[rdfweb-dev] XSLTs for FOAF, Spring v1.3.1 and plans for FOAFspec improvements

Morten Frederiksen mof-rdf at mfd-consult.dk
Tue Jun 24 18:27:17 UTC 2003


On Tuesday 24 June 2003 20:09, James Carlyle wrote:
> > Any chance you could modify it to handle all of the current
> > specification, outputting in a (to be defined) canoncial format?
> I am happy to do this - personally I feel that given the availability of
> XML processing in almost all languages, a canonical XML format is going to
> be easier to work with than plain text Triples.
Indeed, and a seperate transformation from the canonical format to NTriples 
(or N3, for that matter) shouldn't be hard to do. Toolchaining is nice. :)

> But what do you mean by canonicalisation?  For example, one form (the
> intermediate RDF XML form that Max worked on) is unspecific in terms of RDF
> statements, but with all of the shortcuts specified in the XML
> serialisation standard removed.  This canonical form clearly belongs to the
> RDF core.  Or canonicalisation in the context of FOAF might mean shaping
> FOAF specific statements (i.e. knows, Person etc, with a defined and
> consistent nesting structure) and removing other non-FOAF statements.
Having tinkered with this quite a bit, I think that a general minimal RDF/XML 
subset would be preferable - I imagine it's not only in the realm of FoaF 
that this could come in handy.

The subset would have to support the complete RDF model, but without the 
variations in the RDF/XML syntax, and yes, that's supposedly an RDF Core 
issue, but not one I'm sure they'd just pick up for their to-do list?

What I would prefer, is a simple listing of non-repeating resources with 
mandatory rdf:about or rdf:nodeID (perhaps sorted by URI/ID), with no nesting 
and no typed nodes - even though the latter leads to some quite ugly XSLT, 
having to use more elaborate XPath expressions. Used with xslt:key it could 
be quite efficient as well.

I should go check out the productions in the syntax, but it will have to wait 
a little while, too swamped at the moment...


Regards,
Morten



More information about the foaf-dev mailing list