W3C home > Mailing lists > Public > public-swbp-wg@w3.org > October 2005

Re: [SE] more feedback to new note

From: Daniel Oberle <oberle@aifb.uni-karlsruhe.de>
Date: Mon, 10 Oct 2005 15:37:06 +0200
Message-ID: <434A6E82.1080407@aifb.uni-karlsruhe.de>
To: Holger Knublauch <holger@SMI.Stanford.EDU>
CC: public-swbp-wg@w3.org

Hi again,

>> - there is no real introduction/definition of the word ontology.
>>   You explain it shortly in Sec 2. Shouldn't we give it more focus?
>>   Then we could also mention my other comment on the difference
>>   between UML and ontologies in purpose.
> 
> That may be a philosophical question for the paper.  In one of the calls 
> I already mentioned my belief that the term ontology may have a bad 
> sound to some communities, and I would prefer to use a term that OO 
> people already know, such as "domain model".  For the paper, we take 
> knowledge of OO for granted and try to attract OO developers by building 
> bridges into unfamiliar terrain.  A term like ontology does not 
> necessarily help in this respect.  We should definitely use it, but I 
> wouldn't put it to the center.
Agreed. But as soon as we write "Web Ontology Language" everybody
wants to know what an ontology is, so we must clarify its origin,
meaning and purpose.


> Also, I am not sure I agree that ontologies are built for a totally 
> different purpose than OO domain models.  In many cases, even OWL 
> ontologies contain "simple" class definitions to describe data 
> containers (like customer's email adresses).  Such classes don't 
> necessarily need to be completely generic and don't need to define a 
> shared understanding.  They can, but they can also be used to build data 
> structures so that conventional OO classes can be easier hooked in, and 
> to have a nice XML serialization.
It's not a question what you can and what you can't do with ontologies.
You might use a hammer to open a bottle. But the original purpose of a hammer
is to beat on nails. Of course you can (mis)use ontologies to build data
structures. Their original purpose is to specify the intended meaning of
terms, however. At least if we acknowledge Gruber's as well as Guarino's works.
Vice versa one might (mis)use UML class diagrams to specify the intended
meanings of terms...


Greetings
  Daniel
Received on Monday, 10 October 2005 13:36:39 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:09:44 UTC