Constructing Bayesian-network models of software testing and maintenance uncertainties

The lifetime of many software systems is surprisingly long, often far exceeding initial plans and expectations. During software evolution and maintenance, developers and managers frequently gain or lose confidence in software artifacts, especially when existing uncertainties are relieved or new ones are encountered. Fluctuations in developers' confidences may in turn affect process actions or decisions, e.g. in determining the impact of change, whether regression testing is needed or when to stop testing. We present an approach that allows developers' confidences or “beliefs” regarding software components to be modeled and updated directly. This approach is part of an overall strategy that calls for explicit modeling of software uncertainties using the technique called Bayesian belief networks. Initially, we present several kinds of software uncertainty and how they may be modeled. This is followed by introducing Bayesian belief networks and how they may be used to either confirm, evaluate or predict software uncertainties. We discuss our experiences in constructing Bayesian-network models for an existing software system under development at Beckman Instruments. Once constructed, these models may be used by developers and managers in future software understanding, evolution and maintenance activities. We also list several factors that may affect confidence as identified in conjunction with the Beckman study. Finally, we describe the design and implementation of a Java program that allows software systems and associated beliefs to be modeled explicitly