Re: UFDTF Metamodeling Document

From: "Conrad Bock" <conrad.bock@nist.gov>
Subject: UFDTF Metamodeling Document
Date: Wed, 28 Nov 2007 09:34:02 -0500

> 
> 
> UFDTF,
> 
> Some comments about metamodeling in the user-facing documents:
> 
>   - I would expect the draft metamodel document
>     (http://www.w3.org/2007/OWL/wiki/MOF-Based_Metamodel) is part of the
>     user-facing documents, at a similar level and audience as the
>     reference manual.  It's important that this and the OMG's Ontology
>     Definition Metamodel for OWL are aligned, because OMG is perceived
>     as the primary source for MOF-based metamodels.

Is where any particular reason to make the metamodel document a
recommendation track document?  

Also, the document doesn't appear to have a pointer ot the actual
mdetamodel itself.

>   - We should also review and clarify the statement about metamodeling
>     in the submitted overview:
> 
>       In OWL 1.1 a name (such as Person) can be used as any or all of an
>       individual, a class, or a property. The computational problems
>       that would arise if this were treated as in RDF are avoided by
>       ensuring that no aspect of the use of the name as an individual
>       has any effect on the meaning of the name as a class. Such a
>       treatment of metamodeling is often called punning.
> 
>     Even with no metamodeling, the use of a name as an instance of
>     owl:class and as the type for an individual are of course related.
>     For example, if M1:car is an instance of owl:Class, it is also the
>     type for M0:johns-car.  This is also true for metamodeling, see
>     next.

I don't understand this in an OWL 1.1 context.  In the functional syntax
for OWL 1.1 there is no owl:class.

If by M1:car you are referring to some MOF or OMG constructs and not
some IRI, then please include a pointer to some place that provides an
understanding of what these constructs mean.
 
>   - Finally, we should address a common use of metamodeling that extends
>     OWL for modeling languages.  This defines subclasses of owl:Class
>     with additional properties that have values on the instances of the
>     subclass.  For example, uml:Class as a subtype owl:Class would add
>     characteristics of classes peculiar to UML, such as isAbstract.
>     This does not imply any DL reasoning on owl:class its subtypes
>     (these are only used to type instances, which are only created,
>     read/queried, modified, and deleted), or introduce any constraints
>     that would affect DL reasoning on the instances of instances of
>     owl:class.

How would these extensions work in OWL 1.1?  How could it be guaranteed
that there are no effects to the DL reasoning paradigm?

(I can see at least one way of setting up this sort of thing in OWL 1.1,
but I don't know whether it would suit this usage because I don't know
what is supposed to happen.)

> Some earlier discussion of metamodeling in OWL is at
> http://www.w3.org/2007/OWL/wiki/index.php?title=OWLMetamodeling.
> 
> Conrad

peter

Received on Wednesday, 28 November 2007 16:10:38 UTC