Catalog Of Smells Research Articles

A large scale study on how developers discuss code smells and anti-pattern in Stack Exchange sites

Abstract Context: In this paper, we investigate how developers discuss code smells and anti-patterns across three technical Stack Exchange sites. Understanding developers perceptions of these issues is important to inform and align future research efforts and direct tools vendors to design tailored tools that best suit developers. Method: we mined three Stack Exchange sites and used quantitative and qualitative methods to analyse more than 4000 posts that discuss code smells and anti-patterns.Results: results showed that developers often asked their peers to smell their code, thus utilising those sites as an informal, crowd-based code smell/anti-pattern detector. The majority of questions (556) asked were focused on smells like Duplicated Code, Spaghetti Code, God and Data Classes. In terms of languages, most of discussions centred around popular languages such as C # (772 posts), JavaScript (720) and Java (699), however greater support is available for Java compared to other languages (especially modern languages such as Swift and Kotlin). We also found that developers often discuss the downsides of implementing specific design patterns and ‘flag’ them as potential anti-patterns to be avoided. Some well-defined smells and anti-patterns are discussed as potentially being acceptable practice in certain scenarios. In general, developers actively seek to consider trade-offs to decide whether to use a design pattern, an anti-pattern or not.Conclusion: our results suggest that there is a need for: 1) more context and domain sensitive evaluations of code smells and anti-patterns, 2) better guidelines for making trade-offs when applying design patterns or eliminating smells/anti-patterns in industry, and 3) a unified, constantly updated, catalog of smells and anti-patterns. We conjecture that the crowd-based detection approach considers contextual factors and thus tend to be more trusted by developers than automated detection tools.