Probabilistic Modelling of Software Failure Characteristics

Abstract This paper describes an empirical study of the failure characteristics of software defects detected in the programs developed in the Project on Diverse Software (PODS). The results of the empirical study cast doubts on the general validity of the following assumptions:- - similar failure rates for a population of defects, - a constant software failure rate. The observed deviations from these assumptions are explained in terms of a state machine model of software behaviour in conjunction with an assumption that software defects occur as ‘blobs’ in the program state space.