Re: Input sought on datatyping tradeoff

From: Brian McBride <bwm@hplb.hpl.hp.com>
Subject: Re: Input sought on datatyping tradeoff
Date: Mon, 15 Jul 2002 19:47:16 +0100

> At 07:01 13/07/2002 -0400, Peter F. Patel-Schneider wrote:
> >From: Brian McBride <bwm@hplb.hpl.hp.com>
> >Subject: Input sought on datatyping tradeoff
> >Date: Thu, 11 Jul 2002 19:46:49 +0100

[...]

> > > We believe this idiom to be quite straightforward,
> >
> >I disagree.  This idiom requires that blank nodes be able to denote data
> >values, and if the first triple above is not *always* paired with something
> >like the second triple, then there may be computational consequences for
> >languages like OWL.
> 
> This is important, though its another issue, rather than the one we are 
> currently seeking input for.  How can we turn that "may" into a more 
> definite statement?
> 
> Brian


Well there are several ``may''s that can arise here.

First, if the blank nodes result in the resource domain being non-disjoint
from the datatype domain, then there are many unknowns.  There have been
several messages on this in the WebOnt mailing list, including

	http://lists.w3.org/Archives/Public/www-webont-wg/2002Apr/0260.html

If it is possible to refer to an unknown element of some datatype, and make
statements about it, the situation is again murky.  I'm not the best person
to answer these sorts of questions, Carsten Lutz is probably better.

peter

Received on Monday, 15 July 2002 15:37:42 UTC