| [mdt-uml2.dev] About StateInvariant implementation |
Hi,
I’m currently working on Sequence Diagrams and StateMachine Diagrams.
We want to allow to reference a StateMachine’s State into a Sequence Diagram as a StateInvariant that covers a Lifeline.
According to the UML snippet below:
(describing relations between State, StateInvariant and Constraint)
As we can see we can use the Constraint object as a link between both world (Interaction and StateMachine).
Unfortunately the uml2 implementation is done exactly at the inverse I would have expected.
The State’s stateInvariant property has a containment set to false with a cardinality of 0..1
The StateInvariant’s invariant property has a containment set to true with a cardinality of 1
By default the StateInvariant will own an automatically built Constraint.
Whereas the State can have 1 or zero reference of Constraint.
What can I do to share a State stateInvariant on several Lifeline using several StateInvariant invariant ? IMHO the constraint should be owned by the State itself and reuse on any Lifeline by wrapping it with a StateInvariant.
If the containment properties are reversed, I will be able to attach a Constraint to my State and to reference it from several StateInvariant
Is-it a weakness into the implementation ? otherwise how can I link these items simply (without having to use free text)
(I’m using a Luna M6 derived version)
Thank you in advance for your help
Sébastien BORDES