Re: Renewed RDF/XHTML task force - first telecon

Hi Mark,

* Mark Birbeck <mark.birbeck@x-port.net> [2004-08-10 20:52+0100]
> Hi Dan,
> 
> I've read the IRC log of the telecon, and here's a few comments:

Thanks!


> 1. Normative reference to RDF.
> 
> I've already been given a task by the WG to draw out the relationship
> between our syntax and RDF, so don't worry about that. However, whilst we're
> on the subject, I'd be interested to hear a bit more detail about the
> comment that the examples "are a bit '98". That's probably because my
> history with RDF is from that time, but I then moved on to other things
> since it seemed to be moving so slowly ;).

Ah I was hoping that wouldn't cause offence! What I meant was just that
in the early, idealistic/naive RDF years, we sort of accidentally gave
the world that the "RDF world view" was that everything would have a
URI, and that everyone would know the URIs for everything, and that
everytime something interesting was mentioned in RDF/XML it would have
some kind of URI (including URN, etc) there. This turned out not to be
the way things went. In the 2nd wave of RDF specs, the earlier vague 
notion of a blank (or 'anonymous', in 99-REC terminology) node in the
graph was tightened up, and folks who implemented to the RDFCore specs 
began to feel more comfortable with 'blank' (URI-free) nodes in their
RDF/XML graphs.

So "98-era" is shorthand for "URI-heavy", specifically a tendency to rely
on idioms that assume URIs for things like people, companies etc. will
be well known. In the last couple of years (especially since OWL
formalised some supporting concepts, such as 'inverse functional
property' and 'functional property') you tend to see more RDF/XML which
omits URIs for things that don't have widely known URIs, and instead
supplies enough RDF description (via properties such as 'homepage',
'stockTicker', etc) to identify the thing via a mini-description.

Hope this helps. I wrote about this a bit last summer in the context of
FOAF, http://rdfweb.org/mt/foaflog/archives/2003/07/10/12.05.33/ and 
more recently in a www-rdf-interest thread, summarised helpfully on
XML.com last week by Edd Dumbill:
http://www.xml.com/pub/a/2004/08/04/deviant.html

> 
> Anyway, please feel completely free to make any comments you like that will
> bring my understanding up-to-date, and in turn make the examples and prose
> in the XHTML 2 spec better. One thing that stands out is that you seem to be
> using FOAF for a lot of your examples, so we could easily change ours to
> whatever the latest hip thing is ...

FOAF's handy 'cos there's a lot of real data out there that uses it, and
it is designed with the 'reference by description' idiom in mind. But it
would be healthy to have examples using multiple vocabularies, since
that's one of RDF's selling points, and there are a good number of
vocabs now. 

http://www.schemaweb.info/ has a nice directory of
schemas/vocabs/ontologies or whatever they're called these days. I think
one "selling point" for mentioning an explicit relationship to RDF in
the (excellent, btw; did I forget to say that? :) XHTML 2 text is that
we could then say "there are a number of pre-existing property sets /
vocabs you could use - go look <here/>"...

Gotta run; reply to rest later.

cheers,

Dan

Received on Tuesday, 10 August 2004 21:03:31 UTC