Every other workflow system I've worked with has had a Queue domain object, which is a persistent named repository for tasks. They are used in both push and pull assignment models, and provide a point of control and extensibility to address specific task-management needs. Actor pools are not the same, because they are per-task-instance. For example, if a group of people pull tasks from a common task pool fed by multiple inputs, there's no way, beyond modifying the individual task-instances and driving process definitions, to, say, replace Huey and Dewey with Mickey and Minnie when Huey and Dewey fly south for the winter. Am I missing something? -Ed Staub View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4052811#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...