- From: Jakob Voss <jakob.voss@gbv.de>
- Date: Wed, 02 Nov 2011 19:26:04 +0100
- To: Tim Hodson <hodson.tim@gmail.com>
- CC: "<ian.davis@talis.com>" <ian.davis@talis.com>, "<public-lld@w3.org>" <public-lld@w3.org>
Tim Hodson wrote:
> If the BL uses a vocabulary that imposes restrictions on how the term
> is used those restrictions might mean that linking to outside sources
> becomes more difficult, as the scheme definition starts to imply
> things about the data that are not meant.
>
> Much better is to start simply with the easy things to describe, but
> model the domain in a way that is extensible.
I fully agree. That's one reason why FRBR in RDF has not really started
yet: the current constraints make it difficult to reuse only parts of
FRBR. In particular we need general documents or works (as unspecified
as bibo:Book, and dct:BibliographicResource), single copies or holdings,
and particular editions. I drafted a lightweight ontology for
this purpose:
https://gist.github.com/1331983
> So if some organisation interested in the works of Jane Austen
> decides to produce a description of all her works as linked data,
> then there is a good chance there will be a uri for a single work.
> That uri would almost certainly not be described using any
> recognisably frbrish vocabulary.
I searched for Jane Austen's "Pride and Prejudice" in BL and found dozen
of URIs for it. Some are connected with owl:sameAs, but most
have a particular number of pages and other properties like ISBN.
For instance
@prefix blt: <http://data.bl.uk/schema/bibliographic#> .
<http://bnb.data.bl.uk/id/resource/009011483>
a bibo:Book, dct:BibliographicResource ;
dct:title "Pride and Prejudice" ;
dct:creator <http://bnb.data.bl.uk/id/person/AustenJane1775-1817> ;
dct:extend: "309p"@en .
This could either be a physical book, or an edition, or both, but it
would be against common knowledge to say that it is the same as the , so
they unlikely represent the general work "Pride and Prejudice". For the
latter we already have some URIs:
<http://www.librarything.com/work/2773690> a bibo:Book ;
owl:sameAs <http://dbpedia.org/resource/Pride_and_Prejudice> .
Someone else may have an URI for its single physical copy:
<http://example.org/mybooks/Pride_and_Prejudice> a bibo:Book ;
dct:extend: "309p"@en . # let's assume it's the same edition
For many applications we do not need to distinguish the three, but
<http://www.librarything.com/work/2773690>
owl:sameAs
<http://bnb.data.bl.uk/id/resource/009011483> ,
<http://example.org/mybooks/Pride_and_Prejudice> .
Is obviously wrong, isn't it? But how can you connect them?
How about (if BL URIs reference at least single editions):
<http://bnb.data.bl.uk/id/resource/009011483>
sobr:editionOf
<http://www.librarything.com/work/2773690> ;
sobr:exemplar
<http://example.org/mybooks/Pride_and_Prejudice> .
Cheers,
Jakob
P.S: This is the core of Simplified Ontology for Bibliographic Resources
(SOBR): Three non-disjoint classes:
sobr:Document a owl:Class ; owl:equivalentClass
schema:CreativeWork, bibo:Document, foaf:Document, frbr:Endevaour .
obr:Edition a owl:Class ; rdf:subClassOf sobr:Document ;
owl:equivalentClass [ a owl:Class;
owl:unionOf (frbr:Expression frbr:Manifestation)
] .
sobr:Item a owl:Class ; rdf:subClassOf sobr:Document ;
owl:equivalentClass frbr:Item .
--
Jakob Voß <jakob.voss@gbv.de>, skype: nichtich
Verbundzentrale des GBV (VZG) / Common Library Network
Platz der Goettinger Sieben 1, 37073 Göttingen, Germany
+49 (0)551 39-10242, http://www.gbv.de
Received on Wednesday, 2 November 2011 18:26:44 UTC