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

Re: TEST: last call candidate (?)

From: Jeff Heflin <heflin@cse.lehigh.edu>
Date: Mon, 12 May 2003 13:21:55 -0400
Message-ID: <3EBFD833.36A5DC80@cse.lehigh.edu>
To: Jeremy Carroll <jjc@hplb.hpl.hp.com>
CC: Jeremy Carroll <jjc@hpl.hp.com>, www-webont-wg@w3.org

Jeremy Carroll wrote:
> 
> Jeff Heflin wrote:
> 
> > Jeremy Carroll wrote:
> >
> >
> > ... snip ...
> >
> >
> >>Could I ask Jeff to review the imports tests:
> >>http://www.w3.org/2002/03owlt/editors-draft/draft/byFunction#function-imports
> >>
> >
> > Jeremy,
> >
> > I've reviewed the imports tests. A few comments:
> >
> > 1) You use single quotes instead of double quotes around attributes in
> > all of these tests
> 
> XML permits either; I hope that the tests overall use both.
> No change planned.
> (While we have deliberately avoided some of the variability in RDF/XML,
> there is still quite a lot that we are allowing - there is a tension
> between: (a) allowing people to read the tests without being expert in
> everything and (b) having tests that adequately exercise the stack on top
> of which an OWL system sits. I think this particularly case does a bit of
> (b) without adding any confusion)

That's fine if you want tests to demonstrate both forms, but I think
that you should either have a single set of syntax tests that do this
(independent of test other features), or else in each category of tests,
have some tests that do it one way and some that do it the other. As it
is, the test document seems to suggest that imports requires the single
quote form.

> >
> > 2) Test 002 would probably benefit from having a support-002-A document.
> > As it is, it is not clear if the non-entailment is because the support
> > document doesn't exist or because it is not imported. By having such a
> > document, this becomes clearer.
> 
> Hmmm,
> slightly tricky.
> The file does in fact exist if you click on the URL, but you are right it
> is part of the test and should be included in the printout. My code does
> not allow for this case, I think I will have to do a one off hack of the form
>    if ( testURL.equals(.....imports/ 002 )  {
> // put in support-002-A doc
>    }
> 
> No - that's not really sufficient, the zip file and the manifest files
> should show support-002-A as part of the test: I need to get back to you on
> this.
>
> >
> > 2a) The text of section 3.8 should be changed to allow for the above
> > comment. Change "The support documents are in the imports closure of the
> > premises document." to "The support documents may be in the imports
> > closure of the premises document."
> 
> This is in fact a non-import non-entailment test, .... I will propose some
> text later today or tomorrow.

Okay, I will await your proposal. Sorry to throw this wrench into
things...
 
> >
> > 3) I don't quite understand 004. Why is it that imports004 is Full?
> >
> 
> Answered by Sean
>

It might make things clearer if a more obvious OWL Full document was
chosen, perhaps one that uses a class as an instance? The fact that
using rdfs:Class makes a document OWL Full is very subtle and seems more
appropriate for a Language Compliance Level test.
 
> > 4) Test 005 should probably say "If an OWL Lite document imports an OWL
> > DL document then it becomes OWL DL." in the description. This is more
> > consistent with the description of Test 004.
> >
> 
> Will change as suggested.
> 
> > 5) In Test 008, you should make it clear in the description that
> > imports008 is an ordinary RDFS document and that type declarations are
> > added by writing them in a document that imports the RDFS document.
> >
> > Jeff
> >
> 
> Howabout,
> 
> "It is often possible to import an ordinary RDFS document
> unchanged into an OWL Lite document. Aditional type
> declarations may be needed in the importing document."

Sounds good.

> Jeremy
Received on Monday, 12 May 2003 13:22:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:00 GMT