NOTES

NOTES

    Might be a good idea to get ahead of stuff and design "zones"
    into client/server, eg. ClientZone, ServerZone. Each client/server
    zone can be separate, eg. a machine can have multiple client zones,
    and each client zone could be running, connected to some state in 
    the background, eg. like wads in previous engines I've worked on.

    But this is overkill initially. Just start with basic client/server
    with an entity manager each. Make sure client doesn't know about
    the initial state of the world, thus the initial block of data
    needs to come down from the server to initialize this.

    This should *probably* be a base manager type that takes the class to be managed
    into consideration, eg. CubeManager : public Manager<CubeEntity>

    Maybe in this case differentiating edicts from entities make sense again.

    EdictManager (array of indices), EntityManager<T> ?