An Analysis of Automated Technical Debt Measurement

Abstract

Background: Measuring and understanding Technical Debt (TD) is quite complex since there are a number of different definitions and techniques that have been proposed in the last few years and it is not clear which ones should be used in which conditions. The approaches proposed are almost never based on the existing ones and their validation is often performed in a very limited number of projects. For this reasons, practitioners are confused and find difficult to apply such approaches in their projects. Goals: This paper investigates the available techniques for evaluating TD using automated tools aiming at helping practitioners and researcher in understanding the available options and apply them correctly. Method: The study has been performed as a Systematic Literature Review (SLR) applied to 835 studies obtained from the three largest digital libraries and databases. Results: After applying all filtering stages, 38 papers out of 835 have been selected and analyzed in depth. Almost all of them propose novel approaches to measure TD using different criteria and they do not extend or validate existing approaches. Conclusions: The area is not mature and it lacks independent evaluations of the models proposed. Authors focus on proposing new approaches and no consolidation can be identified. Moreover, almost all the approaches proposed are automated only partially and through prototype tools designed just to support the studies analyzed in the paper in which the approach is proposed and rarely maintained. These facts makes difficult the application of such methods by practitioners.