On Wed, Nov 12, 2003 at 09:33:29PM +0000, Andy Powell wrote:
> On Wed, 12 Nov 2003, Pete Johnston wrote:
> > So e.g. the current "namespace document" at
> > http://dublincore.org/2003/03/24/dces could include something like (I'm
> > not sure hasFormat is the right term here, maybe hasVersion is better...
> > anyway, leaving that aside for a moment...)
> >
> > <http://purl.org/dc/elements/1.1/> dcterms:hasFormat
> > <http://dublincore.org/2003/03/24/dces-fr> .
>
> I've never really understood if the W3C-recommended way to handle multiple
> language versions of resources is to assign each of them different URIs
> (as you suggest above with http://dublincore.org/2003/03/24/dces-fr) or
> whether they are all supposed to have the same URI
> (http://dublincore.org/2003/03/24/dces in this case) and that content
> negotiation should be used to ask for the preferred language version.
>
> I thought that W3C advice was to use content negotiation... but the
> obvious problem with this is that it is not possible to make RDF
> assertions about each langauge version unless different URIs are used.

Is there any reason not to use more than one URI to identify
a term?

DCMI currently identifies the term "Title" using the URI
"http://purl.org/dc/elements/1.1/title", which resolves to
"http://dublincore.org/2003/03/24/dces#title".

Additionally, the term declaration at
"http://dublincore.org/2003/03/24/dces#title"
has an additional arc ("hasVersion") pointing to
"http://dublincore.org/usage/terms/history/#title-004" --
a Web page where past versions of the term are documented.

Currently, the persistence of
"http://purl.org/dc/elements/1.1/title" is backed up by the
DCMI namespace policy, while there is no such support for
the URI "http://dublincore.org/usage/terms/history/#title-004".

While we certainly would not want to promote the latter
identifier for general use, it could be appropriate to use
for specific purposes, e.g., in assertions about a particular
historical version of a term.

This principle could apply to translated versions, as well
as to historical versions of those translated versions.

Getting back to your point: If we accept the notion that a
term can be associated with multiple identifiers, I see no
essential incompatibility between having an identifier for a
schema in French http://dublincore.org/2003/03/24/dces-fr and
using the generic identifier http://purl.org/dc/elements/1.1
together with content negotiation.  Whether one cited one URI
or the other would depend on what one wanted to say about what.

Tom

--
Dr. Thomas Baker                                [log in to unmask]
Institutszentrum Schloss Birlinghoven         mobile +49-160-9664-2129
Fraunhofer-Gesellschaft                          work +49-30-8109-9027
53754 Sankt Augustin, Germany                    fax +49-2241-144-2352
Personal email: [log in to unmask]