An Empirical Investigation into Industrial Use of Software Metrics Programs

摘要

Practitioners adopt software metrics programs to support their software development from the perspective of either overall quality, performance, or both. Current literature details and justifies the role of a metrics program in a software organization's software development, but empirical evidence to demonstrate its actual use and concomitant benefits remains scarce. In the context of an EU H2020 Project, we conducted a multiple case study to investigate how two software-intensive Agile companies utilized a metrics program in their software development. We invited practitioners from the two case companies to report on the actual use of the metrics program, the underlying rationale, and any benefits they may have witnessed. We also collected and analyzed metrics data from multiple use cases to explain the reported use of the metrics. The analysis revealed improvements like better code review practices and formalization of quality requirements management, either as a direct consequence or as a byproduct of the use of the metrics. The contrasting contexts like company size, project characteristics, and general perspective towards metrics programs could explain why one company viewed the metrics as a trigger for their reported improvements, while the other company saw metrics as the main driver for their improvements. Empirical evidence from our study should help practitioners adopt a more favorable view towards metrics programs, who were otherwise reluctant due to lack of evidence of their utility and benefits in industrial context.