W3C home > Mailing lists > Public > www-webont-wg@w3.org > September 2002

Re: TEST: Functional and InverseFunctional tests for approval

From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
Date: Tue, 03 Sep 2002 10:40:28 -0400 (EDT)
Message-Id: <20020903.104028.127138776.pfps@research.bell-labs.com>
To: jjc@hpl.hp.com
Cc: www-webont-wg@w3.org

From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: Re: TEST: Functional and InverseFunctional tests for approval
Date: Mon, 2 Sep 2002 20:51:08 +0200

> 
> Peter:

[...]

> The previously accepted test was very like that:

This test was accepted on 1 August, but has not yet appeared in the list of
accepted tests in http://www.w3.org/2002/03owlt/

The only test in this list of accepted tests is one approved on 30May,
which uses the ont: namespace, and thus needs to be changed.  It also has
several other problems.

[...]

> Differences I see are:
> - I use a greater variety of RDF/XML than you do
> - you choose to use a number of owl:Thing's
>   (these may prove a significant difference in the light of Pat's work)

This brings up a point that is currently being discussed in RDF Core.  I
agree with the sentiment that the use of RDF and XML in tests should be
absolutely minimal, which implies, among other things, that XML base not be
used unless necessary and that all URIrefs be absolute except where
necessary, and I can't imagine any cases where OWL testing needs relative
URIrefs, except testing the use of rdf:ID in OWL.

> I thought it worth for both FunctionalProperty and InverseFunctionalProperty 
> to show one test that was semantically simple by relying on sameIndividualAs 
> and one test that was simple from a layering on RDF point of view by only 
> using one owl feature.

Why?  I don't see any point in doubling the tests this way.

> > PS:  Given that you have agreed that there are mistakes in the tests, are
> > you going to propose that they be unapproved?
> 
> Hmmm ...
> (that could have done with a smiley !)

I'm serious about this.

> I think we have identified a very minor mistake in the xml:base, in that it 
> does not impact the graph; and we have been discussing the style of the 
> wording of the descriptions.

Plus points on the use of XML and RDF, plus a point on what kind of tests
to do.

> For both I would tend to view these as within the remit of editorial fixes 
> that need to be notified to the group but not subject to a revote.

I would go along with this if there was a documented process, but in the
absence of such any change is significant.

> I take your criticism of a lack of a documented process seriously. Without 
> that it is not clear how much editorial leeway there should be, nor who 
> should exercise it.

Without a documented process I don't see why we should be approving tests
at all.  

> Jeremy

peter
Received on Tuesday, 3 September 2002 10:41:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:52 GMT