State Diagram

hi ,
Thanks for your answer, I send you more information about my problem.
My system is composed with several subsystem, each component have a state engine, that’s why I need to define ‘gradually’ the global state diagram to avoid the complexity.
I can illustrate my problem with the state of an andrdoid application :
http://developer.android.com/reference/android/app/Activity. html#ActivityLifecycle

  • I use several diagram view as suggested stephane.
    1- “Hide or not internal state”
    I want to define a global view of android application, then I provide three state:
    Now I want to refine the "visible Lifettime state " and I create a new state diagram with three internal state (OnStart, Foreground Lifetime,OnStop), and Of course in the “OnStart” state , I define the state of my android application in the nominal case… now I re-open the first diagram:
    As you see the first simple diagram, is not very simple !!
    2- The second point concerns the available trigger for a entry point , as you see I have define a trigger for the entrypoint “onstart” in the first diagram, when I create the second
    diagram:
    the tools provides all available trigger… but In my case , only one trigger must be available. The tools doesn’t guaranty the consistency between state and internal state.
    thanks for your recommendation.

Hi,
Point 1: As I wrote, I would activate the “Hide Internal States” filter on the state machines diagrams. So that you see only one level.
Point 2: I agree there are a lot of consistency checks missing in Capella state machines. There is some ongoing work in the context of the Clarity consortium, but if you have suggestion of rules, don’t hesitate to propose them. In your case:
If there is only one transition leaving from the initial state on the “VisibleLifeTime” state, I would not repeat the trigger on the transition.
If there is several transitions leaving from this initial state, I understand you need to precise which trigger leads to which state. In that case, yes, it could make sense to be only able to associate one of the trigger associated to the entry points.
I would be interested in discussing this topic with a UML expert