On 5/30/12 12:37 PM, Jonathan Rochkind wrote: > On 5/30/2012 2:30 PM, Karen Coyle wrote: >> >> Basically, using the ISBN as an identifier is a bit of a mess. Not to >> mention that the ISBN only came into being in the late 1960's, so it >> doesn't apply to any books published before then. > > Totally, but so much less of a mess than NOT using ISBN's at all!
Yes, I agree. I think your comment below about not expecting them to be "canonical" hits the mark. They are good hints about sameness, but require some fussing to get what you are aiming at. kc > > I do a lot of work trying to tie together bibliographic data between > different databases from different vendors -- and when there's ISBN's in > both data sets, things become at least _possible_ that are barely > possible when there aren't. > > Same with OCLCnums, but ISBNs are more common. > > The ontology of the bibliographic universe is complicated -- the "things > identified" by identifiers issued by someone else are often going to be > not _quite_ the way you want the domain in your ontology to work. > > ISBN is _sort_ of a manifestation identifier -- if you keep in mind the > book industry considers hardcover/paperback/other-binding to be a > different manifestation (which is legit) while the library world does > not (which is also legit). > > OCLCnum can sometimes be used as a sort of manifestation identifier -- > if you keep in mind that WorldCat increasing has multiple records (with > multiple OCLCnums) for different "language of the cataloger" even for > the same manifestation -- AND that the library world considers paperback > vs hardcover to be the same manifestation. > > Still, between working with data that has ISBNs or OCLCnums vs working > with data that does not --- it ends up being a HUGE advantage to what > software can feasibly do with the data, and linking between different > corpuses (corpi?). > > ISBN might not be suitable for your internal 'canoical' identifiers, if > your ontology doesn't exactly match the book industries (and I don't > think it should) -- but definitely keep ISBNs _around_ if you can, and > keep track of what ISBNs correspond to which of your internal entity > instances, if you can. And allow lookups in your API's by ISBN, etc. > Same for OCLCnum. Significantly increases the utility and value of > bibliographic data. > _______________________________________________ > Ol-tech mailing list > [email protected] > http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech > To unsubscribe from this mailing list, send email to > [email protected] > -- Karen Coyle [email protected] http://kcoyle.net ph: 1-510-540-7596 m: 1-510-435-8234 skype: kcoylenet _______________________________________________ Ol-tech mailing list [email protected] http://mail.archive.org/cgi-bin/mailman/listinfo/ol-tech To unsubscribe from this mailing list, send email to [email protected]
