Thomas,
Just on this point
> The other reason is that for me it is not immediately obvious
> if this defines the same identity as the "Namespaces in XML
> 1.0 (Second Edition)"
> (http://www.w3.org/TR/REC-xml-names/), which probably would
> be desirable.
>
> In any case, the latter document deprecates relative URIs for
> namespace declaration, and given the rigidity of the
> identification, relative references (w.r.t. the base of the
> URI of the given document) as attribute values might cause
> problems for URIs that are used as resource identifiers as
> well. But this probably can be left to the consideration of DC-XML.
Yes. Any mapping between URIs and XML expanded names or XML QNames is
the business of any XML formats which might make use of such mappings -
in much the same way that for RDF it is handled at the level of the
RDF/XML spec and specs for other XML formats for the serialisation of
RDF graphs. And in the previous drafts for DC-XML we tried very hard to
be careful to explain this, as it's an area which has been the source of
an awful lot of confusion in the past.
But, yes, that is a concern for any specifications which make use of
such mappings, not for the Abstract Model.
Pete
|