Software reliability measurements in N-Version software execution environment

The author quantitatively examines the effectiveness of the N-Version programming approach. He looks into the details of an academia/industry joint project employing six programming languages, and studies the properties of the resulting program versions. He explains how exhaustive testing was applied to the project, and measure the error probability in different N-Version Software execution configurations. He also applies mutation testing techniques to measure the safety coverage factor for the N-Version software system. Results from this investigation reveals the potential of N-Version software in improving software reliability. Another observation of this research is that the per fault error rate does not remain constant in this computation-intensive project. The error rates associated with each program fault differ from each other dramatically, and they tend to decrease as testing progresses