I think that we could simplify the current StateEvent/PortletContext handling in the context of state management of portlets. The current design, while based on subclasses, does not lend itself to polymorphism. For example, in FederatedInstanceContext: if (event instanceof PortletClonedEvent) | { | PortletClonedEvent pce = (PortletClonedEvent)event; | PortletContext clonedContext = pce.getClonedContext(); | ctx.onStateEvent(new PortletClonedEvent(reference(clonedContext))); | } | else if (event instanceof PortletModifiedEvent) | { | PortletModifiedEvent pce = (PortletModifiedEvent)event; | PortletContext modifiedContext = pce.getModifiedContext(); | ctx.onStateEvent(new PortletModifiedEvent(reference(modifiedContext))); | } I don't see the benefit of having such a split between different kinds of events. The above code could be simplified if StateEvent was taking on more responsibilities: PortletContext portletContext = event.getPortletContext(); | ctx.onStateEvent(new StateEvent(reference(portletContext)), event.getType()); In this new design, StateEvent would not be subclassed but contain a type field for these cases where the exact nature of the event needs to be known. This would only work with a rework of PortletContext. I think that PortletContext should add a getState method. It's perfectly fine for a PortletContext to have an optional state (this is how WSRP works after all). We could keep the subclass to optimize memory usage and avoid to carry a byte[] on all PortletContexts even these which don't have a state. getState would return null if there is no state, StatefulPortletContext would override getState to return the state. I think that these modifications would make dealing with state events easier and simplify our code. What do people think? View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3992106#... Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...