W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > September 2003

Re: Action needed: subClassOf on datatypes

From: Graham Klyne <gk@ninebynine.org>
Date: Tue, 02 Sep 2003 16:38:17 +0100
Message-Id: <5.1.0.14.2.20030902163310.00b7a918@127.0.0.1>
To: Dan Brickley <danbri@w3.org>
Cc: Dan Connolly <connolly@w3.org>, pat hayes <phayes@ihmc.us>, w3c-rdfcore-wg@w3.org

At 11:22 02/09/03 -0400, Dan Brickley wrote:
[[
re 'who is responsible'?

Nowhere in our SW specs (to my knowledge) do we profess to care about
*who* makes such assertions, or for that matter *when*, *where* or
*Why*. All we care about is whether they are sensible (ie. meaningful)
things to assert, and if so, what they mean.
]]

Well, yes, I guess that's partly why I think that, having chosen an 
intrinsic interpretation of the subclass relation, it's problematic to 
treat datatypes differently.  My use of words "who is responsible" may be 
unfortunate in that it clouds the issue.

But, returning to my example, in the absence of a specific inference (ala 
extrinsic subclass), if one has two independently defined datatypes whose 
value spaces happen to be in subset relation,  how can the semantics of 
such independent definitions include a relation between them?

#g
--

At 11:22 02/09/03 -0400, Dan Brickley wrote:
>* Graham Klyne <gk@ninebynine.org> [2003-09-02 16:10+0100]
> >
> > At 07:57 02/09/03 -0500, Dan Connolly wrote:
> >
> > >(e) note that semantics of datatypes includes subClassOf
> > >relationships.
> >
> > I'm uneasy with that.
> >
> > Who is responsible for defining the relevant triples that cannot be
> > inferred?
> >
> > Suppose I have datatypes
> >
> >   xsd:integer
> >
> > and
> >
> >   foo:EvenNumbers
> >
> > where the latter's value space is subset of the first's.
> >
> > Who is responsible to say:
> >
> >   foo:evennumbers rdfs:subClassOf xsd:integer
> >
> > ?
>
>re 'who is responsible'?
>
>Nowhere in our SW specs (to my knowledge) do we profess to care about
>*who* makes such assertions, or for that matter *when*, *where* or
>*Why*. All we care about is whether they are sensible (ie. meaningful)
>things to assert, and if so, what they mean.
>
>Dan
>
> > I think it's easier to leave it unsaid, and allow anyone who needs the
> > subclass assertion to addf it in the knowledge that it's not inconsistent.
> >
> > #g
> >
> >
> > ------------
> > Graham Klyne
> > GK@NineByNine.org

------------
Graham Klyne
GK@NineByNine.org
Received on Tuesday, 2 September 2003 12:29:05 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:59:53 EDT