Experiences on Managing Technical Debt with Code Smells and AntiPatterns

TL;DR

This paper discusses practical methods for managing technical debt by detecting code smells and AntiPatterns through static analysis and manual inspection, demonstrating their effectiveness in real-world software projects.

Contribution

It presents a combined approach using static analysis and manual inspection to identify and address technical debt via code smells and AntiPatterns in an industrial setting.

Findings

Effective detection of code smells using static analysis.

Identification of AntiPatterns from code smells.

Successful remediation of technical debt in a real-world project.

Abstract

Technical debt has become a common metaphor for the accumulation of software design and implementation choices that seek fast initial gains but that are under par and counterproductive in the long run. However, as a metaphor, technical debt does not offer actionable advice on how to get rid of it. To get to a practical level in solving problems, more focused mechanisms are needed. Commonly used approaches for this include identifying code smells as quick indications of possible problems in the codebase and detecting the presence of AntiPatterns that refer to overt, recurring problems in design. There are known remedies for both code smells and AntiPatterns. In paper, our goal is to show how to effectively use common tools and the existing body of knowledge on code smells and AntiPatterns to detect technical debt and pay it back. We present two main results: (i) How a combination of static code analysis and manual inspection was used to detect code smells in a codebase leading to the discovery of AntiPatterns; and (ii) How AntiPatterns were used to identify, characterize, and fix problems in the software. The experiences stem from a private company and its long-lasting software product development effort.