Modelling Software Fault debugging Complexity under Imperfect Debugging Environment

The fault debugging progress is influenced by various factors all of which may not be deterministic in nature such as the debugging effort, debugging efficiency and debuggers skill, and debugging methods and strategies. In order to address these realistic factors that influencing the debugging process we propose an integrated nonhomogeneous Poisson process (NHPP) based software reliability model. The integrated modelling approach incorporates the effect of imperfect fault debugging environment, fault debugging complexity and learning debuggers’ phenomenon. The debugging phase is assumed to be composed of three processes namely, fault detection, fault isolation and fault removal. The software faults are categorized into three types namely, simple, hard and complex according to their debugging complexity. As the debugging progresses, the fault removal rate changes to capture learning process of the debuggers. In order to relax the ideal debugging environment, two types of imperfect debugging phenomena are incorporated. Incorporating the imperfect fault debugging phenomena in software reliability modelling is very important to the reliability measurement as it is related to the efficiency of the debugging team. Accordingly, the total debugging process is the superposition of the three debugging activities processes. Such modelling approach can capture the variability in the software reliability growth curve due to debugging complexity of the faults depending on the debugging environment which enables the management to plan and control their debugging activities to tackle each type of fault. Actual test datasets cited from real software development projects have been used to demonstrate the proposed model.