RE: problem with N-Triples

I've just hit this one with a parser based on RFC 2396 absolute URI, which
as Peter pointed out doesn't include fragments (hence '#' chokes my parser).
Leaving the definition to the RFC makes sense, but the question remains -
which production(s)? Any uriref (including relative, with queries etc) or
just absolute + frag?

Cheers,
Danny.

> -----Original Message-----
> From: www-rdf-comments-request@w3.org
> [mailto:www-rdf-comments-request@w3.org]On Behalf Of Dave Beckett
> Sent: 04 June 2003 18:12
> To: Peter F. Patel-Schneider
> Cc: www-rdf-comments@w3.org
> Subject: Re: problem with N-Triples
>
>
>
> On Fri, 30 May 2003 10:07:13 -0400 (EDT)
> "Peter F. Patel-Schneider" <pfps@research.bell-labs.com> wrote:
>
> >
> > The situation with respect to absoluteURI is very unclear.
> >
> > Is '::' a valid absoluteURI.  If not, why not?
> >
> > It appears that absoluteURI should be one of the productions
> from RFC 2396,
> > but which one?  It can't be absoluteURI, because that does not allow
> > fragments!
>
> I have already decided here to change the reference to point to the
> definition of RDF URI Reference as given in the RDF Concepts WD:
>   http://www.w3.org/TR/rdf-concepts/#dfn-URI-reference
> and say less here.  The N-Triples doc will then just describe how to
> encode such things into N-Triples.
>
> However, from what I recall of the definition of URIs, an absolute
> URI reference must start with a legal URI 'scheme', therefore there
> must be at least one legal 'scheme' character before the first ':';
> So '::' is not a legal absolute URI (or absolute URI reference).  The
> definitive answer is of course in the URI RFC and I think the BNF
> section defines that answer.
>
> Dave
>

Received on Tuesday, 10 June 2003 07:46:32 UTC