I am constructing a UML 2.0 activity diagram. The activity is to evaluate data from a usability study of a prototype. It can optionally end with just getting the user satisfaction results. But if the company wants to choose which features to implement in the final version, which ones to change and which ones to leave out based on the satisfaction, they also need to do this evaluation. If they choose the additional evaluation, they also need some decision criteria.
As I understand the UML 2.0 formalism, the object node :classification criteria" in the diagram version above must exist. It has to be there, to be available for token generation, etc.
But in my activity, if the company doesn't want to do the "classify features" branch and chooses to end the activity at the flow final node, it doesn't have to prepare any classification criteria before it starts the activity. On the other hand, the criteria are external to the activity, so I can't really put them after an if split, they have to be there at the border.
How do I model this situation correctly?