Re: Typed literals text

[Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com]


----- Original Message ----- 
From: "ext Jeremy Carroll" <jjc@hpl.hp.com>
To: "Patrick Stickler" <patrick.stickler@nokia.com>; <w3c-rdfcore-wg@w3.org>
Sent: 17 October, 2002 11:26
Subject: RE: Typed literals text


> >
> > IMO, XML Schema got it wrong.
> >
> > There's no reason why RDF should repeat that error.
> >
> > And this is one good reason why RDF datatyping should be framework
> > neutral, so that we are free to address such requirements as language
> > qualification of typed literals while still supporting other standards
> > such as XML Schema.
> 
> W3C specs should work together.

Agreed, where reasonable and practical.

> If XSD is wrong, we should tell them, and it is their responsibility to fix
> it - not ours.

Agreed, but we should also be able to 'do the right thing' in our spec
while they are fixing their spec.

> If we go off and do our own thing then we condemn the semantic web to
> further fragmentation and Aaron's tag line will become "RDF - ten years of
> failure"

I see no conflict whatsoever between XML Schema and RDF if RDF allows
additional language qualification of literals.

Exactly what kinds of problems do you see as potentially causing
fragmentation?

We are not saying that XML elements defined to have data content
conforming to anySimpleType must have language qualification. We
are saying that RDF literals may have language qualification,
and furthermore may be constrained by some datatype, which may
very well be a subtype of XML Schema anySimpleType.

I don't see how that is violating anything in XML Schema. RDF property
elements are not simple CDATA based elements. They have mixed content
and the interpretation of that content is defined by RDF, not any
XML Schema datatype.

Where is the problem?

> We *must* be prepared to live with other working groups decisions, even
> those we don't like. (Of course, our implementations may choose to diverge,
> but as a W3C standards body we should not endorse such behaviour).

RDF datatyping is not XML Schema datatyping. There is no reason why
we should enforce a restriction mandated by one particular datatyping
framework which (a) is not requuired by other datatyping schemes and
(b) is used by existing applications (mine) and (c) is simply wrong.

But rather than just the two of us debating this issue, I'd be happy
to hear what other folks think.

Patrick

Received on Thursday, 17 October 2002 06:33:13 UTC