Information Theory-Based Measurement of Software

Abstractions of software may take the form of complex graphs. This chapter presents methods for measurement of hypergraph abstractions of software, using information theory, rather than counting. Even though the context of this chapter is software engineering, the measurement methods are applicable to any hypergraph. The software-metrics literature often assumes that complex software has more bugs than simple software. Thus, if one measures a software system, one can identify which modules are more likely to have bugs. This chapter presents measures of size, complexity, and coupling in terms of the amount of information, building on formal definitions of these software-metric families proposed by Briand, Morasca, and Basili. Kolmogorov complexity and its information theory-based approximations are the foundation for our measure of size. Excess entropy is the foundation for our measures of complexity and coupling. These concepts make the metrics sensitive to the configuration of hypergraph connections, and thus, may be preferred to corresponding counting metrics.