Re: Addiitonal syntax tests

From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: Addiitonal syntax tests
Date: Wed, 24 Sep 2003 18:41:41 +0300

> 
> 
> Hi,
> 
> I am cc-ing Michael and Bijan given their Pellet reports ...
> There is a note at the end about test I5.8/016 which just reiterates what has 
> already been in the WG list.
> 
> Having made slower progress on my own syntax checker than some others, I am 
> only now at the stage of trying to get all the bugs in it.
> 
> There were a few things I knew I had not coded up, but which did not show in 
> failed tests - in particular the optional type triples on descriptions and 
> restrictions.
> 
> So I have now added some tests in this area

OWLP results syntax results

Test			Required	Determined

Class-005		Lite/Lite	Lite/Lite
Class-006		Lite/DL		Lite/DL
Restriction--005	not in the current manifest file?
Restriction-006		not in the current manifest file?

(I'm assuming that there is a uniform mapping from these shorthand names to
official names.  Otherwise I can't figure out which test these correspond
to at all.)

> (these also better exercise the comprehension rules).
> 
> True to my expectations my code fails three out of four of these tests, and 
> the one it passes it passes for the wrong reason.
> 
> ===
> 
> On I5.8-016, this test emerged from discussion in the WG about the syntax of 
> user defined datatypes. I suggested that what the OWL CR says was a bug, 
> Peter seemed to argue it was a feature. 

I actually argued only that it has not yet been determined to be a bug, and
that there is little reason to worry about it, as there is no effective way
of using datatypes, except for the built-in XSD ones, which are built in to
OWL anyway.

> DanC asked what the implementations 
> do. Since we didn't know I wrote a test; the test proposes that this should 
> be in Full; the OWL CR says Lite. OWLP and Jena both say Lite (even though I 
> wrote the Jena syntax checker?).
> 
> I believe that the tests I5.8-013,014,015 to be correct, and to exercise 
> related parts of the problem.

Validated by OWLP.

> Jeremy

peter

Received on Wednesday, 24 September 2003 13:32:24 UTC