RE: [wsbpel] Issue 53 - Motion for consideration by the TC - Satish issue 3, patterns

["Haugen Robert" <Robert.Haugen@choreology.com>]

> 3.  We can already model these stylized patterns
> using existing BPEL mechanisms as I tried to illustrate
> with my presentation at the F2F (just uploaded to TC site
> under "related resources").

I responded to Satish's related issue 2 previously:
http://lists.oasis-open.org/archives/wsbpel/200401/msg00041.html

I agree with his assertion that we can model transaction patterns using
existing BPEL mechanisms.

However, what are these patterns?  Where will they come from?  How can
they be used interoperably among loosely-coupled participants?  What
ws-bpel features are required for successful implementation?

If we mean patterns in the sense of well-accepted and well-formed
solutions to problems, a la software design patterns
(http://hillside.net/patterns/), those things take a lot of work:
harvesting patterns from best practices, documenting, re-implementing,
testing, and adoption by communities of practice.

I wrote one pattern that took me the best part of a year, including
getting shepherded thru a PLoP (http://hillside.net/shepherding.html).
It is not a trivial process.

I contend, in fact, that the closest thing we have to business
transaction patterns is in the specifications of BTP, WS-T, WS-TXM,
RosettaNet and UN/CEFACT BCF.

I don't think that the interaction styles that Satish illustrated rise
to the level of patterns.  I think they are ill-formed, and that a
well-formed transaction pattern would look like an abstraction derived
from the above specifications.

Which I think is remarkably like the Choreology proposal for including
business transaction management in ws-bpel.