H.S.

Thanks for the comments.

Your key problem with the paper seems be the idea of allowing multiple state
machines per object. As you know, I think that allowing multiple state machines
per object is very powerful. The key advantages are:

- It is scalable to complex problems (as there is no state space explosion)
- It allows behaviour to be re-used (by using a given machine in the definition
of multiple objects)
- It allows business rules to be modelled (see
http://www.metamaxim.com/download/documents/Deontic.pdf )
- It allows states to be derived, like derived attributes.

It also works well in practice.

One of my reasons for getting this paper published was to see if others in the
modelling community shared your concerns. At least based on the comments of the
5 peer reviewers of the paper, they do not.

Anyway, I hope the paper was of interest, even if you don't like what it
suggests!

Rgds
Ashley

[Non-text portions of this message have been removed]