W3C home > Mailing lists > Public > public-swbp-wg@w3.org > May 2004

Re: Close to final draft of "classes as values" note

From: Natasha Noy <noy@SMI.Stanford.EDU>
Date: Wed, 19 May 2004 16:52:13 -0700
Message-Id: <8CD34175-A9EF-11D8-A8F8-000A958B5C28@smi.stanford.edu>
Cc: swbp <public-swbp-wg@w3.org>
To: Alan Rector <rector@cs.man.ac.uk>


> Although Approach 3 is my preferred approach for many purposes, I have 
> problems with the details of the explanation.

It's actually approach 4 -- I had a typo at  "Considerations".

> owl:someValuesFrom really means "at least 1" and can include many.   
> What is being said formally would be better identified as 
> "Unidentified Lions" or maybe "Unidentified Lion(s)"

Of course. Fixed.

> There is nothing in the semantics, nor in the approach, that requires 
> the singular.  Likewise further down in the text under 
> "Considerations", bullet 1, "a specific lion" should read something 
> like "one or more specific lions".

Yes, you are right.

> More seriously, I don't think this is approach has anything to do with 
> "Prototypes".  There is nothing in OWL semantics - Lite, DL, or Full - 
> that deals with prototypes in the usual sense of being a 
> representation of the default characteristics of a thing which can be 
> over-ridden by further knowledge - either using classic frame style 
> inheritance with exceptions or some form of default logic.    I think 
> raising the notion here just confuses matters.  If we want to 
> represent prototypes or 'prototypical individuals' and have them 
> behave as in expected ways, then we need additional semantics not 
> available in OWL.  (This is one of the reasons for wanting to be able 
> to combine Protege and OWL to build more expressive "knowledge 
> resources" with OWL ontologies at their core, but that is not an issue 
> for here.) It may be that the standards comes back to reconsider 
> prototypes, defaults, and exceptions at some later date, but that is 
> out of my understanding of the scope of SWBP.)
> My explanation of Approach 3 bullet 2 would simply be that, in 
> practice, any book about lions is really about  'some lions'  - even 
> if that happens to be all lions.  I would omit all mention of 
> "prototypes".

I am not sure the use of "prototype" here in the explanation is all 
that confusing and clashes with the "prototypes: as you talk about it. 
However, it's one more of those words that has too much ingrained 
meaning and best be avoided. I like your rewording suggestion and have 
incorporated it

Received on Wednesday, 19 May 2004 22:31:27 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:09:38 UTC