Fault Trees, Prime Implicants, and Noncoherence

The K&H algorithm for finding the p.i.'s of a noncoherent fault tree is too long and complicated for the purpose. The example that was used by K&H to illustrate the algorithm is reworked with known methods to show that alternatives are available that require less work. This correspondence is a series of four letters published as a sequel to a paper by Kumamoto & Henley (K&H) that presented an ``algorithm based on top-down analysis particularly designed for noncoherent fault trees''. The first of these letters is by Locks, claiming that the article is too complicated and that a shorter procedure is available. This is followed by Reply #1 by Ogunbiyi and Reply #2 by K&H. The note is concluded with a rebuttal by Locks to both replies.