Abstract: This paper compares partition testing and random testing on the assumption that program failure rates are not known with certainty before testing and are, therefore, modeled by random ...