Re: Update semantics LCC

From: Patrick.Stickler@nokia.com
Subject: RE: Update semantics LCC
Date: Thu, 16 Jan 2003 12:06:51 +0200

> > -----Original Message-----
> > From: ext Peter F. Patel-Schneider 
> > [mailto:pfps@research.bell-labs.com]
> > Sent: 15 January, 2003 22:04
> > To: Jan.Grant@bristol.ac.uk
> > Cc: bwm@hplb.hpl.hp.com; w3c-rdfcore-wg@w3.org
> > Subject: Re: Update semantics LCC

[...]

> > Unfortunately, IL is inconsistently defined for rdf:XMLLiteral
> > 
> > In Section 3.1 there is
> > 
> > 	if ... IL("xxx"^^rdf:XMLLiteral) is the XML canonical 
> > form of xxx
> > 	if ... IL("xxx"@yyy^^rdf:XMLLiteral) is the XML 
> > canonical form of
> > 		xxx with the language tag yyy
> > 
> > In Section 3.4 there is
> > 
> > 	for any typed literal L of the form "sss"^^ddd or "sss"@ttt^^ddd in
> > 	G, if I(ddd) is in D and "sss" ("sss"@ttt) is a valid lexical form
> > 	for I(ddd), then IL(L) = L2V(I(ddd))(sss)
> > 
> > which is inconsitent for rdf:XMLLiteral as long as there is any
> > langauge-tagged XML document whose canonical form depends on 
> > the language
> > tag.
> 
> This is not inconsistent, because it is the lexical form of the
> XMLLiteral which contains the language tag. I.e., the lexical form
> of an XMLLiteral can be a pair, not just the XML encoded string.

[...]

Hmm.

Actually, it might be OK, but only because of the wording outside of the
conditions that the extra conditions are ``on all datatypes other than the
built-in datatype''.  This wording is cryptic in the extreme.  There also
needs to be some clarification here as to whether datatypes are members of the
domain or URI references.

The individual conditions in the table need to have the exclusion
explicitly and individually stated by something link, ``if I(ddd) is in
D-I(rdf:XMLLiteral)''.  However, adding this exclusion to the fourth
condition would require adding a subclass relationship between
rdf:XMLLiteral and rdfs:Literal (which should, in any case, go into RDFS
interpretations).

Otherwise, however, these extra conditions *require* that 

IL("..."@en^^rdf:XMLLiteral) = L2V(I(rdf:XMLLiteral))("...")

and

IL("..."@fr^^rdf:XMLLiteral) = L2V(I(rdf:XMLLiteral))("...")

which requires that the denotation of XML literals be independant of the
language tag.

The translation to Lbase ignores language tags for all typed literals.

> (As an aside, I personally consider the exceptional nature of XML literals
> to be a mistake, for lots of reasons, but that's the way they are defined
> so the "fix" at this stage is to simply make the specs clear about it).

Agreed.

> Cheers,
> 
> Patrick

peter

Received on Thursday, 16 January 2003 06:38:41 UTC