Re: simultaneous execution

These are not new objectives. IMO the only issue is, some of the 
facilities you were talking about (e.g. identification of choreography 
instance) are not a standard part of WSDL nor of common message formats. 
So there's potentially a problem there. One possibility (I think this 
was suggested) is that we decide these are abstract properties and we 
basically defer the problem of realizing them at a concrete message 
level, which I could support. If you want to go in a different direction 
than this, then I need convincing.

--Jon

Burdett, David wrote:
> I think we are actually agreed on two objectives:
> 1. The need to create choreography definitions that are indpendent of 
> the message format, and
> 2. To define how choreography definitions are bound to Web Services and 
> WSDL in particular.
>  
> I don't think these objectives are mutually exclusive and we should be 
> able to do both. Does anyone disagree?
>  
> David
> 
>     -----Original Message-----
>     *From:* Ugo Corda [mailto:UCorda@SeeBeyond.com]
>     *Sent:* Wednesday, August 06, 2003 3:06 PM
>     *To:* Cummins, Fred A; Burdett, David
>     *Cc:* public-ws-chor@w3.org
>     *Subject:* RE: simultaneous execution
> 
>>+1 to defining how WS-Choreography binds to Web services.
> 
>>The Charter specifically says: "The language(s) should build upon
>     the foundation of the WSDL 1.2".
> 
>>WSDL 1.2 defines interfaces and end points. If we don't at least
>     define some precise mapping between WS-Choreography and WSDL
>     interfaces, then I don't see in which >way we are building "upon the
>     foundation of WSDL 1.2".
>     [FAC] I believe we can do that without sacrificing broader
>     applicability of the choreography.  I'm more concerned that we not
>     link the choreography to the message formats. 
> 
>     If your concern is about about linking the choreography to the
>     message formats on the wire, I would like to point out, as I have
>     done before, that a portType/interface message structure does not
>     imply any commitment to a format on the wire. It's only when the
>     portType/interface is bound to a particular service that the wire
>     format is defined.
> 
>     Ugo
>

Received on Wednesday, 6 August 2003 18:51:45 UTC