Re: Resource references/httpRange-14 (was: [Minutes] 1 July TAG teleconf)

At 10:32 AM 7/2/02 -0400, Jonathan Borden wrote:
>Paul Grosso wrote:
>
> >
> > But can an application such as RDF redefine the semantic of
> > an embedded URI reference in such a way that it predetermines
> > (overrides) the MIME type of the referenced resource regardless
> > of how the resource is labeled?
> >
> > And if it can't, then RDF cannot redefine how to interpret the
> > fragment identifier; rather, the fragment identifier must be
> > interpreted according to how the MIME type says it should.
>
>Presumably the "referenced resource" doesn't itself have a MIME type, rather
>the MIME type is assigned during the process of returning a representation
>of the resource back to the user agent which is resolving a URI.

Quite;  that is what I was trying to say.

>In RDF, there is no such URI resolution going on. Given an arbitrary URI
>reference which is just sitting there (i.e. as a string) how can one
>interpret the fragment identifier w.r.t the URI when there is no MIME type?
>(Remember: the URI is not being resolved.)

Yes.  If one presumes an RDF representation, that provides the 
MIME-type-relative interpretation according to rules of the RDF MIME 
type.  And if an RDF document representation is retrievable at the 
indicated URI-without-fragment, then it should be regarded as the defining 
document.

#g


-------------------
Graham Klyne
<GK@NineByNine.org>