[rdfweb-dev] foaf:publications

Peter Mika pmika at cs.vu.nl
Thu Feb 23 13:27:44 UTC 2006


> sound like a argument for dropping RDF in favor of CSV ;-)

CSV is at least easy to write :)
 
> I'm not proposing to mandate content-negotiation, but URI-schemas like
> HTTP are very spread for naming foaf:Document and for this you may use
> content negotiation. Not only semweb oriented server like knobot, even
> apache supports it. You still may indicate the content-type of your list
> of publications, but this should not be constrained by the
> foaf:published property. If you prefer not use content negotiation and
> feel no need for "Cool URIs" you can have the following triples in your
> foaf-file:
> [ a foaf:Person;
>   foaf:published [ = <http://example.org/pubs.xhtml>; eg:mediaType
> "application/xhtml+xml"],
>                            [ = <http://example.org/pubs.bibtex>;
> eg:mediaType "text/x-bibtex"]
> ].

The property is called foaf:publications :)

Again, my feeling is that people (who have problems understanding the
current two-liner of a definition) will certainly not go into the trouble of
specifying content-types the way you suggest and/or configuring servers.

The implicit argument of trying to keep foaf simple (intentionally
underspecified) is something I have sympathy with. What I'm trying to argue
is that it would be important to support the possibility of pointing to
individual publications (which can be used for disambiguation) as well as
collections of publications (which can be used for presentation purposes or
to acquire information about individual publications after further
processing). So I would suggest minimally to refer to the foaf:made property
in the definition of foaf:publications as a way to point to individual
publications or even introduce a separate foaf:published property. For
pointing to BibTeX files I guess my approach will be to create a subproperty
of foaf:publications.

Best,
Peter





More information about the foaf-dev mailing list