Cyclomatic Complexity Metrics Revisited: an Empirical Study of Software Development and Maintenance Comments Were Received from Bill Curtis on an Earlier Draft

While the need for software metrics to aid in the assessment of software complexity for both development and maintenance has been widely argued, little agreement has been reached on the appropriateness and value of any single metric. McCabe's cyclomatic complexity metric, a measure of the maximum number of linearly independent circuits in a program control graph has been widely used in research. However, despite the widespread interest and popularity of this metric, it has not been without criticism, both analytical (the Myers and Hansen variants) and empirical (the high correlation of cyclomatic complexity with size measures). The current research tested both types of critiques on a newly collected dataset of real world software development and maintenance projects. The analytical research questions were tested on a set of 834 software modules from a number of existing real-time systems. Neither the Myers nor Hansen variants were found to be significantly different from the original value as computed by McCabe. Therefore, these particular theoretical criticism seem to have little or no practical impact, as represented by the data collected in this study. In regard to the empirical research questions, previous concerns were validated on this new dataset. However, the current research proposes a simple transformation of the metric whereby the cyclomatic complexity is divided by the size of system in source statements, thereby determining a "complexity density" ratio. This complexity density ratio is demonstrated to be a useful predictor of software maintenance productivity on a small pilot sample of actual maintenance projects. CR