[Top] [All Lists]

[ontolog-forum] Re: SUO KIF testing

To: Adam Pease <adampease@xxxxxxxxxxxxx>
Cc: ontolog-forum@xxxxxxxxxxxxxxxx
From: Peter Denno <peter.denno@xxxxxxxx>
Date: Thu, 10 Mar 2005 11:50:51 -0500
Message-id: <200503101150.51953.peter.denno@xxxxxxxx>
On Thursday 10 March 2005 11:28, you wrote:
> Hi Peter,
>    Many thanks for doing this.  Is the code Java?  If so, we could include
> it in Sigma as a call from the Diagnostics.jsp page.    (01)

No. Unfortunately it is in lisp, but I might include it in my UBL validation 
tool, if we find that it is useful.    (02)

> I'll look at the 
> report and see what needs to be addressed.    (03)

There are of course some spurious reports. (e.g. Entity is not in the type 
hierarchy of Entity). And the motivation behind the test of ensuring that 
things in a disjointDecomposition are subclass of the thing decomposed is 
probably flawed.    (04)

Also, I just noticed that the "instance of something" test is not thorough. I 
only test on predicate/function symbols that I pick as the first symbol in an 
axiom. Thus I pick up p in (p ...)  but not q nor r in (=> (q...) (r...)). I 
have a clausifer for SUO KIF. I'll see if I can't pick up the symbols from 
the clause normal form.     (05)

> Adam
> At 07:58 AM 3/10/2005, Peter Denno wrote:
> >Adam, et al.,
> >
> >I wrote a small tool to check a few aspects of the 'structural integrity'
> > of ontologies written in SUO KIF. The attached file is the output against
> > SUMO 1.72. The tool is not well tested (it's a day's work) but I thought
> > that getting this out there might start a dialog to improve it. (For that
> > matter, I was looking for the correct forum for this dialog. Ontolog
> > might not be it, but scanning old email I couldn't think who else to send
> > it to!).
> >
> >Regarding what the tool reports, there are several reports under the test:
> >
> >  *** Test: ?x in (domain <Relation> <num> ?x) isn't one of the other
> > classes that are in the disjointDecomposition of Abstract.
> >
> >This concerns, as we have discussed, the problems of the axioms:
> >
> >(disjointDecomposition Abstract Quantity Attribute SetOrClass Relation
> >Proposition)
> >
> >(domain domain 3 SetOrClass)
> >
> >The test reports only those instances where the disjointDecomposition is
> >violated. There are plenty more situations where something other than an
> >Abstract (e.g. Agent) is used in (domain domain 3...). I suppose that what
> > is intended in those situations is "SetORClass of"  (e.g. (domain
> > experiencer 2 Agent) means "SetOrClass of Agent") ... but that seems like
> > an
> >extra-theoretical interpretation.
> >
> >
> >--
> >Best Regards,
> >- Peter
> >
> >Peter Denno
> >National Institute of Standards and Technology,
> >Manufacturing System Integration Division,
> >100 Bureau Drive, Mail Stop 8264             Tel: +1 301-975-3595
> >Gaithersburg, MD, USA 20899-8264          FAX: +1 301-975-4694
> ----------------------------
> Adam Pease
> http://www.ontologyportal.org - Free ontologies and tools    (06)

Best Regards,
 - Peter
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (07)

<Prev in Thread] Current Thread [Next in Thread>