Abstract: Early studies of random versus partition testing used the probability of detecting at least one failure as a measure of test effectiveness and indicated that partition testing is not ...