« Update: State and Polymorphism Subsystems | Main | Updating API for Postgresql 9.1 »

Polymorphic creation events?

It seems the model permits a Creation Event to be polymorphic, which has been rightly pointed out as nonsense.  Consider the bank account example.  You might want to send a close signal to an instance of Account without knowing or caring whether it is checking or savings, even though the resultant behavior might be handled in only one subclass.  But you wouldn't ever try to create an instance of Account without knowing what type it should be.  (Otherwise you would create only a portion of the object, which is illegal in miUML).

You can see on the current State and Polymorphism subsystem class diagrams that a Creation Event is an Effective Event which may be an Inherited Effective Event, which would be wrong in the case of a Creation Event.  Fix pending.

Reader Comments

There are no comments for this journal entry. To create a new comment, use the form below.
Member Account Required
Welcome to miUML. We want your input, but to ward off evil spam, we request that you create a user account first.