Author Topic: Acitvity + instance classifier  (Read 463 times)

Torsten Binias

  • EA User
  • **
  • Posts: 31
  • Karma: +0/-0
    • View Profile
Acitvity + instance classifier
« on: December 01, 2008, 09:48:30 pm »
Hello,

EA (7.1.832) offers to set the instance classifer of an activity.
But what is the instance classifier of an activity?

Best regards,
Torsten

Frank Horn

  • EA User
  • **
  • Posts: 535
  • Karma: +0/-0
    • View Profile
Re: Acitvity + instance classifier
« Reply #1 on: December 02, 2008, 07:16:35 pm »
Well, the dialog offers classes and interfaces, and when you choose one, it will show in the project browser ("Activity1 :Class1") but not in the diagrams.

But I've no idea what it's supposed to mean, an activity being an instance of a class.

When I export this to XMI 2.1 I can see that the packagedElement node for the activity has no reference to the so called classifier.  The relation is completely absent below the uml:Model node.

There's only a classifier attribute on the xmi:Extension/elements/element node for the activity, so maybe it's just nonsense.

Or some UML 1 legacy thing.

Paolo F Cantoni

  • EA Guru
  • *****
  • Posts: 6041
  • Karma: +73/-83
  • Inconsistently correct systems DON'T EXIST!
    • View Profile
Re: Activity + instance classifier
« Reply #2 on: December 03, 2008, 10:10:35 am »
Quote
Hello,

EA (7.1.832) offers to set the instance classifier of an activity.
But what is the instance classifier of an activity?

Best regards,
Torsten
Hi Torsten,

If my ageing memory serves me well, the last time I played with this (over 12 months ago), the classifier of an activity turned out to be the composite holonym activity in a composite structure diagram for an activity.

Remember an activity diagram IS an activity (although EA's support for it is somewhat vexed).  So, if you decompose an activity, you need to store the reference to the composite.  Since it's a composition, the meronym can only be in one so the notion of an "instance" is reasonable.

HTH,
Paolo
Inconsistently correct systems DON'T EXIST!
... Therefore, aim for consistency; in the expectation of achieving correctness....
-Semantica-
Helsinki Principle Rules!