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

Re: TEST: harder DL tests (3 SAT)

From: Sean Bechhofer <seanb@cs.man.ac.uk>
Date: Fri, 28 Feb 2003 10:13:57 +0000 (GMT Standard Time)
To: Jeremy Carroll <jjc@hpl.hp.com>
cc: www-webont-wg@w3.org
Message-ID: <Pine.WNT.4.44.0302280941320.2140-100000@potato>

On Fri, 28 Feb 2003, Jeremy Carroll wrote:

> Sean's Tests
> ==========
> I have added Sean's tests - one question for Sean or Ian in
> http://www.w3.org/2002/03owlt/editors-draft/snapshot#dl-000-satisfiability-025
> The description reads:
> [[
> DL Test: t5f.1 Non-finite model example from paper The concept should be
> coherent but has no finite model
> ]]
> Which paper? I will add it to the references.
> (These are a great addition)

Ian? This was a comment from the original lisp format....

> OWL Lite and UnionOf
> ==================
> I have also added Ian's example from
> http://lists.w3.org/Archives/Public/www-webont-wg/2003Jan/0401.html
> which is found in 6 tests under:
> http://www.w3.org/2002/03owlt/editors-draft/snapshot#proposedIssue-I5.2-Language-Compliance-Levels
> I note that many of Sean's tests can be reformulated as OWL Lite tests using
> Ian's techniques, and I hope to do that (copying the tests into a new
> subsection "Harder OWL Lite Tests")

Yes. I didn't attempt to make any distinction between DL/Lite.

> 3 SAT and individuals
> ==================
> I was a little worried at the lack of exercise for the distinctive DL feature
> of individuals as well as classes and properties; I am also not convinced
> that Sean's tests adequately exercised the OWL DL cardinality constructs.

The tests I submitted this week were the basic reasoning tests that I've
been using for some regression testing and that were already in roughly
the right format. There are also a number of test problems from the DL98
systems comparison exercise [1] which we should be able to use. Some of
these involve A-box (individual) reasoning and the application KB tests
also use cardinality restrictions, although a quick glance suggests that
in the main they're of the form atmost 1/atleast 1. I'll make a stab at
converting those into OWL tests too.


[1] http://dl.kr.org/dl98/comparison/data.html

Sean Bechhofer
Received on Friday, 28 February 2003 05:14:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:56:51 UTC