> Yes, I agree. We only need the dcterms:URI DC encoding
> scheme - the URI scheme and the structure of the URI tell the
> processing application everything else it needs to know.
... and further in many (maybe not all, but many) cases where URIs
appear in representations of DC metadata, we won't even need to deploy
the dcterms:URI encoding scheme to indicate that those strings are in
fact URIs. Their "URI-ness" will be signalled by some syntactic
mechanism which indicates that those strings are "resource URIs" or
"value URIs", _not_ (IMHO) by the use of the dcterms:URI encoding
scheme.
As I think Douglas noted, RDF/XML does this perfectly well through
syntactic constructs like rdf:about and rdf:resource, and you rarely
(never?) see a reference to the class http://purl.org/dc/terms/URI in
RDF.
But the current DC-in-XML syntax does _not_ do this satisfactorily, and
that is one of the things that needs fixing.
Pete
|