On the generality of power diagrams