One interesting thing to consider here is the fact that you probably won't have a transition to the end state from every node in the process design. If you want to send the token to the end state at once, you have to take into consideration the whole path needed to get there, or you can "fly" there changing the state. Nevertheless, I think that the "flying" option must be adopted carefully cause strange things can start happening. For example (at least in jbpm 3.1.2) there is no restriction for a subtoken to escape from a fork-join segment to the parent process. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4054817#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...