Re: metamodelling feature of OWL 1.1 / OWL 2.0

Hiya,

First, I strongly recommend sticking to the OWL2 docs:
	http://www.w3.org/2007/OWL/wiki/OWL_Working_Group

Second, there are two forms of  syntactic higherorderness in OWL 2:  
Punning and annotations (the latter are under flux).

Punning lets you use a term in one syntactic category (e.g., a class  
name) in a syntactic position different from the legal ones for that  
category (e.g., in the subject position of a type triple).

Thus, the following assertion is legal in OWL 2 (though not in OWL 1  
DL):
	:SomeClass rdf:type :SomeClass.

Currently, there is no property punning in OWL 2.

The main difference between this sort of syntactic higher orderness  
and that of OWL Full (aside from the syntactic restrictions on  
properties) is some difference in the semantic conditions. (Here I do  
not speak of redefining the builtin logical vocabulary, but only of  
using user defined classes etc. as instances etc.) In OWL Full, if  
you equate two terms then you can derive that they are equivalent.  
(i.e. sameAs entails equivalentTo).

I recommend Boris Motik's paper:
	http://dip.semanticweb.org/documents/Boris-Motik-On-the-Properties- 
of-Metamodeling-in-OWL.pdf

Punning corresponds to the contextual semantics; OWL full to the  
hilog semantics.

Annotations are different because they do not (in principle,  
necessarily) affect "object" or domain level entailments at all. The  
annotation facilities are still under development:	
	http://www.w3.org/2007/OWL/wiki/Annotation_System

Cheers,
Bijan.

Received on Tuesday, 6 May 2008 08:11:49 UTC