Investigating factors that affect the human perception on god class detection: an analysis based on a family of four controlled experiments

Abstract

Context:Evaluation of design problems in object oriented systems, which we call code smells, is mostly a human-based task. Several studies have investigated the impact of code smells in practice. Studies focusing on human identification of code smells have shown low agreement among developers. Unfortunately, those studies do not attempt to investigate the reasons behind this phenomenon. Objective: This paper aims to investigate factors affecting human perception of code smells. Specifically, it focuses on factors affecting god class detection, one of the most known code smells. Method: The investigation encompassed a family of four controlled experiments, covering potential factors affecting human detection of code smells. The method is incremental. In other words, each experiment produces insights to the next one. This allows the investigators to control specific factors affecting the agreement on god class detection. The factors addressed in this study are: i) developer experience, ii) developer knowledge, iii) developer training, iv) tool support for design comprehension, and v) software size. Result: Our findings show that tool support for design comprehension is the only factor that does not affect the human perception of god class. The other factors impact this perception in some way. Conclusion: The area still needs more investigation and discussion on what we call the code smell conceptualization problem, to ensure similar criteria and thresholds on human-based code smell detection.

Building empirical knowledge on the relationship between code smells and design patterns: An exploratory study

“…The adoption of both code smell and design pattern metaphors affects common activities of software development, such as estimation of the effort related to design activity, diagnostics in code inspection, and refactoring or maintenance decisions. While practitioners are impacted by the choice of tools and team training, a research effort might be misdirected by lack of knowledge on code smell and design pattern practical effect 27–30 …”

“…While practitioners are impacted by the choice of tools and team training, a research effort might be misdirected by lack of knowledge on code smell and design pattern practical effect. [27][28][29][30] 3.1.2 | RQ2 rationale: How significant is the variation in number of smelly classes linked and not linked to design patterns?…”

Towards a systematic approach to manual annotation of code smells

“…For annotating the Long Method and Large Class code smells we only needed to considered code components in isolation without analyzing other parts of the project. Santos et al (2017) also concluded that the tools supporting design comprehension does not affect the human perception of the Large Class. However, to annotate smells such as Feature Envy and Refused (parent) Bequest, the annotator must also consider related code components.…”

“…We performed error analysis to derive the underlying reasons and found that inconsistency of the annotations and the fact that many instances were not cross-checked (and therefore, these annotations are susceptible to subjectivity) may be the underlying reasons. Santos et al (2017) showed that training the annotators improves Large Class detection. They recommended training the annotators using "golden" examples and conducting discussions between them to align their conceptualization of the analyzed code smells.…”