What Happened to Software Metrics?

The software community of the 1980s was abuzz with seemingly endless approaches to producing higher-quality software. At the forefront was software metrics and its corresponding techniques, tools, and process-improvement schemes. Cyclomatic complexity, Halstead metrics, source lines of code (SLOC), Fagan inspection, defect counting, number of defects prediction, reliability estimation and modeling, and other metric-oriented ideas were oated as solutions to the software quality “quagmire.” To elucidate what went wrong—and right—in software metrics over the past 40 years and to explore future opportunities for new or hybrid metrics, we interviewed a panel of seven experts: Alain Abran (University of Quebec), Victor Basili (University of Maryland), Jim Bieman (Colorado State University), Ram Chillarege (Chillarege Inc.), Taghi Khoshgoftaar (Florida Atlantic University), Edward F. Miller (Software Research Inc.), and Adam Porter (Fraunhofer USA). See the “Roundtable Panelists” sidebar for information about the panel. The panelists’ individual insights are presented below. STATIC AND DYNAMIC SOFTWARE METRICS Computer: If you could only recommend one static software metric and one dynamic software metric, what would they be, and why?