This week's book giveaway is in the OCMJEA forum. We're giving away four copies of OCM Java EE 6 Enterprise Architect Exam Guide and have Paul Allen & Joseph Bambara on-line! See this thread for details.
I've read some threads recently about creating Service interfaces that can be used in networked and standalone modes.
My GUI controller will have a reference to a Service object that could be a remote stub or a local object. Does it violate spec to use the same class for both modes? I would like to use the Remote object in standalone mode. I would not use any networking but the object I create would be a Remote (object instanceof Remote). Since my assignment says that none of the network server code may be used in standalone mode I'm worried that this might be against the spec.
The alternative I've been reading about is to create a separate RemoteService interface that extends Service. So Service would not be a Remote but its methods would throw RemoteException. Can anyone help me understand why I can't just use the one Remote object in both modes?
I implemented the alternative, because my approach needed 2 different implementations: the method implementations are different in local and rmi mode.
According to my instructions:
Keep in mind that networking must be entirely bypassed in the non-networked mode.
The ultimate question: what is meant by the word "entirely"? No use of Remote interface or just no network traffic?
As far as I know I don't know of someone using your described approach, so following this approach might be risking automatic failure.
i think Mike's second alternative would look like Roberto explained here. So, following Roberto's approach is valid and not against the specs i think. It has to be valid, because i changed my client implementation from thick to thin and did it in a very similar way to Roberto...