Christopher Spottiswoode comments from a Metaset/MACK point of view
(on Texas Instruments Business Object Architecture Team (Ed. Tom Digre), Business Object Facility):

(Please first read my Introduction to these comments on your papers.)

The major problem I have here is that you wish (in the Foreword on page 5) to facilitate Identifying, structuring, and aligning our specifications consistent with the OMG roadmap. It must be clear by now that I regard that present roadmap as really leading out onto a minor limb of the evolutionary tree!

However, I agree fully with this:

The primary objectives of this specification are to facilitate:

• Simplicity in the development, deployment, change and use of business objects for application users and developers.
• Interoperability of independently developed business objects.
• Implementation of business object configurations as "plug and play business components" of the information system.
• A direct correspondence between the business object model, in understandable business terms, and the business components of the information system.
• Isolation of infrastructure and business objects from tool or presentation technologies.
• Isolation of technology from business logic.
• Transactional integrity across distributed business objects.
• Direct coupling between a Business Object and its defining meta business object.
• Consistency of business object specification across the development lifecycle.
• Flexible and dynamic business object models.
• Enforced business semantics.

If I were to adopt those objectives as MACK’s (and I could certainly state at least those!), then I would only qualify the two starting with the word "Isolation". I would try to make it clear that such isolation merely implies the ability - where applicable - to present and manipulate the one without the other.

I would steer clear of any implication that MACK has any n-tier architectural dogma. As I have explained in [Ralph, Jeff, Wolfgang], MACK implies, needs and fully implements a full integration of the internal model, while yet being able to expose appropriate perspectives to the various kinds of user.

Anything short of such integration tends to oversimplification through loss of generality and easy extensibility, and greatly impedes many opportunities for rationalization, optimization, functionality and - in short - power with simplicity.

MACK has learnt to benefit from the demands for coherence with consistency, and is greatly helped - not hindered - by the increasing interconnectedness of our abstract models.

At the same time such abstraction is safe, as it is stabilized both by its "RE" empirical facets and by its democratic orientation via a boosted demand-driven market with reinforced assistance by supply.

Considering your acceptance and my rejection of the OMG roadmap, it would not be relevant for me to go into the detail of your commendably thorough and well-presented report.

I would prefer to end with this quote (p.5):

Commercial products based on the OMG-adopted Business Object Facilty specification will not be available until at least 1998.

And add the comment that Metaset could well be on the market before the end of 1997 [Andrew], will make MACK’s philosophy, design, structure and all kinds of use very clear to all concerned, and at the latest by then will replace the entire OMA as it is presently conceived.

Maybe you can help it happen sooner? See the "Strategy" section in my introduction to these comments.