A note on two-stage software testing by two teams

Abstract

This paper reports our experience of software reliability evaluation in the actual software development which employs somewhat uncommon testing procedure. The procedure can be called as a two-stage testing by two teams. First we discuss the problems underlying the software testing. We explain the concept of the procedure of the two-stage testing by two teams, and then, a stochastic model which traces the procedure is derived. The model analyzes a data set which were obtained from the real testing activity. By using the result of data analysis, we show some numerical illustrations for the software reliability assessment and the evaluation of the skill level of the test teams. © 2011 Springer-Verlag.