Secure computer systems use both mandatory and discretionary access controls to restrict the flow of information through legitimate communication channels such as files, shared memory and process signals. Unfortunately, in practice one finds that computer systems are built such that users are not limited to communicating only through the intended communication channels. As a result, a well-founded concern of security-conscious system designers is the potential exploitation of system storage locations and timing facilities to provide unforeseen communication channels to users. These illegitimate channels are known as covert storage and timing channels. Prior to the presentation of this paper twenty years ago the covert channel analysis that took place was mostly ad hoc. Methods for discovering and dealing with these channels were mostly informal, and the formal methods were restricted to a particular specification language. This paper presents a methodology for discovering storage and timing channels that can be used through all phases of the software life cycle to increase confidence that all channels have been identified. In the original paper the methodology was presented and applied to an example system having three different descriptions: English, formal specification, and high-order language implementation. In this paper only the English requirements are considered. However, the paper also presents how the methodology has evolved and the influence it had on other work.