RE: [wsbpel] Groups - BPEL Goals and Nuances for 1st F2F.ppt uploaded

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

Message text written by INTERNET:jevdemon@microsoft.com
>
Document Description:
Presented by Satish Thatte (Microsoft) at the 5/28 F2F in Armonk, NY.
<<<<

Satish,

Thanks for providing this.    Finally we have something that 
approximates to some requirements - except at one or two
levels down into the technology layers, below the behaviours
and scope layers.

What your analysis emphasizes is the need to have those 
higher levels of requirements - as a means to resolve the
issues in the technology layers.

And then there is the implementation models - we appear
to have at least two technical mechanisms - and that makes 
one nervous in terms of the robustness and interoperability.

BPSS solved this beautifully for the eBusiness domain by
constraining the model to actions, steps and binary couplings,
so that interaction between partners is instantly discernable,
and the scope of impacts is limited within an action only.

It seems to me - that if you adopt this BPSS model for external
interactions - at least for V1.0 of BPEL - you will dramatically
reduce the complex of the spec', and also solve a lot of 
your requirements related issues.

Then internally - people can exploit BPEL (which probably
is where they want to do this anyway) - while constraining the
external effects of this.

Again decoupling from WSDL will also greatly reduce the
points-of-failure worries - and make for a much cleaner
process model.  Seems like you've made a start in
that direction already - defining the messaging 
behaviour requirements in a neutral way would 
solidify this.

If the F2F can get closer to providing just one thing:

requirements, requirements, and requirements,

then I believe we'll have a much easier task 
ahead of us in unravelling the technology 
Gordonian Knot that is currently in front of us.

Cheers, DW.