Re: Fragment identifiers and intermediaries

Mark Baker wrote:

>
> On Fri, Jul 26, 2002 at 04:04:18PM -0400, Christopher B Ferris wrote:
> > No, it is not.
>
> Sorry, I thought it was obvious from the context of this discussion
> that we were limiting ourselves to RDF.
>

If that is the case, then when RDF says "resource" read "thing" and
URIreferences are perfectly good opaque identifiers for "things". I see the
architectural issue, but realize that RDF itself (in isolation) has no
problem to the extent that a URIreference is being used as an opaque
identifier, and to the extent that RDF itself doesn't dereference any
URIreferences -- RDF doesn't directly deal with representations, just
"resources" which are actually "things"***

Jonathan

*** a "thing" is the union of RFC 2396 defined "resources" and "nodes"
within an abstract document which are identified with fragment ids