Re: defconcept versus assertion

Hi Thomas,

The reason for my question is that I'm attempting to describe the PowerLoom 
ontology using CLIF which does not directly support things like defconcept 
and the like.  Therefore, I would like to define concepts using something 
like (subsetOf Dog Animal).  Based on what you've said I should be able map 
this into PowerLoom using (assuming Animal is already defined within 
PowerLoom as a concept):

(defconcept Dog)
(assert (subset-of Dog Animal))

Thanks much,
Cameron.


----- Original Message ----- 
From: "Thomas Russ" <[email protected]>
To: "Cameron Ross" <[email protected]>
Sent: Wednesday, August 22, 2007 12:38 PM
Subject: Re: [PowerLoom Forum] defconcept versus assertion


>
> On Aug 21, 2007, at 7:48 PM, Cameron Ross wrote:
>
>> Are the following statements equivalent in PowerLoom?
>>
>> (defconcept dog)
>> (assert (instance-of dog concept))
>
> Not really.
>
> PowerLoom needs to have the defconcept form because that triggers 
> additional setup that will not happen with just the assertion.
>
> In fact, if you try just the latter, PowerLoom will not have built  the 
> necessary data structures for treating DOG as a concept.
>
> PowerLoom will try to work with undeclared items, but you will print 
> warnings when it uses them.  You will get better results by declaring  all 
> of the concepts and relations before using them.
>
>