Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: SUO: OMG, UML and Ontologies




Pat,

At 06:50 PM 2/28/2001 -0600, pat hayes wrote:

>  Adam Pease <apease@ks.teknowledge.com>:
>>Francis,
>>  I'm not sure where our disagreement lies then.  Maybe in fact we 
>> agree.  Let me try to restate some points which I believe are tied together.
>>
>>1.  Logical languages like KIF are very expressive and ensure that we can 
>>define precisely the semantics of the terms on our ontology.
>>
>>2.  Logical languages like KIF can be translated easily into less 
>>expressive forms such as XML or UML.
>
>Er...that isnt quite true. These less expressive forms can be translated 
>into KIF, but not (in general) the reverse. What is true is that anything 
>that can be said in UML (etc.) can be said in KIF, or automatically 
>translated into KIF.


I was speaking too casually.  You're exactly right, of course.  As a 
practical matter though, folks who like "frame" representations will be 
happy with the (diminished) translated result since if they liked axioms 
they'd be happy with KIF.


>XML isnt really in the same league since it has no semantics: it is really 
>only a notation for arbitrary labelled directed graph structures. KIF can 
>be directly rendered into XML-compliant syntax in several ways, of which 
>the most obvious is to replace each pair of nested (  ) brackets with XML 
>start and end tags <....> </...> labelled with the syntactic class of the 
>KIF expression. <Metacomment> Why anyone would actually prefer this 
>barbaric clumsiness is beyond my comprehension, but apparently many people 
>do. Perhaps their heads don't have enough capacity to maintain a pushdown 
>stack? </Metacomment> I believe there is a standardisation effort under 
>way right now to settle on a generally useful such notational convention 
>for XML-KIF.
>
>>3.  Many people are not familiar with KIF and so they should be able to 
>>work with the SUO in a language in which they feel comfortable.
>
>The very real problems the SUO faces will not be made any easier or more 
>difficult by the choice of notation, which should be a minor matter that 
>we could profitably ignore on this listing. As you know, Adam (having done 
>a good deal of it yourself :-)  there is a great deal of work now in hand 
>on translations between alternative notations, most of which can be done 
>mechanically.

I agree, but folks tend to feel very tied to their particular 
representations, and when they're not familiar with someone else's, they 
worry about whether it will meet their needs.

Adam

>Pat Hayes
>
>---------------------------------------------------------------------
>IHMC                                    (850)434 8903   home
>40 South Alcaniz St.                    (850)202 4416   office
>Pensacola,  FL 32501                    (850)202 4440   fax
>phayes@ai.uwf.edu http://www.coginst.uwf.edu/~phayes

-----------------
Adam Pease
Teknowledge
(650) 424-0500 x571