I'm working on yanking my bibtex parser out of Haystack for inclusion into
the Simile RDFizers. I've found three different Bibtex ontologies on the
web with namespaces:
*
http://www.ontoweb.org/ontology/1
*
http://www.isi.edu/webscripter/bibtex.o.daml
*
http://purl.oclc.org/NET/nknouf/ns/bibtex
Of these, the first appears to be the defacto "standard", due mainly to
the existance of a perl-based bibtex-to-rdf translator, which can be found
at
http://www.cs.vu.nl/~mcaklein/bib2rdf/ .
This tool makes a couple of design decisions: it reifies persons and
institutes (losing author order); assigns author/institution URIs relative
to the file URI; and finally it strips bibtex character formatting
(\lambda becomes lambda).
As it stands, my tool is fairly similar in its approach, except it uses a
Haystack ontology (I'm going to change this, but it won't be too hard to
change it again). However, I think it would be better to approach
rdfizing bibtex in two stages: first create RDF in an ontology that is as
close to bibtex as possible and then second reify authors, institutions,
conferences, journals, months, years, etc. It would be kind of cool to
have an ontology that could encompass the idea of a "conference series"
and a "conference" as seperate entities.
Nick
Received on Tue Nov 02 2004 - 16:55:37 EST