Re: pfps-02

From: Brian McBride <bwm@hplb.hpl.hp.com>
Subject: pfps-02
Date: 21 Jul 2003 14:12:04 +0100

> Peter,
> 
> with reference to your comment
> 
> http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-02
> 
> The RDFCore WG resolved to accept your comment:
> 
>  http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Jul/0236.html
> 
> As you know, there have been extensive modifications to the treatment
>  of XML literals since your comment was written.
> 
> The current editor's draft, which is now stable enough to review,
> 
> http://www.w3.org/2001/sw/RDFCore/TR/WD-rdf-mt-20030117/
> 
> now contains a version of the Lbase translation rules which we believe 
> adequately covers the case you mention. XML Literals are now treated 
> exactly uniformly with other typed literals in the Lbase axioms.  Please
> note that the LBase reference is to the published note.
> 
> Please reply to this message, copying www-rdf-comments@w3.org, 
> indicating whether this response adequately addresses your comment.
> 
> Brian

My comment does not have to do with just the cases I provided.  My concern
was about the translation in general.  

I note that the current document does not correctly treat many strings
correctly.  For example the string
	"'a\'b'"
is translated to
	"'a\\\'b'"
which denotes
	'a\\\'b'
not the original string.

XML Literals may be treated exactly uniformly with other literals in the
LBase translation, but this is not in accordance with the RDF semantics.
For example, it does not appear that the translation of 
	ex:a ex:b "a"^^rdf:XMLLiteral .
plus the RDF and RDFS axioms implies that the translation of 
"a"^^rdf:XMLLiteral is in the extension of rdfs:Literal.  This only comes
out of the LBase version of a datatype theory that includes rdf:XMLLiteral.


I also note that the document has other serious problems.  

Peter F. Patel-Schneider

Received on Wednesday, 23 July 2003 13:09:58 UTC