Towards Assessing the Technical Debt of Undesired Software Behaviors in Design Patterns

Abstract

Providing software developers and researchers with useful technical debt analysis tools is an instrumental outcome of software engineering and technical debt research. Such tools aggregate industry best practices to provide users with organized and quantifiable metrics that can help minimize the time it takes to synthesize and make an intelligent future decision regarding a system. Today, most tools rely primarily on structural measurements from static analysis to generate results. However, it is also necessary to consider measurements that capture the behavior of software, as these represent additional complexities within a system that structural measurements are incapable of detecting. Herein, we present our position, that more effort needs to be placed towards understanding software behavior so that technical debt analysis tools can begin supporting them, in order to provide tool users with a more accurate and complete view of their system. In this paper, we describe this problem in the context of design patterns and outline an effective method to talk about behaviors in the future. We create and classify two example behaviors using our method, both of which increase the technical debt in their respective design pattern applications.