Assessing the fault-detecting ability of testing methods

It is important to be able to assess the fault detection capabilities of proposed software testing techniques. It is not enough for a researcher to introduce new techniques and assure us that they are good. There must be precise ways of comparing criteria and rating them based on their fault detection ability. With this in mind, several relationships between software testing criteria are studied in this paper. For each of these relations R, we investigate whether R(C1, G’z) guarantees that criterion Cl is better at detecting faults than C’2, according to various probabilistic measures.