> Though, I think DCMI and its users would be better off in the
> long run if the present statement that "http://purl.org/dc/elements/1.1/"
> is a "namespace" simply means that it functions as an XML namespace
> name, but no commitment is made regarding its actual denotation.
And to preemptively answer a question that the above is sure to
raise: yes, even if the above URI is not used to denote a vocabulary
of terms, DCMI can still have the URI resolve to some body of
information which can help DC users understand how best to
use the DC terms (e.g. if a user decides to try to dereference
the namespace name URI, regardless of the fact that the
specs don't require that it be resolvable to anything).
Thus, just because DCMI provides a representation of "something"
accessible via a URI used as a namespace name, that does *not*
obligate DCMI to say what those URIs used as namespace names
actually denote (the fact that they resolve to a representation
indicates that they denote "something", but insofar as the
web infrastructure is concerned, DCMI doesn't have to
say what that something is ;-)
So DCMI can continue to provide information (e.g. an RDF schema)
via the URIs used as namespace names, as it does at present, while
still saying nothing whatsoever about what those URIs denote. This
is OK because the web architecture (layer) is fully agnostic about
the actual denotation/meaning of URIs. And if there is, otherwise,
no need at the semantic web layer to say what those URIs might mean
(and I don't see any need at present to do so), then don't bother.
DCMI is then free to mint new, distinct URIs which are used to
denote the various vocabularies -- and these URIs might very well
resolve to the very same RDF schema(s) as the URIs used as namespace
names; which is fine, since there is no reason why a particular
representation (the RDF schema instance) can't be used as a valid
representation of more than one resource ;-)
Later, should it be deemed useful/necessary to do so, DCMI can
state what those URIs used as namespace names might denote, and
the nature of the resources denoted (and if necessary, adjust the
particular representations provided via those URIs), but that's
probably best left until the whole "namespace" and
"namespace document" business is worked out (if ever).
Cheers,
Patrick
--
Patrick Stickler
Nokia, Finland
[log in to unmask]
|