W3C home > Mailing lists > Public > www-webont-wg@w3.org > January 2003

RE: ISSUE 5.2 Language Compliance Levels - proposed clarification

From: Ian Horrocks <horrocks@cs.man.ac.uk>
Date: Mon, 27 Jan 2003 11:12:49 +0000
Message-ID: <15925.5169.789855.57193@merlin.horrocks.net>
To: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>
Cc: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, <www-webont-wg@w3.org>

On January 24, Jeremy Carroll writes:
> 
> 
> OK - I'll rephrase the proposal.
> 
> (I was just trying to do the minimal change on the current resolution
> closing the issue).
> 
> How about:
> PROPOSE
>   - to reopen issue 5.2
>   - to retract the endorsement of existing OWL lite language subset.
>   -  to remove modality = complete from the OWL Lite Class Axioms in the
> Abstract Syntax
>   -  to endorse the modified OWL Lite language subset
>   -  to close issue 5.2
> 
> This is more in keeping with seeing AS&S as the definitive doc.

Jeremy,

This is madness - surely you can't be suggesting that we re-open the
discussion on OWL Lite *AGAIN*:

1. The specification of the language has been discussed at great
length at more than one f2f meeting and has been agreed by the whole
group *including you*.

2. Apart from the AS&S specification, which has been stable for
*months*, the expressive power of OWL Lite was precisely characterised
in [1], which clearly states that it contains ALC, is contained in
SHIQ(D+) and that satisfiability/subsumption is thus decidable in
ExpTime. You clearly saw that because you corrected one of the
references! (see [2]).

3. Simply tinkering with the syntax of OWL Lite almost certainly
*wont* achieve the result you want/expect (whatever that is).

4. If you really did succeed in eliminating the ability to express
"complete" classes in OWL Lite, you would make it useless in a wide
range of important applications (e.g., see [3]).


To get this far with OWL we have all made compromises w.r.t. what
would have been our preferred outcome. I was under the impression that,
after a great deal of hard work, we had finally managed to reach a
position that we can all live with (just), and that apart from some
relatively minor details we were ready to move forward to last call.

If we agree to re-open a core issue (i.e., the basic specification of
one of the OWL species), then I suppose we can expect everyone to
demand the re-opening of issues were they had formerly agreed to a
compromise. I.e., we would set back the work of the WG by months, if
not sabotage it altogether.

Regards, Ian

[1] http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0239.html
[2] http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0243.html
[3] http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0088.html


> 
> Jeremy
> 
> 
> > -----Original Message-----
> > From: Peter F. Patel-Schneider [mailto:pfps@research.bell-labs.com]
> > Sent: 24 January 2003 01:07
> > To: jjc@hpl.hp.com
> > Cc: www-webont-wg@w3.org
> > Subject: Re: ISSUE 5.2 Language Compliance Levels - proposed
> > clarification
> >
> >
> > From: Jeremy Carroll <jjc@hpl.hp.com>
> > Subject: ISSUE 5.2 Language Compliance Levels - proposed clarification
> > Date: Thu, 23 Jan 2003 21:27:50 +0100
> >
> > [...]
> >
> > >   - to endorse the existing OWL Lite language subset in the OWL
> > Overview of
> > > 20 Jan 2003
> > >
> > http://lists.w3.org/Archives/Public/www-webont-wg/2003Jan/att-0327
> > /01-OWLOverview
> >
> > [...]
> >
> > Hmm.
> >
> > To endorse the language described in this document requires:
> >
> > For OWL Lite:
> > - prohibit owl:DatatypeProperty and owl:ObjectProperty
> > - not use datatypes (yet)
> > - not use owl:AllDifferent
> > - allow owl:inverseOf, owl:TransitiveProperty, owl:SymmetricProperty,
> >   and owl:InverseFunctionalProperty on any property
> > - apply restrictions to classes (somehow)
> >
> > For OWL DL:
> > - prohibit owl:DatatypeProperty and owl:ObjectProperty
> > - require that all properties belong to either owl:DatatypeProperty and
> >   owl:ObjectProperty
> > - not use datatypes (yet)
> > - not use owl:AllDifferent
> > - not allow owl:oneOf for data values
> > - allow owl:inverseOf, owl:TransitiveProperty, owl:SymmetricProperty,
> >   and owl:InverseFunctionalProperty on any property
> > - apply restrictions to classes (somehow)
> >
> > I do not think that these are good ideas.
> >
> > Peter F. Patel-Schneider
> > Bell Labs Research
> > Lucent Technologies
> >
> 
Received on Monday, 27 January 2003 06:13:35 GMT

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