Data model and formal semantics

I see by our charter that we are required to supply a (fairly  
constrained by the charter) policy data model with attendant  
processing model. I would like that we met these charter requirements  
by means of a formal semantics, preferably denotational, perhaps by  
translation to an appropriate expressive logic.

I have done this for an older version of WS-Policy, see the papers  
and presentations available here:
	<http://www.mindswap.org/2005/services-policies/>

(I do not, at all, insist on OWL as the formalism of choice. What we  
use is determined by the expressivity we desire to standardize.  
However, I think it's extremely valuable to have a clear and clean,  
preferably machine verifiable, semantics. For one, it eliminates  
ambiguity and confusion. Second, it makes defining interesting  
services easier (see, for example, <http://www.mindswap.org/papers/ 
2005/Policy-ISWC05.pdf> for such services as policy containment,  
equivalence, incompatibility, and incoherence). Third, it helps  
suggest new functionality and expressivity (e.g., in <http:// 
www.mindswap.org/papers/2005/Policy-ISWC05.pdf> we point out that  
being able to express policies as *refinements* of more general  
policies is valuable, esp. in collaborative settings).)

Manchester is willing to supply expertise and effort for a  
formalization.

I'm afraid I'm not at the F2F and am sort of on vacation. But if  
folks wanted to discuss it, I'd happily call in at a specific time  
for it. Otherwise, it's hit or miss.

Cheers,
Bijan.

Received on Tuesday, 11 July 2006 15:17:42 UTC