Experimental assessment of random testing for object-oriented software

Progress in testing requires that we evaluate the effectiveness of testing strategies on the basis of hard experimental evidence, not just intuition or a priori arguments. Random testing, the use of randomly generated test data, is an example of a strategy that the literature often deprecates because of such preconceptions. This view is worth revisiting since random testing otherwise offers several attractive properties: simplicity of implementation, speed of execution, absence of human bias. We performed an intensive experimental analysis of the efficiency of random testing on an existing industrial-grade code base. The use of a large-scale cluster of computers, for a total of 1500 hours of CPU time, allowed a fine-grain analysis of the individual effect of the various parameters involved in the random testing strategy, such as the choice of seed for a random number generator. The results provide insights into the effectiveness of random testing and a number of lessons for testing researchers and practitioners.