> As far as I can see, there's no rdf:type specified for the
> resource <http://purl.org/dc/elements/1.1/>
Correct, though a type rdfs:Resource is rdf-entailed per the RDFS
entailments listed in the RDF Semantics WD [1].
> My understanding of the outcome of the long debate about
> rdfs:isDefinedBy was that the value can be _any_ resource
> which in some unspecified way or other "defines" the subject
> resource.
Yep. According to the RDF Schema for RDF Schema [2] the range and domain
for rdfs:isDefinedBy is rdfs:Resource, and the semantics presented for the
property [3] in the RDF VDL 1.0 WD are quite loose. Hmm... note that it
does say "indicates *the* resource defining the subject resource" [emphasis
added], which seems to imply that you can only specify a single
rdfs:isDefinedBy property for a resource. Might be worth an rdf-comments
mail, but I reckon it's a typo (should be "a resource").
> if we did specify a type for <http://purl.org/dc/elements/1.1/>,
> it would be a resource of type dcmitype:Text
<mild_aside>
I don't think so; i.e. not necessarily, and it might be bad practice to do
so. TAG have deferred the overriding issue (the range of HTTP [4]) that
overrides this discussion, but if the range of HTTP is decided (by what or
whom I now have no idea) to be some kind of generic document thingy as
TimBL proposes, I'm quite sure that it couldn't be dcmitype:Text--that
dcmitype:Text is disjoint with timbl:GenericDocument. If, on the other
hand, the range of HTTP can be anything, then you could call it text, but
then you'd never be able to change the representation of the schema :-)
</mild_aside>
But that's irrelevant: I don't think that you need to worry about
specifying an rdf:type for that resource. Why would it be useful to do so?
What's the justification?
> I _could_ have a non-RDF/XML document as the object of
> an rdfs:isDefinedBy statement: an HTML document, a text file,
> a graphic, even a non-Web retrievable resource. Is that correct
> please?
Yes, you could.
> leaving aside the issue of persistence, that second URL
> could equally well be used in the schemas: something like:
>
> <http://purl.org/dc/elements/1.1/title>
> <http://www.w3.org/2000/01/rdf-schema#isDefinedBy>
> <http://dublincore.org/2002/08/13/dces#> .
Yes. I'd take the hash off, though. I've no idea what <HTTP_URI#> denotes
for any of the possible MIME types that <HTTP_URI> could be served as. Cf.
[5].
> I'm _not_ suggesting we should do that for these schemas:
> I just want to clarify that my understanding is correct and it
> would be a valid use of rdfs:isDefinedBy ;-)
Yes, but don't confuse "valid" with "best" :-) I think that persistence is
the main issue that would bug me if this were a suggestion, but also the
fact that people come to expect the <namespace...term> conventions built up
on the semantic web. Conventions are just conventions, but they should be
followed unless there's a good reason not to, IMO.
Cheers,
[1] http://www.w3.org/TR/2002/WD-rdf-mt-20021112/#rdfs_entail
[2] http://www.w3.org/2000/01/rdf-schema
[3] http://www.w3.org/TR/2002/WD-rdf-schema-20021112/#ch_isdefinedby
[4] http://www.w3.org/2001/tag/ilist#httpRange-14
[5] http://www.ietf.org/rfc/rfc2396.txt The dreaded "The semantics of a
fragment identifier is a property of the data resulting from a retrieval
action, regardless of the type of URI used in the reference", RFC 2396
section 4.1.
--
Sean B. Palmer, <http://purl.org/net/sbp/>
"phenomicity by the bucketful" - http://miscoranda.com/
|