RE: Comments on process MetaModel

[Bob Haugen <linkage@interaccess.com>]

Karsten Riemer wrote:
>I agree with most of Cory's comments.

>Specifically - I too believe that the REA model is in some sense a subset of
>a generic model for interaction between organizations, namely that subset that
>produces true economic events with full duality.

Do you also agree with the proposal that the classes that model business
relationships be removed from the metamodel?  (Repeating this
circular argument?)

Needless to say, I disagree.  I don't think you can support 
any contractual business relationships without all of them.
How would you support the auto component procurement
example?

Here are the classes in the metamodel that model economic
relationships between Partners:

Partner
Partner Role
Economic Event
Economic Resource
Economic Resource Type
Contract
Contract Type

Do you propose removing all of them?

It is true that Economic Events are a subset of Business Events,
as the current metamodel indicates.  But generic Business
Events do not have constraining relationships with one another,
nor do they fulfill contracts, so if you get rid of Economic Events,
you get rid of contract fulfillment relationships and actual
economic exchanges.

You could collapse Partner into Party, I suppose, but would
need the relationship between the collapsed class and
Economic Resources (and I think another relationship is
required between Partner and Economic Resource Type,
e.g. a Partner offers Economic Resource Types, e.g.
products.

Nothing else in the list of classes above is a subtype or
subset of anything else in the metamodel, and all
are needed to define contractual business processes.
In that sense, I think they are already minimal. 

What is your design for a generic model of interaction
between organizations that can express the same 
relationships?

And will that design hold up?  REA has been peer-
reviewed and implemented in software for 20 years; 
should we be inventing new models on the fly that 
are not similarly proven?

(The same comment goes to tinkering with the
RosettaNet model of interactions.  I don't see
any problem with a principled merger of eCo 
and RosettaNet, agreed to by proponents of
both models, but designing something new
on the fly - by committee - does not seem 
like a good idea.)

-Bob Haugen