W3C home > Mailing lists > Public > www-rdf-validator@w3.org > November 2001

RE: RDF Validation Question

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Fri, 23 Nov 2001 17:34:13 -0000
To: "Stephen Williams" <stephenswilliams@yahoo.com>, <Jeremy_Carroll@hp.com>, <www-rdf-validator@w3.org>
Message-ID: <JAEBJCLMIFLKLOJGMELDOEIPCCAA.jjc@hplb.hpl.hp.com>
[--aside who is www-rdf-validator nowadays?--]

Sorry Stephen for taking so long to reply to this.

My understanding of your query is that:

- If we replace both occurrences of cr:test with crtest in your example then
the picture coming out of the validator is different.

- The equivalent experiment with Beckett's code is that the triples coming
out are identical (except for the change in label).


I am clear that you have identified an ARP defect here, and have entered
this into Bugzilla. I typically clear the bugs before each release.
I guess I am looking at another ARP release before christmas?

The rdf:resource="#cr:test" and the rdf:ID="cr:test" are I think identical
and ARP should behave identically towards both.

However, what it does is to treat the first as an illegal URI and the second
as a legal fragment ID.

In default mode, ARP responds to illegal URIs by issuing a warning message
and then treating the exact text string as if it were a legal absolute URI.
Hence "#cr:test" is not treated as a fragment ID.

I will need to refer to RFC 2396 to determine whether or not "#cr:test" is a
legal relative fragment URI or not. I will do this when I fix the bug.

So the fix will either:
- allow the first occurrence
or
- forbid the second occurrence

I don't know which yet.

Jeremy







> -----Original Message-----
> From: Stephen Williams [mailto:stephenswilliams@yahoo.com]
> Sent: 08 November 2001 22:47
> To: Jeremy_Carroll@hp.com; www-rdf-validator@w3.org
> Subject: RDF Validation Question
>
>
> Hi Jeremy
>
> If I reference a description identified with a
> namespace qualified ID then the RDF validator does not
> make this association, E.g.:
>
> 	<rdf:Description rdf:ID="one">
> 		<sw:name rdf:resource="#cr:test"/>
> 	</rdf:Description>
> 	<rdf:Description rdf:ID="cr:test">
> 		<sw:first_name>xxx</sw:first_name>
> 		<sw:last_name>yyy</sw:last_name>
> 	</rdf:Description>
>
> However, if the ID is not namespace qualified then the
> validator makes the association.
>
> However, Dave Beckett's Redland RDF Database Demo at
> http://www.redland.opensource.ac.uk/demo appears to
> make the association when namespace ID's are used.
>
> Is it correct to use namespaces with ID's with in RDF?
>
> Thanks for your assistance.
>
> Stephen Williams
>
> __________________________________________________
> Do You Yahoo!?
> Find a job, post your resume.
> http://careers.yahoo.com
>
>
Received on Friday, 23 November 2001 12:35:00 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 1 May 2011 06:14:53 GMT