RE: "Resource" (RDF vocabulary definitions)

> What about URI references which include fragment identifiers? My
> reading of RFC 2396 is that <http://example.org/> and
> <http://example.org/#foo> refer to the same resource.

In the case of http://example.org/#foo the URI (http://example.org/) is a
URI identifying a resource. The fragment identifier (#foo) is either:

a. meaningless.
b. An identifier of a resource that is part of the resource identified by
the URI.

Since which of these is the case can only be determined once the
representation is obtained (and even then not with 100% certainty; for we
may only be receiving one of a potentially limitless number of possible
representations, so if we receive a representation which doesn't contain an
identifiable representation of the subresource we can't assume the
subresource doesn't exist) any system that uses URIRefs in an opaque manner
(such as RDF) must assume that a URIRef with a fragment identifier refers to
a different resource to the same URIRef without the fragment (that the
former is a component of the latter is a reasonable inference, but just what
that means could be pretty vague).

Received on Friday, 22 November 2002 06:05:33 UTC