WIT Press

Identification Of Experimental Parameters For Automatic Software Evaluation And Testing


Free (open access)

Paper DOI





1,430 kb


A. Montanari & M. Pighin


Identification of experimental parameters for automatic software evaluation and testing A. Montanari & M. Pighin Dipartimento di Matematica e Informatica, Universitd di Udine, Via Zanon, 6, 33100 Udine, Italy ABSTRACT The paper presents an approach to software development methodology re- vision and testing based on the evaluation of a number of experimental parameters characterizing the structure of the resulting software product. The values of such parameters are statistical in nature and can only be determined on the basis of concrete practice. In our work, we focus on two main values of each parameter: mean value M and standard deviation (j. However, these values are not sufficient by themselves to establish the criticality of a parameter and must be correlated to the statistics of sig- nalled errors of programs to identify the actual weight of each parameter. In the paper, we describe the basic steps of a procedure that determines p