Quotient states and probabilistic channels

Abstract

Restrictiveness is interpreted in terms of a quotient set of the states of a machine. Consideration is given to how restrictiveness can still allow certain probabilistic effects to open up a communication channel between high and low users of a computer system. Specifically, assigning probabilities to transitions in order to look for simple probabilistic channels is examined. The theory is then extended to deal with extended transitions. How Shannon's work on information theory can be used to analyze a system that is restrictive but nonetheless has a probabilistic channel is discussed.<>