Very tentative plan for ebTWG Commitments project

[bhaugen <linkage@interaccess.com>]

By "very tentative", I mean these are my ideas as of this
moment, subject to all kinds of revisions.  I thought it
would be good to put something down now, though,
so people could have something to think about.

These tasks are not intended be followed in
waterfall fashion.  There are obvious opportunities for
parallelism (note dup numbers) and certainly spirals 
will happen.

1. Gather business examples.
Don't worry about form.  Text is ok.  Any diagram style.
Basically, get examples in whatever form they are in now.
Make it easy. 
  
Sooner or later, get test data to go with leading examples.
Use the examples and test data to guide all subsequent
tasks.

2. Convert business examples to UMM use cases.
Led by expert modelers.  

Focus on the commitment-fulfillment aspects of 
the business examples; lose whatever is off-topic.

2.  Do roughcut commonality and variation analsis
on examples to harvest first-draft list of patterns.

3. Take business examples one-by-one in priority
sequence thru UMM views from BOM to BTV.
Use concepts from REA and ISO-Open EDI as
well as UMM.  

4. Generalize a metamodel and business object model
for commitments and fulfillments and their main
collaborators (as in my ppt slide).

4. Refine the patterns and write them up as documents.
I expect there will be basic patterns and variations.
For each pattern, identify the context and conditions
where it would be useful (and where it would not).

5. Implement the UMM model using one or more
runtime choreography.  This is where I think that 
the Commitments project meets the BPSS and 
BSI and other related projects.

Of course, there should be lots of cross-talk
all the time.  And the Collaboration cross-talk
certainly starts earlier.

6. Get one or more software vendor to do a
runtime version.  The test cases shd be ready.

Make sense?
Suggestions?

Respectfully,
Bob Haugen