Hi Peter,
Many thanks for your continued work on this. I think the path to Entity
check still has a problem. For example (01)
Entity
Abstract
SetOrClass
Class
InheritableRelation (02)
Arity checking is a little tricky. For example (03)
BinaryRelation
CaseRole
path (04)
BinaryRelation->valence=2 (05)
Adam (06)
At 12:33 PM 3/13/2005, Peter Denno wrote:
>Hi Adam,
>
>Attached is a new testing report that includes arity checking. Nothing failed
>the arity check, though as you will see, it appears that the arity of some
>relations is not known. I haven't investigated this, so perhaps it is
>spurious.
>
>I think I have addressed the other bugs we found, but if not, let me know.
>
>
>On Friday 11 March 2005 12:58, you wrote:
> > Hi Peter,
> > Another useful diagnostic would be to check that relations are of the
> > proper arity. A common typo I've made is, for example
> >
> > (sublcass Foo Object)
> > (documentation "A Foo is a roly-poly creature often found in tall grass
> > ...")
> >
> > where documentation takes two arguments and I've left off the first one.
> >
> > Adam
> >
> > At 03:16 PM 3/10/2005, you wrote:
> > >On Thursday 10 March 2005 13:16, you wrote:
> > > > Hi Peter,
> > > > I think there's a bug. You'll need to get the transitive relation
> > > > through "subRelation". For example
> > > >
> > > > BinaryPredicate
> > > > measure
> > > > length
> > > > height
> > >
> > >Hi Adam,
> > >
> > >Good point. I'll write a function all-subrelations. There are actually two
> > >ways it has to be applied I suppose:
> > >
> > >1)
> > >K> (superrelations 'ki::|height|)
> > >(KI::|length| KI::|measure|)
> > >
> > >2)
> > >K> (all-subrelations 'ki::|instance|)
> > >(KI::|element| KI::|immediateInstance|)
> > >
> > >So I have to check whether height or one of its subrelation is an instance
> > > or one of its subrelations. Here is the new output. "whatever" was a
> > > false-positive test I added and "and" is a bug in my clausifier. So the
> > > "everything is an instance of something" test comes up clean. More output
> > > attached.
> > >
> > >I'd look at it more, but I've gotta run.
> > >
> > > > Adam
> > > >
> > > > At 10:11 AM 3/10/2005, Peter Denno wrote:
> > > > >Adam,
> > > > >
> > > > >I ran the "Everything is an instance of something test" with the wider
> > > > >collection of predicate/function symbols provided by looking through
> > > > > the clausal form. Doing that, I get:
> > > > >
> > > > > *** Test: Everything is an instance of something...
> > > > > pointOfFigure is not an instance of anything.
> > > > > connectsEngineeringComponents is not an instance of anything.
> > > > > engineeringSubcomponent is not an instance of anything.
> > > > > stays is not an instance of anything.
> > > > > home is not an instance of anything.
> > > > > attends is not an instance of anything.
> > > > > length is not an instance of anything.
> > > > > height is not an instance of anything.
> > > > > bottom is not an instance of anything.
> > > > > top is not an instance of anything.
> > > > > subset is not an instance of anything.
> > > > > and is not an instance of anything.
> > > > > exactlyLocated is not an instance of anything.
> > > > > resource is not an instance of anything.
> > > > > true is not an instance of anything.
> > > > > instrument is not an instance of anything.
> > > > > represents is not an instance of anything.
> > > > > component is not an instance of anything.
> > > > > result is not an instance of anything.
> > > > > piece is not an instance of anything.
> > > > > side is not an instance of anything.
> > > > >
> > > > >"and" being on the list shows that my clausifier has a bug!
> > > > >
> > > > >--
> > > > >Best Regards,
> > > > > - Peter
> > > >
> > > > ----------------------------
> > > > Adam Pease
> > > > http://www.ontologyportal.org - Free ontologies and tools
> > >
> > >--
> > >Best Regards,
> > > - Peter
> >
> > ----------------------------
> > Adam Pease
> > http://www.ontologyportal.org - Free ontologies and tools
>
>--
>Best Regards,
> - Peter
> (07)
----------------------------
Adam Pease
http://www.ontologyportal.org - Free ontologies and tools (08)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-dev/
Subscribe/Unsubscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontolog-dev/
Shared Files: http://ontolog.cim3.net/file/work/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-dev@xxxxxxxxxxxxxxxx (09)
|