W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > March 2003

Re: Suggestion williams-02 - *huge* editorial change

From: Graham Klyne <GK@NineByNine.org>
Date: Fri, 28 Mar 2003 10:52:56 +0000
Message-Id: <>
To: Jeremy Carroll <jjc@hpl.hp.com>, w3c-rdfcore-wg@w3.org

I would be *very* wary about adopting the term IRI at this time.

I was at the IETF URI BOF last week, and there was some discussion of the 
IRI spec.  Minutes at: 

One issue in particular that is not currently clear regarding IRIs is the 
status of certain US-ASCII characters not allowed in URIs:
(<, >, ", SPACE, {, }, |, \, ^, `) are not allowed in URIs; in the
current draft, they are allowed in IRIs.


XML namespaces 1.1: http://www.w3.org/TR/xml-names11/

I also note that this spec does not *define* IRIs, but refers to an 
Internet draft:
IRI draft
     Internationalized Resource Identifiers (IRIs). Available at 

I would say that this doesn't have a cat-in-hell's chance of being on the 
IETF standards track by the time (we hope) RDFcore goes to recommendation 

Hmmm... there's also http://www.w3.org/TR/xml-names11/#IRIs, but I don't 
think I'd know exactly how to include an IRI containing (say) '>' in an RDF 
document based on that specification.


My conclusion:  this whole area is very messy at present, and I'm not sure 
how we can be confident of getting any alignment with IRI's completely 
right.  I'm inclined use some text of our own (possibly copy xml-names11 
and clarify the use of "additional" characters), with a note that this may 
be replaced with a reference to a future IRI spec, rather than indirect 
through some transient specification in yet another document.

I note that adopting the IRI per xml-names11 would require a change to 
N-triples, which has:

absoluteURI  ::=  ( character - ( '<' | '>' | space ) )+
-- http://www.w3.org/2001/sw/RDFCore/ntriples/#absoluteURI


At 11:19 27/03/2003 +0100, Jeremy Carroll wrote:

>Williams-02 is the suggestion that we should:
>- defer to Namespaces 1.1 for the definition of IRI
>- and globally substitute "[RDF] URI Reference"  by "IRI"
>There is a small substantive differences which is Unicode Normal Form C.
>On the advice of I18N WG in Cannes we required RDF URI References to be in
>Unicode NFC; Martin Dürst informs me that they have changed their minds on
>this one.
>Draft proposal:
>Change concepts to defer to Namespaces 1.1 defn of IRI rather than have a
>definition of RDF URI reference.
>Globally substitute "[RDF] URI Reference", "uriref" etc.  by "IRI".
>This needs discussion!
>Also, we should e-mail i18n-ig and ask for an official comment from the wg
>retracting their advice in Cannes.

Graham Klyne
PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E
Received on Friday, 28 March 2003 07:10:37 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:24:21 UTC