RE: Definition of Choreography

> -----Original Message-----
> From: David Orchard [mailto:dorchard@bea.com]
> Sent: Saturday, October 19, 2002 6:59 PM
> To: 'Burdett, David'; 'Champion, Mike'; www-ws-arch@w3.org
> Subject: RE: Definition of Choreography
>
 
>  We're  certainly getting
> lots of interest in customers for further and ongoing work on 
> bpel4ws, like big bank kind of customers that get our attention.

Do you have any sense of what the "big bank kind of customers" want to see
in the way of STANDARDS for choreography?  Would the declarative/public
interface level being standardized do them any good, or do they need
something at the level of BPEL4WS, "turing completeness" and all?

I keep hoping that there's a "minimum needed to declare victory" somewhere
between BPEL4WS and Paul's precondition/postcondition suggestion that
doesn't seem to meet the needs of the industry [or at least what the
industry perceives to be its needs, but that's something for Paul to educate
our *customers* about before most of us jump on the bandwagon]. 

On the other hand, there are two reasons for wanting some technology
standardized.  The first is the obvious -- people want to be able to
standardize on interfaces and interoperate with (or be able to replace with)
other implementations of those interfaces.  If so, standardizing on the
declarative/interface aspects of choreography is sufficient.

The other is less obvious, but maybe what many customers really want -- to
reap the benefits of the network effect that agreement on a standard brings,
in terms of the availability and cost of tools, expertise, tutorials, and
the general desire to make the easy decision to go along with the crowd
rather than make hard decisions about costs, benefits, and alternatives.  If
so, a standard executable implementation language is necessary to declare
victory.

Thoughts anyone?

Received on Saturday, 19 October 2002 19:30:29 UTC