We present an overview of key results on the control of partially-observed discrete-event systems. Both centralized and decentralized control architectures are considered. The control framework adopted is that of the theory of supervisory control of discrete event systems, initiated by Ramadge & Wonham in the 1980's. Four key properties of this discrete-event system theory are: controllability, nonconflicting, observability, and co-observability. We focus on the properties of observability and co-observability. The first part of the talk presents computational complexity and decidability results regarding the synthesis of safe and nonblocking partial-observation supervisors for centralized and decentralized systems. The second part of the talk presents a decentralized control architecture where the control actions of the individual supervisors are combined by ``fusion by union'' (of enabled events) for certain controllable events and ``fusion by intersection'' for the other controllable events. The properties of this architecture are described and contrasted to previous architectures considered in the discrete event systems literature.