W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > April 2001

Re: Issue http://www.w3.org/2000/03/rdf-tracking/#rdf-ns-prefix-confusion

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Mon, 23 Apr 2001 11:10:11 +0100
To: RDF Core <w3c-rdfcore-wg@w3.org>
Message-ID: <13839.988020611@tatooine.ilrt.bris.ac.uk>
>>>Brian McBride said:
> Hi Dave,
> 
> Good proposal.  Like the use of test cases.

That email is now [1]

> 
> When you said:
>  
> >   <Description xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
> >                about="http://example.org/">
> 
> did you mean:
> 
>  
> >   <Description xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>                  ^^^^^
> >                about="http://example.org/">


Yes, sorry - I should have checked with a parser (!) before I sent
it.  Just to repeat, the above was example #2 of what would be allowed
since the Description element was impliclity in the RDF namespace
(with the correction above) and the about attribute could therefore
be interpreted.

> 
> Regarding handling non-syntactic properties, I'd suggest that whatever
> we
> do, we do it consistently.
> 
> One thing to note is that the current grammar specifies that syntactic
> attributes can always be unqualified.  If we are going to change that
> we should check what, if anything will break and perhaps we should
> deprecate rather than immediately outlaw.

(Again using the IETF words for requirement levels from RFC 2119
 http://www.ietf.org/rfc/rfc2119.txt.  I propose we use these.)

Deprecation is a stronger alternative to what I proposed but I would
support that as an alternative change.  Something like this:

  In RDF 1.1 namespaced attributes MUST be accepted on input and MUST
  be emitted on output.  Non-namespaced attributes MUST be accepted
  on input but MUST NOT be emitted.  In the next version of the
  syntax, non-namespaced attributes SHALL be rejected RDF 1.1 input.

> 
> I've written up some test cases as an XML file (text at the end of this
> message in lieu of getting a directory set up to hold test cases), and
> run them through the parsers I have access to.

<snip/>

I have similar tests I used for Rapier to make sure I got this
support working but I also did some adhoc checking with other parsers
and found inconsistencies.  Some more strict ones took the grammar
literally and forbid rdf:resource etc. (this is from memory) which is
a valid interpretation of the spec, if you ignore the M&S examples
which use rdf:about :-)

Dave

[1] http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Apr/0031.html
Received on Monday, 23 April 2001 06:10:14 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:35:04 EDT