Towards proactive management of technical debt by software metrics

Large software development organizations put enormous amount of effort not only for responding to continuous requests of customers but also for reengineering and refactoring activities to keep their product maintainable. Often rapid and immature feature deliveries over long period of time gradually decrease the product quality, and therefore the refactoring activities become costly and effort-intensive. This situation is described by the concept of “technical debt”, which represents the accumulated rework that organization has to do in order to prevent the slowdown of the development. In this paper we report results of a case study at Ericsson on using software metrics for moving towards proactive management of technical debt. Our observations show that there are four distinguishable maturity phases of quality management over the eight years of development time of two large products: Start-n-stop, Reactive, Systematic, and Proactive quality management. Three sophisticated metrics are applied to help the organizations to move towards Proactive management of technical debt. These metrics are used on a systematic basis to provide information on the areas of the product that have tendency of accumulating technical debt. Software engineers use this information for making decisions on whether or not the pinpointed areas should be refactored.