Re: Cutting the Patrician datatype knot

----- Original Message -----
From: <Patrick.Stickler@nokia.com>
To: <pfps@research.bell-labs.com>; <geoff@sover.net>
Cc: <www-rdf-interest@w3.org>
Sent: Saturday, November 24, 2001 6:11 AM
Subject: RE: Cutting the Patrician datatype knot


>
>
> > -----Original Message-----
> > From: ext Peter F. Patel-Schneider
> > [mailto:pfps@research.bell-labs.com]
> > Sent: 22 November, 2001 18:35
> > To: geoff@sover.net
> > Cc: www-rdf-interest@w3.org
> > Subject: RE: Cutting the Patrician datatype knot
> >
> >
> > From: "Geoff Chappell" <geoff@sover.net>
> > Subject: RE: Cutting the Patrician datatype knot
> > Date: Thu, 22 Nov 2001 08:42:30 -0500
> >
> > > > -----Original Message-----
> > > > From: www-rdf-interest-request@w3.org
> > > > [mailto:www-rdf-interest-request@w3.org]On Behalf Of Peter F.
> > > > Patel-Schneider
> > > > Sent: Wednesday, November 21, 2001 1:34 PM
> > > > To: www-rdf-interest@w3.org
> > > > Cc: joint-committee@daml.org
> > > > Subject: Cutting the Patrician datatype knot
> > > >
> > > > An RDF/XML serialization of an RDF graph element of the form
> > > > < s , p , v > for v a data value
> > > > is of the form
> > > > <... s ...>
> > > >   ...
> > > >   <p xsi:type="du">x</p>
> > > >   ...
> > > > </...>
> > > > where d is some datatype with URI du
> > > > for which v in DTC(d) and x is a lexical form for v in d.
> > > >
> > > > Thus in the serialization we need access to the
> > lexical-to-value mapping,
> > > > but not in the model theory.
> > >
> > > If you're going to put the mapping in the parsing, why not just use
> > > 'parseType=' to make clear it's a parser directive?
> >
> > Precisely because XML Schema has a perfectly good way of
> > doing it, so why
> > not use that way?
>
> Because, although we definitely want to ensure that RDF is fully
> compatible with, and takes advantage of the existing XML Schema
> data types, we cannot and should not preclude nor discriminate
> against the use of other data typing schemes which are not
> defined in XML Schema. That's not an anti-XML Schema position
> (I'm actually very much pro-XML Schema).
>
> I don't read the charter as saying that RDF must use XML Schema
> data type *definition mechanisms*. Only that it should make best
> use of XML Schema defined data types. Those are not quite the
> same thing.

Of course if parseType is used for this purpose and the parser does not do
the lex->val mapping (perhaps because the parseType is unrecognized, or...),
the parseType would need to be carried into the graph in some form. Then I
guess you're stuck with another meta-level statement like "lang=" where the
statement is really about the lexical form of the literal and not about some
thing denoted by the literal. And not being able to make statements about
literal instances...

Patrick, maybe something like your proposed URV mechansim could be used to
preserve these meta-statements (parseType, lang, ...). I think you'd want to
keep it clear that any statements made inside the URV were about the lexical
form, and not the value (which I guess isn't really the URV proposal?). With
that additional information, you could happily use datatyping proposals like
P (the original) without fears of ambiguity in lex->val mappings.


>
> Cheers,
>
> Patrick
>

Regards,

Geoff

Received on Friday, 23 November 2001 15:40:44 UTC