[openbiblio-dev] Metadata format for Bibliographica

[William Waites]

Hello Primavera,

There is no detailed specification. However we are using the
bibo ontology for books/works simple dublin core for the main
metadata and foaf for people, so an entry might look like this:

<http://bibliographica.org/entry/abc123> a bibo:Book ;
    dc:title "my first book of abcs";
    dc:contributor <http://bibliographica/entity/def456>;
    dc:publisher <http://bibliographica.org/entity/ghi789>;
    dc:subject [ rdfs:label "letters";
    bibo:isbn "1234567890".

<http://bibliographica.org/entity/def456> a foaf:Person;
    foaf:name "big bird";
    bio:event [ a bio:Birth; bio:date "1970"^^xsd:gYear ].

<http://bibliographica.org/entity/ghi789> a foaf:Agent;
    foaf:name "big publisher".

series and collections and such are handled with dc:isPartOf.

There is also linkage between the "abstract entity", the
book, author, publisher URIs above to the documents that contain
their representation using foaf:primaryTopic.

A "live" example is http://bnb.bibliographica.org/entry/GB8514494.n3

Some care must be taken with creating records, and which graphs
they live in, we store one graph for each book and one graph
for each author, publisher and duplicate some data between them.

The simplest way to create information is to just send the 
serialised RDF graph via HTTP POST to the appropriate URI. The
way that this is done, and the automatic generation of URIs
will be shortly revisited, but the basic mechanism will remain
the same.

All of this said, the "best" way to represent this kind of data
in RDF is still very much an open question. The W3C incubator
group for the topic will finish soon and is likely to spin off
other working groups to try to hammer out some standards and 
best practices, but at the moment there is no real standard 
for this.

So if Robert Myers is interested in modelling, and already has
RDF data in Freebase, I would suggest just to think on the best
representation that Freebase can have. Then we take the best
attributes of that and improve our model, and write some 
inference rules to convert back and forth. This is worthwhile
in the case of Freebase because there is a lot of data there
that we may benefit from. After some more discussion and when
this settles, this kind of functionality can be simply exposed
as a button in the UI that says, "import from freebase" which
will understand Freebase's conventions and run the inference
rules to do the transformation. Because the inference rules 
are written down in a standard form, N3, we can publish those
easily enough as well and would indeed include links to them
in the provenance information for our record along with links
to the source.

For interest, not that Mr. Myers has to worry about this, 
such a block of provenance information would look like 
this:

<http://bibliographica.org/entry/abc123> a opmv:Artifact;
    opmv:wasGeneratedBy [
        a opmv:Process;
        opmv:used <http://bibliographica.org/rules/freebase.n3>,
                  <http:/rdf.freebase.com/...>;
        opmv:wasControlledBy <http://bibliographica.org/account/user...>;
        ....
    ].

Hope this helps,

Cheers,
-w


* [2011-03-16 20:34:49 +0100] Primavera De Filippi <primavera.defilippi at okfn.org> écrit:

] Hi all,
] I have been talking with Robert Myers concerning possibility to integrate
] metadata from Freebase into PublicDomainWorks.net.
] Robert is grabbing information on visual artists and their works out of the
] freebase database, and we would like to make this information accessible to
] PublicDomainWorks.net.
] Right now, this can only be done by re-formatting freebase metadata into the
] format currently used by PublicDomainWorks.net.
] 
] However, after a discussion with Rufus and Tatiana, we have agreed that the
] whole PDW database should be eliminated, and the interface should be
] re-coded so as to go and query directly the Bibliographica database. This
] operation will be done in the following months.
] 
] Before we can do this, we need the technical specification of the format
] used by Bibliographica for their database, so that we can (1) import new
] data directly into Bibliographica, and (2) implement the code for
] PublicDomainWorks to directly query Bibliographica database.
] 
] If anyone could provide a proper documentation on the format used for the
] metadata and the procedure to insert it into Bibliographica, that would be
] really helpful !
] 
] Looking forward to your answer,
] Primavera

] _______________________________________________
] openbiblio-dev mailing list
] openbiblio-dev at lists.okfn.org
] http://lists.okfn.org/mailman/listinfo/openbiblio-dev


-- 
William Waites                <mailto:ww at styx.org>
http://river.styx.org/ww/        <sip:ww at styx.org>
F4B3 39BF E775 CF42 0BAB  3DF0 BE40 A6DF B06F FD45