> If you consider extra info used by binding b1 to be part of the
> service interface then how could you switch to using binding b2
> of the service?
I don't fully understand your question. If you are referring to the fact that binding b1 specifies a non-portType part while binding b2 does not, the same can happen to a portType part: binding b1 can include it and binding b2 can leave it out. Where is the difference?
By the way, there is never a guarantee that the variables currently specified in BPEL activities are meaningful with any possible binding. In some bindings, parts belonging to the corresponding message will be missing. As a special case, the whole contents of the variable might be missing on the wire.
Ugo
> -----Original Message-----
> From: Sanjiva Weerawarana [mailto:]
> Sent: Monday, November 24, 2003 7:48 PM
> To: Ugo Corda; Satish Thatte; Francisco Curbera
> Cc: Ron Ten-Hove;
> Subject: Re: [wsbpel] Issue - 77 - Under specified operation
> definitions
>
>
> "Ugo Corda" <> writes:
> >
> > Yes, I understand your position. My point is that it is just your
> > personal position, not supported by anything said in the spec. You
> > cannot expect designers to base their choices on anything outside
> > the spec, no matter how reasonable those external
> argumentations are.
> > That's one of the rules of the standards game ...
> >
> > Ugo
>
> This is not my personal position.
>
> What you pointed to is a section of the spec that talks about
> bindings and what they can do. I agree with you that bindings can
> refer to other messages. However, those are *not* part of the
> abstract interface of the service; those are simply extra information
> that a given binding may use.
>
> If you consider extra info used by binding b1 to be part of the
> service interface then how could you switch to using binding b2
> of the service? Your approach will make BPEL be tied down to a
> specific binding at the time of modeling - which I would consider
> extremely bad practice.
>
> Sanjiva.
>
>
