Re: annotation properties

On Jul 11, 2007, at 10:17 PM, Michel_Dumontier wrote:

> Hi,
>
> I downloaded the latest Protˇgˇ 4 so as to further enhance my  
> ontologies with new OWL 1.1 features, and have found myself  
> wondering why I canÕt specify or use my OWL 1.0 defined annotation  
> properties. In brief conversations with Matthew Horridge, heÕs told  
> me something along the lines that OWL 1.1 calls for annotation URIs  
> rather than annotation properties as they were previously defined.
Something in your recounting isn't right (I know know whether it was  
a mis-statement by Matt or a garble in the reporting). In the mapping  
to RDF, annotationUris get mapped to annotation properties:
	http://www.webont.org/owl/1.1/rdf_mapping.html

	annotationURI
gets mapped to:
	annotationURI rdf:type owl:AnnotationProperty

Annotation assertions:

	EntityAnnotation(OWLClass(cID)
   		Annotation(apID1 ct1) ... Annotation(apIDn ctn))

get mapped to
	dID T(apIDi) T(cti)   1 ² i ² n

(hmm. Something a bit wacked in this table entry; been a while since  
I worked with the mapping doc but I presume the subject of the  
triples is implied...oh I see, there are some cut and paste errors.  
Replace dID with cID.

So, let's follow the translation
	EntityAnnotation(OWLClass(Person)
		Annotation(dc:creator "Bijan"))
So, we make a new annotation triple:
	dc:creator rdf:type owl:AnnotationProperty

Then translate the content
	Person dc:creator :Bijan.
	Person rdf:type owl:Class.

>  I thought it was agreed that OWL 1.0 documents would be valid OWL  
> 1.1 documentsÉ whatÕs going on here?

That is the goal; there may be bugs and misunderstandings. If you  
find a bug in the specs (e.g., an OWL DL ontology that is not an OWL  
1.1 ontology), please file an issue on it:
	http://code.google.com/p/owl1-1/issues/list

(I just did on the above typo.)

Cheers,
Bijan.

Received on Wednesday, 11 July 2007 22:57:18 UTC