Making mutation testing a more feasible proposition for the industry

Abstract

Software engineering firms find themselves developing systems for customers whose need to compete often leads to situations whereby requirements are vague and/or prone to change. One of the prevalent ways with which the industry deals with this situation is through the adoption of so-called Agile development processes. Such processes enable the evolutionary delivery of software systems in small increments, frequent customer feedback, and, ultimately, software which continuously adapts to changing requirements. In this fluid scenario, developers rely on automated unit tests to gain confidence that any regressions resulting from code changes will be detected. Consequently, trust in the software system can only follow from the quality of the tests. Unfortunately, the industry tends to rely on tools that calculate primitive measures such as statement coverage, a measure which has been shown to provide a false sense of security.