W3C home > Mailing lists > Public > public-owl-wg@w3.org > November 2007

Re: ISSUE-52 (Explanations): Specification of OWL equivalences and rewriting rules for explaining inferences

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Mon, 05 Nov 2007 12:53:15 -0500 (EST)
Message-Id: <20071105.125315.55856658.pfps@research.bell-labs.com>
Cc: public-owl-wg@w3.org

From: "Kashyap, Vipul" <VKASHYAP1@PARTNERS.ORG>
Subject: RE: ISSUE-52 (Explanations): Specification of OWL equivalences and rewriting rules for explaining inferences
Date: Mon, 5 Nov 2007 12:21:24 -0500

> >I really don't see a need for the OWL 1.1 spec to include a way of
> >passing around even just proofs. 
> [VK] I realize that I seem to have a different perspective on this from a few
> members of this group. May I request your feedback on the following questions to
> clearly understand where the dissonance lies?
> 1. Is there a need for proofs and explanations to the ontology builder as he is
> developing and debugging large scale ontologies?

Maybe.  There is certainly need for some help in building large
ontologies.  Whether proofs and explanations are part of the solution
is still unclear, at least to me.  

> 2. If the ontology builder is collaborating with a set of ontology builders
> using different ontology building tools and reasoners, how would he or she share
> these explanations across his/her team?

This presupposes that explanations are useful.  Even if this is so, it
may be that the sharing is completely dependant on the particular tools
being used and thus not even suitable for standardization.

> 3. Is it likely some folks might [want] their own proprietary format as a
> standardization scheme is not supported by the OWL/Semantic Web community.

Maybe, or maybe the format would not be proprietary, just idiosyncratic
to a tool or user group or ....

> 4. Could this then lead to potential fragmentation of the OWL 1.1 standard
> around this issue?

How?  It is not as if the language depends on how explanations are
handed around. 

I think that this may be where our disconnect is happening.  I just
don't see proofs and explanations as relevant to the language itself.
Therefore proofs and explanations are not germane to the the WG.

> 5. Also, could this have a negative impact on acceptance and adoption of OWL 1.1
> due to lack of what I believe is an important user need?

Well, given the above, I don't see that there could possibly be a
negative impact.

> >Consider the problems in making a proof language part of OWL 1.1.  If
> >the proof language has meaning as part of an ontology then the WG has to
> >design syntax, devise semantics, translate it into RDF, etc.  Some of
> >these are hard problems.  If the proof language doesn't have meaning as
> >part of an ontology, then it shouldn't be in an ontology at all, and the
> >WG shouldn't be working on it.
> [VK] The above is a scope issue and I agree in large part with this. Is it
> possible to scope down the effort in the following manner?
> 1. The Proof Language is not part of the ontology (extra logical construct?)
> 2. The specifications of proofs could be supported in a limited manner using
> annotation properties as strings?
> 3. Maybe develop a limited format for proofs around a small extension to OWL
> 1.1?

I don't see this as a way of bringing proofs and explanations into the WG.

> Thanks and Regards,
> ---Vipul

I don't have any problem with work on proofs and explanations.  There
should be more research on proofs and explanations and on support tools
for ontology building.  There even should be more support of this sort
of thing - I'm still dissapointed that DARPA didn't follow up on the
DAML program, where such work could have flourished.

If you, or anyone else in the WG, is interested in pursuing work in this
area, related to OWL or not, related to the semantic web or not, feel
free.  I'll even provide what limited (moral?) support I can.  W3CD may
even be able to provide some limited resources, like
public-owl-dev@w3.org (which is "for discussion [related to] the Web
Ontology Language (OWL), from specifications to tools to applications,
tutorials, books, and so on" and thus should be entirely appropriate),
or even semantic-web@w3.org.


PS:  I think I've made my position on the topic quite clear, so I'll bow
out of the discussion until and if the issue is actually accepted.
Received on Monday, 5 November 2007 18:04:41 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:00 UTC