W3C home > Mailing lists > Public > www-webont-wg@w3.org > October 2002

RE: Guide: review - boring (this review not the guide)

From: Smith, Michael K <michael.smith@eds.com>
Date: Tue, 1 Oct 2002 08:14:59 -0500
Message-ID: <B8E84F4D9F65D411803500508BE3221410D446D8@USPLM207>
To: Jeremy Carroll <jjc@hplb.hpl.hp.com>, www-webont-wg@w3.org

Thanks, Jeremy.

This current RDF status is something I hope OWL can avoid.  For instance,
this is the first time I have even heard of the document Jeremy references, 

 Resource Description Framework (RDF):
 Concepts and Abstract Data Model

and I am still unclear as to what its weight will be vis-a-vis RDF syntax
and semantics even when it becomes a recommendation.  

The good news from Jeremy's note is that RDF has already charted most of the
murky, low level XML/XML Base/XML Namespace/URL waters for us.

Having an operational definition in the form of a validator is great, but it
is not a big help if it embodies syntax and semantics that are not recorded
in the standard.  

- Mike

-----Original Message-----
From: Jeremy Carroll [mailto:jjc@hplb.hpl.hp.com]
Sent: Tuesday, October 01, 2002 6:41 AM
To: www-webont-wg@w3.org
Subject: FW: Guide: review - boring (this review not the guide)

>
>  rdf:ID="VIN:FOO"     (IDsymbol is an XML Name.)
>
> Where is this ruled out?  Or is it permitted and if so what does it
> mean?

This one is probably not yet in any published doc really. M&S is at best
ambiguous, it really ought to refer to XML Namespaces NC Colon but doesn't.

See

http://www.w3.org/2000/10/rdf-tests/rdfcore/rdfms-rdf-id/error003.rdf

(this is not yet approved).

>
>  rdf:about="FOO"      (rel_path)
>  rdf:about="FOO#BOO"  (rel_path + fragment)
>  rdf:about="VIN:FOO"  (absolute URI with opaque part)
>
>

The newer specs are clear that:

1) The RDF graph requires absolute URIs
http://www.w3.org/TR/rdf-concepts/#section-Graph-URIref
2) That relative URIs in RDF/XML are converted to absolute URIs using
RFC2396, and xml:base, if any.
3) That rdf:about takes a URI not a qname.
Thus all three examples are legal, but VIN:FOO is a distinct uri from an
unregistered scheme VIN, rather than the uri which is formed from the qname
VIN:FOO.

Dan is right to point out that the RDF validator does encapsulate most of my
knowledge. Things missing at the moment are:
 - the illegality of " 1 " when an integer is required (depends on RDF
datatyping that is not finished)
 - rdf:parseType="Collection" support (present in the Jena CVS but not in
any released version - next release this week or next, probably a little
time before it gets into the validator).

But it does report an error on the obscure namespace erratum!

Jeremy
Received on Tuesday, 1 October 2002 09:17:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:53 GMT