For UML cracks: StateMachine MetaModel - Why does Transition **aggregate** Behavior?

[willybrandt]

Hi,

I have a (as I think) pretty sophesticated question about the meta
model of UML state machines as described in
UML 2.1.1 superstructure, Figure 15.2 (see http://www.omg.org/spec/UML/2.1.1/Superstructure/PDF/
).
The class diagram describes the relations between
* Transition and Behavior
* Transition and Trigger
* Transition and Constraint and between
* State and Behavior
as aggregation. Why? Obviously this prevents several transitions from
referencing the same behavior (or trigger, constraint). Is there any
reason why this would be problematic?

I'm currently developing a tool which handles UML state machines and
I'd rather implement the above relations as references, but I don't
want to get in trouble, only because I can't see the reasons for
modeling it as aggregations. Hopefully someone can give me a leg up
here...

So long,
Alfred