RE: [wsbpel] Topics for the "Review input from TC members" session of the F2FSome

[David RR Webber - XML ebusiness <Gnosis_@compuserve.com>]

Mike,

These are interesting assertions.

First of all this sound like something out of
"General Hospital" - "It is not a problem for stateless 
idempotent WSDL operations" - translation into simple
business english, please!?!

Then second of all - there's a lot of assumptions here about the
"whole point" of BPEL - obviously this is fodder for the F2F,
what exactly are we deciding here, if anything!?!

We need some ground rules.  The TC scope of course
completely glosses over these minor details - but we definately
need something drafted for the F2F as a start point - is
someone going to collate these items into something
coherent and formal?

Thanks, DW.
======================================================
Message text written by "Marin, Mike"
>
Well, yes and no. It is not a problem for stateless idempotent WSDL
operations. However, the whole point of BPEL is to support state full
and potentially long running services to run real business stuff.
Therefore, BPEL should suggest a solution or at least identify the
problem for somebody to solve. There are several solutions to this
issue, including the one mentioned by Fred (except that it does not work
under the common case of SOAP under WSDL). Others include simple
variations of two-phase commit to guarantee message delivery.
<