On Wed, 13 Feb 2002, Rachel Heery wrote:
> I for one think it is vital that DCMI can reach consensus on a
> comprehensive and accurate expression of the semantics of the DCMI
> vocabulary in RDFS (and other syntaxes).
Yes, agreed.
One immediate problem is that (as far as I can tell) the namespace URI for
the DCTERMS namespace
http://purl.org/dc/terms/
currently resolves to
http://dublincore.org/2001/08/14/dcq#
Unfortunately, this schema is incorrect :-( For example it contains
entries for 'note' and 'release', neither of which have ever been agreed
by DCMI (as far as I recall).
I'm not sure when or how this schema got put in place but I would strongly
suggest that the purl.org resolver configuration is changed so that the
DCTERMS namespace URI does *not* resolve to this schema. Please can we
make the DCTERMS namespace URI not resolve to anything until we have
agreed what it should resolve to. This should happen now (or as soon as
possible). Apologies if I missed a decision somewhere!?
I recall lots of discussion about what the DCMI namespace URIs should
resolve to (RDDL vs. RDF Schemas vs. nothing)... I don't recall any
agreement that they should resolve to incorrect RDF schemas :-)
I think there are similar problems with the schema that the DCMITYPE
namespace URI resolves to
http://purl.org/dc/dcmitype/ -> http://dublincore.org/2001/08/14/dctype#
This schema contains the term 'Resource' which has never been approved by
DCMI.
Again, this resolution of the namespace URI should be removed as soon as
possible.
> A schema that defines only qualifiers (as opposed to a schema that
> attempts to define all terms in the terms namespace, which are a mix of
> qualifiers and 'new elements') may be most appropriate if we accept that
> DCMI schemas should express the relationship between qualifiers and
> elements. And if the resulting schema achieves this.
Why is a schema that only includes qualifiers most appropriate? Can't a
single schema express the relationships betwgeen qualifiers and elements?
The namespace spec says that 'All DCMI namespace URIs will resolve to a
machine-processable DCMI term declaration for all the terms within that
namespace.' If we assume that RDF Schemas provides the machine-processable
declaration, then we have to cover all terms, at least for that particular
representation.
Andy.
> ---------- Forwarded message ----------
> Date: Wed, 13 Feb 2002 16:51:50 -0500
> From: Eric Miller <[log in to unmask]>
> To: [log in to unmask]
> Subject: Re: Updated charter and WG web page
>
> At 09:07 PM 2/13/2002 +0000, you wrote:
>
> >I think this perhaps illustrates Rachel's point about controlling the
> >content of the RDF Schemas indexed by the registry - and that is equally
> >vital whether they are static text files edited "by hand" or files
> >generated from a database.
>
> I absolutely agree with Rachel and Pete about this, but I think we have
> demonstrable proof of creeping into out-of-topic discussion threads :). I'd
> appreciate feedback from the chairs if this is in scope or out. Seems to me
> the focus of this group should be on registry requirements (which are
> largely content independent).
>
> If its the view of the chairs to have this discussion here (rather than
> dc-usage, etc.) I took a crack at re-representing the DCQ document in a
> more machine processable manner... if appropriate, comments are welcome.
>
> - http://www.w3.org/2001/11/26/dcq
> -
> http://www.w3.org/2001/10/navigate/view?subject=http%3a//www.w3.org/2001/11/26/dcq%23
>
> If the chairs feel that this should discussion should be done elsewhere
> they have my permission to forward this message to the appropriate list.
>
> --
> eric miller http://www.w3.org/people/em/
> semantic web activity lead mailto:[log in to unmask]
> w3c world wide web consortium tel:1.614.763.1100
>
Andy
--
Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK
http://www.ukoln.ac.uk/ukoln/staff/a.powell +44 1225 383933
Resource Discovery Network http://www.rdn.ac.uk/
|