@Ronald, Just so I can think about the difference in approaches, what are your thoughts on why it would be more difficult to do certain exit strategies with my suggested dynamic task-instance approach? I guess the join would make sure everyone completed before leaving the fork-task-join block, but doesn't signal="last" on the task-node accomplish the same thing? Additionally, doesn't using the signal= on the task node give more flexibility than a join would, which would fix the user to a pattern of waiting for everyone to complete before moving forward? The last point of my implementation is that the graph would not need to change with the dynamic task-instances since it would be an action attached to the task node. But then again, I haven't tried to implement this myself. So, I am wondering if there is a technical hurdle here, or what your thoughts are for why it would be more difficult. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4216077#... Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&a...