RE: [wsbpel] A Topic for the F2F?

["J. Matthew Pryor" <matthew_pryor@versata.com>]

Considerable time has been spent on this subject at many BPMI meetings

In my mind, it boils down to a discussion of whether or not you need a
different meta-model to support the kind of modeling you *want* to
support.

To Steve's point, if you want to allow more business focused people to
use a highly productive design environment to specify business
processes, then you almost certainly need a different meta-model than
that which underlies BPEL. In this case, it would be sensible to then
develop [canonical?] mappings between the high-level notation meta-model
and the execution focused BPEL (this is what is currently being done
with BPMN for BPEL & BPML - sorry about all the acronyms). This allows
for dealing with issues such as graph oriented versus block structured
etc. I encourage people to talk with the folks who have done a lot of
excellent work on the BPMN to tap into their experience. The output of
the OMG RFP that Fred mentioned would also fill this bill.

If on the other hand, if what you want is a standardized way to
represent the actual BPEL meta-model, then yes there is almost certainly
other information that will be needed (text annotations, icons, position
information) and there is possibly some value in standardizing this;
there will be valid arguments for and against storing this additional,
visually oriented information inside a BPEL file or as an external
document that references the BPEL. This could also be left to vendors to
implements, especially if there is a simple extension mechanism for
adding tools specific data into a BPEL file.

Perhaps these thoughts will help focus discussion about which if any of
these issues might be addressed in this TC.

Matthew