MEASURING SOFTWARE DEVELOPER PRODUCTIVITY: REVIEW OF APPROACHES AND THEIR LIMITATIONS

Абстракт

This article examines modern approaches for measuring software developer productivity, focusing on the tools used, the available data, and the applied metrics. The main sources of information in the development process are analyzed, including task management systems and version control platforms, as well as different groups of metrics – classical, DevOps, and experimental.

Special attention is given to the limitations of these approaches. It is shown that a significant part of the real work of developers remains outside the available data and is not reflected in the used metrics. Key problems are discussed, such as “invisible work”, lack of context in the data, dependence on team dynamics, and the influence of metrics on behavior.

Based on the analysis, it is concluded that productivity cannot be adequately measured using single quantitative indicators. Instead, a broader and more integrated approach is needed, which takes into account both the technical and the social aspects of work in software teams.