[rdfweb-dev] Evaluating FOAF against Dublin Core Agent WG's draft Functional Requirements

[danbri at w3.org]

I'm at the Dublin Core Metadata Initiative's Advisory Board meeting, at
Shanghai Library.
Here, and in conference and hallways, I've had a few discussions about
person/agent description in the context of Dublin Core. There is an Agents
WG within DC, and for as long as I can
remember DC folk have been discussing how to 'do' agent stuff in Dublin
Core (ie. 7+ years).
I believe the FOAF scene has more experience than most on how to do this
stuff in RDF/XML, and
would like to start some dialog between folks here and in DC about
approaches that could be taken.

A good starting point is the draft functional requirements doc,
http://www.dublincore.org/groups/agents/
-> http://www.dublincore.org/groups/agents/agentFRdraft2-2.html

...and to start thinking about what those requirements might look like in
an RDF/FOAF context.
I think much of what DC is looking for can be done with FOAF and
extensions (eg. bio for dates) that people have designed around FOAF. The
biggest gaps are probably those around Groups.

The FOAF approach to identification,
http://rdfweb.org/mt/foaflog/archives/2003/07/10/12.05.33/ means that
we've a slightly different approach to IDs than a simple "ID" field.

Also btw see my recent weblog notes on DC-architecture work to clarify the
different ways in
which dc:creator dc:contributor dc:publisher are used;
http://danbri.org/words/?p=63 ...as DC tightens those things up, it might
be that we find less need for foaf:maker and foaf:made. At the
Architecture-WG meeting there seemed to be some enthusiasm for not
allowing rdf:Seq, rdf:Bag, rdf:Alt constructs to be used with dc:creator
etc., which'd simplify things a bit.

So I'd encourage folks here to take a look at the functional requirements,
esp as we think about
extensions to, and refinements of, the contents of the FOAF vocabulary.
Also of course it might
inspire people who are working on extensions and applications...

cheers,

Dan