Activity Diagram with multiple end points

Question

My Question may be simple, but I am confused because I have no idea about activity diagrams.

My question- Are multiple end points ever acceptable in an activity diagram?

Solution 1

It's a bit touchy to reference IBM though they are one of the big OMG parents. The "truth" is written in OMG's superstructures.

Actually UML2.5 talks about Final Node in the context of activity diagrams:

Final Nodes

A FinalNode is a ControlNode at which a flow in an Activity stops. A FinalNode shall not have outgoing ActivityEdges. A FinalNode accepts all tokens offered to it on its incoming ActivityEdges.

There are two kinds of FinalNode:

1. A FlowFinalNode is a FinalNode that terminates a flow. All tokens accepted by a FlowFinalNode are destroyed. This has no effect on other flows in the Activity.
2. An ActivityFinalNode is a FinalNode that stops all flows in an Activity ...

A ControlNode is a kind of ActivityNode and those form Activities. The definition except in the context of the FinalNode does not tell how the single ActivityNodes can be related. So from that you can have as many as you like/need.

Solution 2

Yes, there may be cases where multiple endpoints are acceptable.

It is possible for an activity diagram to show multiple final states. Unlike initial state symbols, of which there can be only one on an activity diagram, final state symbols can represent the termination of one of many branches in the logic -- in other words, the activity may terminate in different manners.

https://www.ibm.com/developerworks/rational/library/content/RationalEdge/sep03/f_umlbasics_db.pdf (Page 5)

Other Resources-

http://iswwwup.com/t/08d7e57730dc/uml-do-activity-diagrams-always-end-in-one-endpoint.html

https://softwareengineering.stackexchange.com/questions/215071/do-activity-diagrams-always-end-in-one-endpoint