Re: SEMANTICS: Agenda

Francis McCabe wrote:

>
> This is a kickoff message for the formal models/semantics task force. 
> My aim is to lay down a simple agenda for the work of the task force.
>
> While semantic issues are long and deep (never mind piranhas, think 
> Great White Sharks), I see two key areas that relate to 
> choreographies, neither of which needs to end up as shark bait:
>
> 1. Semantic models of choreography.
>
> The idea being to adopt model(s) that help to inform us in our choice 
> of choreography language constructs. It does not seem to be necessary 
> that any of the constructs of a modeling notation be directly 
> reflected in the choreography language; rather the modeling notation 
> should help us to be precise about the features of the choreography 
> language.
>
> 2. Semantics of Web services.
>
> The idea here is to suggest ways in which information which relates to 
> the semantics of Web services should be integrated with any 
> choreography language. The purposes of such information might be:
>
> a. To permit automatic discovery of Web services based on the 
> semantics of the input, output and behavior of the Web services.
> b. To permit semi-automatic composition of Web services based on the 
> semantics of the input, output and behavior of the Web services.

mm1: (a) and at least a portion of (b) look to be in the scope of a 
registry.  How do we differentiate this from any scope we would consider?

> c. To permit tools to verify certain aspects of executing 
> choreographies (such as whether privacy tokens are properly propagated).
> d. To permit better documentation of Web services, especially in the 
> context of complex choreographies.
>
> Note that the W3C already has a strong commitment to enhanced 
> semantics -- via the Semantic Web initiative. Therefore, the real 
> effort in WS-CHOR in relation to 2. would be adapting the choreography 
> language to enable an integrated use of OWL/RDF etc. Something similar 
> is happening in WSD at the moment.
>
> The first of these goals is probably higher priority at the moment; 
> although both are essential. <duck/>
>
> To get the ball rolling, I invite interested parties to submit a short 
> (2 page) easily digestible summary of their favorite formal model; 
> together with some proposal for its use. This would be used to help us 
> to get educated about the various modeling approaches and help us make 
> the best choices.
>
> Frank McCabe
>

Received on Sunday, 6 July 2003 16:12:05 UTC