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

Re: Possible syntax restriction on encoding rdf:XMLLiteral

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Thu, 7 Aug 2003 14:55:25 +0100
To: Patrick Stickler <patrick.stickler@nokia.com>
Cc: ext Brian McBride <bwm@hplb.hpl.hp.com>, RDF Core <w3c-rdfcore-wg@w3.org>, Martin Duerst <duerst@w3.org>
Message-Id: <20030807145525.46a592b8.dave.beckett@bristol.ac.uk>

On Thu, 7 Aug 2003 10:55:40 +0300
Patrick Stickler <patrick.stickler@nokia.com> wrote:

>   ----- Original Message ----- 
>   From: ext Brian McBride 
>   To: Dave Beckett 
>   Cc: RDF Core ; Martin Duerst 
>   Sent: 06 August, 2003 16:52
>   Subject: Re: Possible syntax restriction on encoding rdf:XMLLiteral
Brian said: 
>   3) it doesn't achieve its goal.  there is nothing to stop folks defining 
>   synonyms for rdf:XMLLiteral, so I don't see how one can prevent it.

Patrick said:
> I think this particular point is sufficient to justify not 
> making any such syntactic restriction.
> Better to view parseType="Literal" as syntactic sugar for the 
> general form of typed literal expression, as a convenience to
> users -- not as either a primary or (as suggested) exclusive
> form of expression.
> I.e. typed literals are expressed using rdf:datatype and a valid
> lexical form. Because RDF/XML *is* XML, and because the canonical
> lexical form of rdfs:XMLLiterals is complicated, we provide a
> more convenient way to express them, just as various contractions
> are provided by the RDF/XML syntax.

The point is, we do NO checking on the value of the rdf:datatype content.
So you can encode something that might be a non exc-C14N
Unicode string (caveats, after octets->de-utf8ing).  That means you
can introduce ill-formed XML into the graph and if you get into
entailments, these have no literal denotation.

So if this remains this way, applications of RDF that just deal with the
graph need to have XML parsers and exc-C14N checkers, even if they
never used rdf/xml.  Not that it's hard, XML tools are everywhere.

Received on Thursday, 7 August 2003 10:00:36 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:54:07 UTC