Technical Debt as an External Software Attribute

Background: Technical debt is currently receiving increasing attention from practitioners and researchers. Several metaphors, concepts, and indications concerning technical debt have been introduced, but no agreement exists about a solid definition of technical debt. Objective: We aim at providing a solid basis to the definition of technical debt and the way it should be quantified. Method: We view technical debt as a software quality attribute and therefore we use Measurement Theory, the general reference framework for the quantification of attributes, to define technical debt and its characteristics in a rigorous way. Results: We show that technical debt should be defined as an external software quality attribute. Therefore, it should be quantified via statistical and machine-learning models whose independent variables are internal software quality attributes. Different models may exist, depending on the specific needs and goals of the software product and development environment. Also, technical debt is a multifaceted concept, so different kinds of technical debt exist, related to different quality attributes, such as performance, usability, and maintainability. These different kinds of technical debt should be evaluated individually, so one can better focus on the specific quality issues that need to be addressed. Conclusions: We show that, to provide it with a rigorous basis, technical debt should be considered and measured as an external software attribute. Researchers and practitioners should build models for technical debt and use them to (1) assess the extent of the technical debt and (2) investigate and assess different ways of modifying software to repay technical debt.