be provided by Google 3) libraries like Michigan already have the
Google scanned books in their catalogs associated with the original
full MARC data 4) the Open Library is using Z39.50 to retrieve full
MARC records for all of the Google-scanned public domain books in
order to add them to the catalog. I don't think there are any excuses
for not having good metadata.
Karen Coyle
kcoyle@kcoyle.net http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet
Received on Fri Sep 18 2009 - 12:02:24 EDT