RE: [wsbpel-abstract] Thoughts on abstract BPEL, choreography, etc.

["Satish Thatte" <satisht@microsoft.com>]

On second thought we may want to simply accept that both positions are
valid in certain circumstances, as you also point out in your second
scenario, though you make the statement that the data dependent behavior
requirement is rare.  Perhaps so and perhaps not, but this is the "sweet
spot" for abstract BPEL as I think you are also suggesting.  If that is
agreed to by all then for the technical discussion we can just focus on
how to facilitate this class of usage.

Satish

P.S  You can probably tell that I want to avoid as many "ratholes" as I
possibly can ;-)

-----Original Message-----
From: Satish Thatte [mailto:satisht@microsoft.com] 
Sent: Sunday, June 20, 2004 8:55 PM
To: ygoland@bea.com; wsbpel-abstract@lists.oasis-open.org
Subject: RE: [wsbpel-abstract] Thoughts on abstract BPEL, choreography,
etc.

Interesting.  One of the core discussions is the expressive power needed
for external views.  There are two schools:  data dependent behavior
must be expressible vs mechanical verification of conformance must be
possible.  They don't sound like opposites but they are.  BPEL has taken
the former position.  Your requirements advocate the latter.  This is
the issue I suggested earlier we take up in this group, but it may be
that we will have to try the possibilities in the real world to see
which one is good for what.  In that case, BPEL would be the obvious
candidate for trying the "richly expressive external views" option since
it is naturally there already.

Satish

 
-----Original Message-----
From: Yaron Y. Goland [mailto:ygoland@bea.com] 
Sent: Friday, June 18, 2004 10:22 AM
To: wsbpel-abstract@lists.oasis-open.org
Subject: [wsbpel-abstract] Thoughts on abstract BPEL, choreography, etc.

In our discussions on abstract BPEL I will inevitably refer to a paper I

wrote which deals with issues related to abstract BPELs - 
<http://lists.w3.org/Archives/Public/www-archive/2003May/att-0029/chor.h
tm>.

The paper consists of a set of requirements, three fully fleshed out use

cases and a bunch of smaller articles detailing particular technical 
issues that abstract BPEL will inevitably run into. It then examines 
those issues and makes recommendations for how to deal with them.

	Thanks,

		Yaron