|
Re: ActionExecutionSpecification vs. BehaviorExecutionSpecification [message #1005290 is a reply to message #1004604] |
Thu, 24 January 2013 19:36 |
Eclipse User |
|
|
|
Andrea,
you can completely neglect them for modeling operation calls (although I
know the drawing facilities of Papyrus automatically create an
ActionExecutionSpecification when you create a synchronous message; even
worse the created model by Papyrus is invalid, since, due to whatever
reasons, ActionExecutionSpecification requires an Action, in contrast to
BehaviorExecutionSpecification that has an optional association with
Behavior).
Both are intended to invoke additional things in addition to the operation
that has already been invoked by the message itself. So, they are mostly
used for graphical convenience, i.e., to visualize the execution dutation of
an Operation, but this is semantically not correct. When I generate
Interactions, I completely avoid those ExecutionSpecifications, since they
do not add any value for my purpose.
It can be seen similar to StateMachine, where a transition may refer to any
additional behavior in addition to the behavior that was invoked by the
firing event.
Regards,
Marc-Florian
"Andrea Sindico" schrieb im Newsbeitrag
news:kdos86$3h8$1@xxxxxxxxe.org...
What is the actual difference between ActionExecutionSpecification and
BehaviorExecutionSpecification? I mean, I know they are both concrete
realization of the ExecutionSpecification meta-class and that the former
relate to actions while the latter to behaviors. However it is not clear to
me what to choose in a sequence diagram when referring to the execution
specification related to an Operation call.
It seems I am not the one having such misunderstanding
http://stackoverflow.com/questions/14202361/uml-sequence-diagram-in-omg-specificationsuperstructure
|
|
|
Powered by
FUDForum. Page generated in 0.02836 seconds