Re: Linked Library Holdings/Items

On Fri, Oct 21, 2011 at 1:11 PM, Jakob Voss <jakob.voss@gbv.de> wrote:

> So how you relate <http://dbpedia.org/page/The_Diary_of_a_Young_Girl>, which
> is not in FRBR, for instance to
>
> http://uri.gbv.de/item/opac-de-b1549:epn:878552081 (its frbr:Item)
> http://lobid.org/resource/HT008722600 (frbr:Manifestation of the Diary)
> http://www.librarything.com/work/3032251 (its frbr:Work)
>
> Or is this not the use case of ov:commonThing ontology?

No, that's exactly the use case.

So, rather than use the document URI, let's use the non-information
resource uri (http://dbpedia.org/resource/The_Diary_of_a_Young_Girl).

I think the most granular frbr entity you can hope for in that
resource is a Work (there are multiple, conflicting, expression
statements: language, englishPublicationDate, etc.), which is probably
generally appropriate for dbpedia resources.

So, in this case:

<http://dbpedia.org/resource/The_Diary_of_a_Young_Girl>
  <ov:commonWork>
<http://uri.gbv.de/item/opac-de-b1549:epn:878552081>,
<http://lobid.org/resource/HT008722600>,
<http://www.librarything.com/work/3032251>,
<http://openlibrary.org/works/OL2931460W> .

So this at least states that all 4 resources are referring to the same
frbr:Work (whether it's called "Diary of a Young Girl", "Journal
d'Anne Frank", or the "Diary of Anne Frank").

ov:commonWork is pretty broad, so:

<http://lobid.org/resource/HT008722600>
  <ov:commonWork> <http://openlibrary.org/books/OL6120127M>,
<http://purl.org/NET/lccn/88195896#i>,
<http://dilettantes.code4lib.org/LODThing/isbns/0670916676#book> .

A work doesn't ever have to be explicitly mentioned.

I *think* you mentioned your daia:Item is for the equivalent
Manifestation in your catalog as hbz's (same edition, etc.), so in
that case, you can:

<http://lobid.org/resource/HT008722600>
  <ov:commonManifestation>
<http://uri.gbv.de/item/opac-de-b1549:epn:878552081> .

What you're trying to do, rather than explicitly model the frbr
hierarchy, take the data you have and visualize the frbr stack that
would be represented in it.  Then you make your (simple - no
derivativeWorks and the like) FRBR relationships between them based on
that implied frbr stack.  If the FRBR entities ever *do* get formally
modeled, then the ov:commonThing properties can get mapped to that, as
well (since there's no other obvious way to link a bibo:Book or
dct:BibliographicResource to the FRBR entities).

I figure the most useful property is ov:commonWork, since it is the
easiest to imagine the implementation (something like OCLC's
x-identifier service, for example).

Hopefully, that will clear it up a little better.
-Ross.

Received on Friday, 21 October 2011 13:29:29 UTC