W3C home > Mailing lists > Public > public-owl-wg@w3.org > May 2009

Re: CR Exit Criteria

From: Sandro Hawke <sandro@w3.org>
Date: Wed, 27 May 2009 13:03:33 -0400
To: "Michael Schneider" <schneid@fzi.de>
cc: "Jim Hendler" <hendler@cs.rpi.edu>, "OWL 1.1" <public-owl-wg@w3.org>
Message-ID: <5848.1243443813@ubehebe>

> >So I suggest that we make the only OWL 2 Full criterion:
> >
> > At least one implementation passing each of the non-DL test cases that
> >support features of the OWL 2 Full design that were not in OWL 1.
> I consider "one for each" to be a *very* hard criterion for the given =
> time frame of our WG, and the restriction to "non-DL" and "OWL 2-only" =
> tests to be hardly any better than asking for a complete implementation =
> of OWL 2 Full. This criterion is also very dependent on the actually =
> used test suite and, in particular, won't be realistically achievable =
> within the few weeks of the CR phase with the test suite that I am =
> creating at the moment.
> Just to be clear: I don't claim that the criterion is too hard to meet =
> *per se*, which would really be an, erm, "exit criterion" for OWL 2 =
> Full. I don't see anything obviously "un-implementable" in OWL 2 Full, =
> in particular now that the so called comprehension conditions of OWL 1 =
> Full are gone. And the often stated theoretical problem of =
> undecidability is, of course, not relevant for the task of implementing =
> a concrete test suite - in fact, this is always trivial in principle =
> (one can write a script that generates such a reasoner from a given test =
> suite having the test cases hard encoded :)). But if one is about =
> creating a serious implementation of OWL 2 Full, then this will almost =
> certainly take much too long for this working group to build one. The =
> same would probably be true for an OWL 2 /DL/ reasoner, if one were =
> about implementing it from scratch at the beginning of the CR phase.

This looks challenging.   :-(

I've heard the claim that people actually want to use OWL 2 Full....  If
that's the case, do we know what they want to do with it?  How long will
it take to implement reasoners for just the bits they want?  Is that
still months and months, or is it much less because they want fewer
bits? If it's a smaller subset, then let's designate some subset of the
tests as "key element tests" or something, corresponding to the bits in
OWL 2 Full that we know people want to use, and then use Jim's proposed
criterion for those tests.

     -- Sandro
Received on Wednesday, 27 May 2009 17:03:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:12 UTC