A semiformal correctness proof of a network broadcast algorithm

In past years, a large number of published distributed algorithms have been shown to be incorrect. Unfortunately, designers of distributed algorithms typically use informal correctness proofs, which tend to be unreliable. Formal correctness proofs offer a much higher degree of reliability, but they are not popular among algorithm designers because they are too mathematical and they typically assume synchronous message communication or some other abstract notation, and are therefore not easily applicable to the asynchronous message passing environment, the environment commonly assumed by many algorithm designers. To address this problem, we have developed a semiformal correctness proof method for the asynchronous message passing environment, using ideas from well known formal correctness proof methods. We illustrate part of the proof method by proving the safety property of a simple network broadcast algorithm