More on the E-measure of subdomain testing strategies

Abstract

The expected number of failures detected (the E-measure) has been found to be a useful measure of the effectiveness of testing strategies. This paper takes a fresh perspective on the formulation of the E-measure. From this, we deduce new sufficient conditions for subdomain testing to be better than random testing. These conditions are simpler and more easily applicable than many of those previously found. Moreover, we obtain new characterisations of subdomain testing strategies in terms of the E-measure.