RE: Definition of Choreography

> -----Original Message-----
> From: Paul Prescod [mailto:paul@prescod.net]
> Sent: Saturday, October 19, 2002 1:53 AM
> To: Burdett, David
> Cc: Champion, Mike; www-ws-arch@w3.org
> Subject: Re: Definition of Choreography
> 
>
> the state where they meet your preconditions.
> 
> Customer: "Garcon, your service was terrible."
> Waiter: "You asked for a coffee and a salad. I brought them to you."
> Customer: "No, I asked for a salad and a coffee. But you brought me a 
> coffee and THEN a salad."

Hmm ... in the absurdly hypothetical situation where I come into a
restaurant and let my Bluetooth or WiFi-enabled PDA deal with an automated
waiter via SOAP messages, I would disagree.  I might want water and salad,
then a a main course with wine, and then dessert with coffee.  I for one
would be highly annoyed if dinner started with dessert and wine, then I got
salad and coffee, and ended with a main course and water. I would think it a
bug in the Restaurant Service if there was no way to specify my preferred
ordering.

Oops, I see that Chris has made a similar point in more detail.  Oh well, +1
to Chris.  I also liked the "deli" vs "fine restaurant" distinction.  Right
now, it sounds like web services standards are up to the "deli" level ...

> 
> Tim Bray says that Web Services are in danger of having more 
> specifications than running applications. 

Tim has a point, and to pick up another of Tim's mantras, I think we should
be focussed on doing the "minimum required to declare victory" here.  Not to
draw too close a parallel with current events :-) there are obviously some
battles that one wins by staying out of them.  On the other hand there are
some battles that you can win if you act quickly and decisively before the
situation degenerates into polarization and chaos. In the case of
choreography, I think we're in the latter situation.  

I'm (taking co-chair hat off!) not too impressed by either WSCI or BPEL.
It's clear from this discussion that we need to separate out the
WSDL-extensions that would *declare* what an external consumer needs to know
about the choreography of a complex service before we worry about
standardizing the process by which they are executed.  It's also so hard to
keep all the tangled-up concepts in these specs straight that I'm beginning
to hope that OWL or some similar formalization can help us maintain order,
especially since it's a little hard to visiualize diagrams in
higher-dimensional space or project them onto 2 dimensional drawings.

So, I think that web services choreography is a battle that has to be fought
in order to declare victory, and we'd better fight it soon or there WILL be
polarization and chaos. On the other hand, the "minimum required to declare
victory" may be considerably less that what is in BPEL or WSCI, and I agree
that a focus on some use cases could help us (and the AC) identify what that
minimum is.

Received on Saturday, 19 October 2002 11:44:24 UTC