Re: IRIs vs. URIs in SPARQL spec

On Mon, 2011-03-07 at 09:19 -0500, Bob DuCharme wrote:
> Thanks Dave... so if
> http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#section-Graph-URIref says that "RDF URI references are compatible with International Resource Identifiers as defined by [XML Namespaces 1.1]", does this mean that URI References do allow non-US-ASCII characters, so that http://www.example.org/Montréal is a valid URI Reference but not a valid URI? 

A valid RDF URI Reference (and IRI), yes, if you encode that string in
UTF-8 then %-encode that you would arrive at a syntactically legal URI.

> And that if "isURI is an alternate spelling for the isIRI operator",
> as the SPARQL spec says, then the latter refers to URI references and
> not to URIs? 

to RDF URI References, yes I assume so. I've no inside knowledge here :)

Cheers,
Dave

> On 3/7/2011 3:34 AM, Dave Reynolds wrote: 
> > On Sun, 2011-03-06 at 13:45 -0500, Bob DuCharme wrote: 
> > > The SPARQL specs says in two places that "IRIs are a subset of RDF URI 
> > > References that omits spaces." I have trouble seeing it as a subset for 
> > > two reasons:
> > > 
> > > 1. The  http://www.ietf.org/rfc/rfc3987.txt document that it references 
> > > says that it "defines a new protocol element called Internationalized 
> > > Resource Identifier (IRI) by extending the syntax of URIs to a much 
> > > wider repertoire of characters."
> > Note that those statements in the SPARQL spec are not about URIs but
> > about "RDF URI References" [1], these are not the same thing. RDF URI
> > References were an attempt by the RDF Core WG to be compatible with IRIs
> > before the IRI spec was finalized - an attempt which was very successful
> > apart from the the handling of spaces.
> > 
> > Dave
> > 
> > [1]
> > http://www.w3.org/TR/2004/REC-rdf-concepts-20040210/#section-Graph-URIref
> > 
> > 

Received on Monday, 7 March 2011 15:58:32 UTC