W3C home > Mailing lists > Public > public-webont-comments@w3.org > September 2003

Re[2]: OWL CR feedback: owl:Class 'vs' rdfs:Class causing pain. Is owl:Class really needed?

From: Benjamin Nowack <office@e-senses.de>
Date: Thu, 18 Sep 2003 11:59:04 +0200
To: Jim Hendler <hendler@cs.umd.edu>
Cc: public-webont-comments@w3.org
Message-ID: <PM-EH.20030918115904.CB1E0.1.1D@192.168.27.2>


Hi,

in my opinion, it would fit well into the Guide and Reference.
For people who understand the AS document the issue may be
obvious but the non DL folks will probably start with the Guide
and continue with the Reference. There is already a NOTE in 
chapter 3.1 of the Reference that is quite close to this stating
that in OWL Lite and OWL DL "owl:Class" must be used for all
class descriptions. The NOTE below explains that owl:Class is 
a subclass of rdfs:Class. Maybe you could just add if this 
means that vocabularies defined in RDF Schema can not be used
(imported, linked) in OWL (Lite, DL) ontologies, neither 
their terms nor their classes (I still hope that I'm wrong..).

In section 7.1 a Dublin Core term is reused as AnnotationProperty

<owl:AnnotationProperty rdf:about="&dc;creator" />

If this is not allowed for rdf schema classes or in no context 
other than AnnotationProperty it could be helpful to explicitly
say that. I saw several ontologies that use something like e.g.

<owl:Class rdf:about="&foaf;Person" />

or even

<owl:Class rdf:ID="Person">
   <owl:equivalentClass rdf:resource="&foaf;Person" />
</owl:Class>

trying to "make" the rdfs:Class of the FOAF spec an owl:Class.
Same for properties:

<owl:DatatypeProperty rdf:about="&dc;creator" />

or (even more confusing)

<owl:ObjectProperty rdf:about="&foaf;mbox">
   <rdfs:range rdf:resource="&rdfs;Resource" />
</owl:ObjectProperty>
(in order to be able to make use of 
  owl:InverseFunctionalProperty)

Whether this was allowed/not allowed could be mentioned 
somewhere in the Guide or the Reference. Maybe in section
3.2.2 (owl:equivalentClass) as people will probably look 
here when they try to map different ontologies/vocabularies.

(I personally don't need a more detailed explanation of why 
rdfs:Class and owl:Class aren't equivalent in OWL Lite and 
OWL DL. Just some notes on what this means for people trying
to build OWL tools or write OWL ontologies.)


greetings,
benjamin

___________________________
benjamin nowack

am exerzierplatz 1
de-97072 wuerzburg




Jim Hendler (hendler@cs.umd.edu) schrieb am 18.09.2003:
>
>Dan-
>  you asked a good question for which, as you see from Pat's message, 
>the answer is more complex than you might think.  From the point of 
>view of our documents/CR let me ask you this - which document would 
>you expect to look in to see a "human readable" answer?  Seems to me 
>my WG should add a paragraph or two that answers in an "informal" way 
>to one of our informative documents
>  -JH
>
>-- 
>Professor James Hendler				  hendler@cs.umd.edu
>Director, Semantic Web and Agent Technologies	  301-405-2696
>Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
>Univ of Maryland, College Park, MD 20742	  *** 240-277-3388 (Cell)
>http://www.cs.umd.edu/users/hendler      *** NOTE CHANGED CELL NUMBER ***
>
>
Received on Thursday, 18 September 2003 05:59:49 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:43:29 GMT