W3C home > Mailing lists > Public > www-webont-wg@w3.org > November 2002

RE: RDF datatyping summary

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Thu, 21 Nov 2002 15:06:22 +0100
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, <jjc@hplb.hpl.hp.com>
Cc: <www-webont-wg@w3.org>
Message-ID: <BHEGLCKMOHGLGNOKPGHDOELICAAA.jjc@hpl.hp.com>



> Calling this an error is misleading, in my opinion.  An invalid literal
> just denotes some junk value.

I think that's how Pat does it, you probably understand his text better than
I do.

> On the contrary, this is precisely the sort of thing that makes a useful
> test case.  It is an entailment that can easily be missed by implementors.

You may choose to push RDF Core to be decisive on this one then. The group
consensus seems to be lifting the carpet up, ready to brush this test case
under it.

> Well, I don't see why this is non-fixable.  WebOnt could just state
> something like
>
> 	A datatype URI that is a fragment in an XML Schema document refers
> 	to the top-level simple datatype with that fragment as its local
> 	name, if any.


If we believe that fixes it, I suggest we should make that as an official
comment to RDF Core, preferrably this week. It makes more sense to let RDF
Core (do the work/take the flak) on datatyping. It perhaps works ... DanC
understands the difficulties best.

Perhaps even,

  In the absence of any other indication from the owner, a datatype URI
  reference that is a fragment in an XML Schema document SHOULD be
  read as refering to the top-level simple datatype with that
  fragment as its local name, if any.

Personally, I would be minded to accept either version, but since it is a
substantive change it would need a WG decision.

Jeremy
Received on Thursday, 21 November 2002 09:06:38 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:55 GMT