[rdfweb-dev] RDF all the way on XML.com

Dan Brickley danbri at w3.org
Sat Aug 23 12:07:05 UTC 2003


* Nick Gibbins <nmg at ecs.soton.ac.uk> [2003-08-23 11:14+0100]
> 
> Victor Lindesay <victor at vicsoft.co.uk> writes:
> 
> >> One way out of this dilemma is for a foaf-like standard to 
> >> appear that 
> >> is constrained to be XML hierarchical and which is a superset of 
> >> foaf-RDF. An Atom or RSS 2.0 to FOAF's RSS 1.0.
> 
> > This seems like a good idea. Perhaps with RDF support built into the XML
> > Schema so it can be consumed as XML or as RDF. This would satisfy
> > everyone, surely.
> 
> There's an existing precedent for this approach; the Gene Ontology
> Consortium, PRISM and RSS 1.0 all effectively have simplified profiles
> of RDF/XML for the reasons you give.
> 

The problem being that they all have different simplified profiles. Also 
that this technique works best for data formats that are basically 
homogenous, and not open to unexpected extensions of any substance.

I think we should do a simplified syntax for FOAF, but do it as an XHTML 
profile, ie. a convention for annotating your homepage markup with
enough info to allow RDF to be extracted. Dan Connolly showed a design
for this 3 years back, 
http://rdfweb.org/pipermail/rdfweb-dev/2000-July/010125.html based on my 
wishlist from the 1st message to this list. Probably time to revisit the
idea!

I am less keen on a syntactic profile of RDF for FOAF, since it will
make things a lot harder from the RDF side of the fence: instead of just 
writing out RDF/XML with an off-the-shelf RDF serializer, we would need 
custom serializers that knew the FOAF-specific rules. This could get 
quite fiddly if extensions were being used. I believe folk have had
similar frustrations with RSS1.

I am, on the other hand, interested to try out alternate RDF syntaxes on 
FOAF. For example, see http://esw.w3.org/topic/SpotOfDrama for
experiments with unstriped syntax. (I think I circulated that already)
These are completely general, and should work for all RDF, not just
FOAF, so seem to me to be a better use of coding and spec-writing time.

Dan



More information about the foaf-dev mailing list