Code churn, the intensity of software changes, can threaten software quality and has been empirically associated with many quality attributes. However, few studies have explored its relationship with code smells, which are widely recognized as indicators of poor object-oriented design quality. This study presents a preliminary investigation into the relationship between code churn and code smells. We use the number of commits and edited lines as proxies for code churn and the presence of code smells as a measure of software quality. We mined a large software repository comprising 31 open-source projects, 749 developers, and 153,994 commits to examine how development activity intensity relates to the incidence of code smells. Our findings suggest that code churn is related to the presence of code smells. Notably, developers with lower activity levels tend to have a proportionally higher impact on the introduction or persistence of code smells. This preliminary study highlights a correlation between code churn and code smells, providing a foundation for further research. Future work may explore variations such as different types of smells or commit categories to deepen our understanding of this relationship.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Investigating the Relationship Between Churning and Code Smells

  • Kevin Cerqueira Gomes,
  • Elivelton Ramos Cerqueira,
  • Gabriel Moraes,
  • Lidiany Cerqueira,
  • Glauco Carneiro,
  • Rodrigo Spínola,
  • Manoel Mendonça,
  • José Amancio Macedo Santos

摘要

Code churn, the intensity of software changes, can threaten software quality and has been empirically associated with many quality attributes. However, few studies have explored its relationship with code smells, which are widely recognized as indicators of poor object-oriented design quality. This study presents a preliminary investigation into the relationship between code churn and code smells. We use the number of commits and edited lines as proxies for code churn and the presence of code smells as a measure of software quality. We mined a large software repository comprising 31 open-source projects, 749 developers, and 153,994 commits to examine how development activity intensity relates to the incidence of code smells. Our findings suggest that code churn is related to the presence of code smells. Notably, developers with lower activity levels tend to have a proportionally higher impact on the introduction or persistence of code smells. This preliminary study highlights a correlation between code churn and code smells, providing a foundation for further research. Future work may explore variations such as different types of smells or commit categories to deepen our understanding of this relationship.