RMI: constructing the server side class (Compute Engine)

Question

I took the Sun RMI tutorial, which was very beneficial. However the understanding now of the RMI interface
from the tutorial has me to come with two new questions. When building your ComputeEngine class (inheriting from Compute which inherits from Remote) do you need to encapsulate the static main function inside ? If not how does
the ComputeEngine process continue without a run function ?
Let say on of the overriding functions in ComputeEngine is:
void doThis(Task t) throws IOExcpetion() {
MyTask mytask = (MyTask) t;
t.execute();
MyTask is a subclass of Task that has been already preinstalled. When the remote client calls the doThis function
how does mytask get the de-serialized data from the client ? Does this function need to do anything to retreive the serialized data ? Also when t.execute is implemented has the Task thread started already ? Do I have to call t.start() ? And how does the execute function survive without being in the run function ?
Thanks Michael

Answer 1

do you need to encapsulate the static main function inside ? If not how does
the ComputeEngine process continue without a run function ?The main() function can be anywhere in your application.
Does this function need to do anything to retreive the serialized data ?No. It just arrives in the method as a parameter.
Also when t.execute is implemented has the Task thread started already?Probably not. It's in whatever state it is in when it's deserialized.
Do I have to call t.start() ?See above.
And how does the execute function survive without being in the run function ?I can't answer the question without knowing anything about the Task class you are using.