ISSUE-16 (entity annotations)

In ISUE-16, Jeremy points out an asymmetry in the treatement of
annotations in OWL 1.1.

There are two kinds of top-level "annotations" in OWL 1.1, 
1/ Entity Annotations, which are axioms
2/ Annotations on Ontologies, which are not axioms

I agree that this asymmetry is not ideal.

There appear to be two ways to recover symmetry here:

1/ Make Entity Annotations not be axioms, which, to retain symmetry,
   would probably end up not allowing annotations on entity annotations
   (but leaving annotations on entities), so one could say:

   EntityAnnotation(OWLClass(Person)
     Label("People") Comment("The Class of People"))

   but *not*

   EntityAnnotation(Annotation(creatorsName "Peter F. Patel-Schneider")
     OWLClass(Person) Label("People") Comment("The Class of People"))

2/ Make Annotations on Ontologies be axioms, which would probably allow
   annotations on entity annotation axioms, and also allow annotations
   on ontology annotations, as in

   Ontology(<http://foo.ex/foo>
     Comment(Annotation(creatorsName "Peter F. Patel-Schneider")
             "An ontology about nothing at all"))

2a/ The second option could be extended to allow annotations on any
    annotation, as in 

   EntityAnnotation(Annotation(creatorsName "Peter F. Patel-Schneider")
     OWLClass(Person)
     Label(Annotation(creatorsName "Peter F. Patel-Schneider") "People")
     Comment(Annotation(creatorsName "Bijan Parsia") "The Class of People"))
    
I think that I would vote for option 2a, even though it is the largest
changed to the current situation.

Peter F. Patel-Schneider
Bell Labs Research

PS:  I make not claims about the suitability of the actual annotations
     in the examples above.

PPS: I expect that there is a missing "not" after "should" in the
     description of the issue in the issue list.

Received on Monday, 14 January 2008 15:20:25 UTC