Re: dealing with strings in RDF

From: Graham Klyne <GK@ninebynine.org>
Subject: Re: dealing with strings in RDF
Date: Thu, 19 Jun 2003 11:10:58 +0100

> At 14:50 18/06/03 -0400, Peter F. Patel-Schneider wrote:
> 
> >Hi:
> >
> >Does
> >
> >         <http://foo.ex/a#b> <http://foo.ex/a#b> "aa" .
> >
> >rdf-entail
> >
> >         <http://foo.ex/a#b> <http://foo.ex/a#b> 
> > "aa"^^http://www.w3.org/2001/XMLSchema#string .
> >
> >I believe that it is up to the RDF Core working group to provide an answer
> >to this question.
> 
> Not speaking for RDFcore, but as I think this is an important point to 
> clarify in some way, my view is this:
> 
> No, this is not an rdf-entailment.

Yes, of course.  In my drive to be more precise, I changed entail to
rdf-entail, forgetting that this needs XSD-entailment (or, at least, some
datatype entailment that includes xsd:string.  Silly me.

> But there's a related question:  is it an rdfd {xsd:string}-entailment?
> 
> In this case I think the answer is yes.  And vice versa.  But if a language 
> tag is on the plain literal, then the entailment is off.

I wholeheartedly agree.  However, the current editor's draft of RDF
Semantics contains, in Section 1.2:

	The semantics described here is deliberately agnostic on the
	question of whether or not the values of xsd:string are idential to
	character strings.

> My rationale for this is that both xsd:string values and plain literals 
> (without lang tags) are defined to denote UNIcode character sequences 
> derived straightforwardly from the literal content.  I think the above 
> conclusions follow from the abstract syntax and semantics as given (but I 
> don't have proof of this).

I agree with this reasoning, which is why I think that the wording I quoted
above should be removed.

> #g
> 
> 
> -------------------
> Graham Klyne
> <GK@NineByNine.org>
> PGP: 0FAA 69FF C083 000B A2E9  A131 01B9 1C7A DBCA CB5E


Peter F. Patel-Schneider
Bell Labs Research
Lucent Technologies

Received on Thursday, 19 June 2003 08:50:24 UTC