RE: Using fragment identifiers with URNs

Roy Fielding wrote:
> Note, however, that your suggested change would
> restrict the applicability of a URI-reference beyond what the 
> specification currently requires, to the point where it conflicts
> with the Web.  You are suggesting that an application-specific
> requirement be placed on an existing protocol element in order
> to satisfy some restriction that somebody wants to use within RDF.
> I don't see any reason why we should make that change.

Actually, I'm not proposing making a change, just finding out how
compatible the IETF notion of a URI is with the use of a URI scheme
to represent abstract names with no retrieval semantics.  The answer
seems to be that it's not compatible.  For the record though, I don't
see how such a change would "conflict with the Web".

- Stephen