Software complexity and testing effectiveness: An empirical study

This paper presents a simple method for evaluating the effectiveness of testing strategies by using the relation between software reliability and software complexity. A wide range of metrics from the literature are being tested for their suitability in measuring changes in software complexity of software code at different levels of system partitioning. Reliability data from software tests and operational data of professional systems from the field are examined for their dependency on software complexity. Does more complex software require more rigorous testing? Is current testing adequate for taking the complexity factor out of the equation or does it still remain relevant for software reliability in the field? The results of the statistical analysis show that software complexity is relevant during testing, affecting the amount of discovered errors. It does not affect software field reliability once the code has been debugged. In addition it is suggested that software complexity metrics when used to measure software changes, should be applied to low level software partitioning before integration.