Re: exchanging OWL through RIF

Hi, being a bit late in this discussion, let me try to briefly
describe my point of view:

- it is not RIF's job to encode OWL ontologies, but
- we need to be careful how RIF interacts with OWL ontologies since
there are various such interactions, and they differ in their  
consequences.
I suggest to make these interaction explicit (for Phase 1, we might  
want to
restrict ourselves to one such interaction mode):

1) [bi-directional interaction, see http://www.springerlink.com/index/ 
3AH2YPJ3P628FT4M and http://www.dis.uniroma1.it/~rosati/publications/ 
Rosa05.htm]

In this interaction mode, OWL and rules live in the same "world",
consequences from rules impact on consequences of OWL and, vice versa,
consequences from OWL impact on consequences from rules. An example  
is SWRL,
others (including decidable, non-monotonic ones are linked above) are  
possible and
well-understood.

We can distinguish 2 sub-classes:
1a) variables in rules stand for named objects only (ie, those  
explicit mentioned in our ontology/gound facts). Rosati's and DL-safe  
rules are of this kind
1b) additionally, so-called "may-bind variables" stand for any  
individual, including those not mentioned explicitly. SWRL is of this  
kind

2) [uni-directional such as the INFOMIX approach, Thomas Eiter,  
Thomas Lukasiewicz, Roman Schindlauer, Hans Tompits: Combining Answer  
Set Programming with Description Logics for the Semantic Web. 141-151]

In this interaction mode, rules only query an OWL ontology. Hence
consequences from OWL impact on consequences from rules,  but
not the other way round. An example is INFOMIX.

So, can we capture these differences? I would suggest that, for phase  
1, we restrict our attention
to 1a or 1b: we simply need to allow "imports" statements in a RIF  
ruleset that imports an OWL ontology.
For phase 2, we might want to extend to (2) which could be realized  
by allowing literals in a rule body to call out/import
from an OWL ontology...

Cheers, Uli