This study addresses the challenges introduced by the shift from procedural-oriented to object-oriented paradigms, focusing on the need to maintain consistent design while satisfying market demands, particularly in achieving software stability. While previous research has explored various factors affecting software stability, a gap remains in understanding the correlation between class stability and C&K and evolution-based metrics. To fill this gap, we conducted an empirical investigation using two open-source Java projects, Android and Eclipse, across three versions each. Our approach involved collecting C&K and evolution metrics using two different tools, calculating the stability metric for classes, and analyzing the data with SPSS to determine correlations. The results show a significant negative correlation between the class stability metric (CSM) and nine other metrics, confirming that C&K and evolution metrics are generally negatively correlated with CSM, although three C&K metrics exhibit weak correlations. These findings enhance our understanding of the relationship between various metrics and class stability, contributing to more stable and maintainable software systems.

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

Analyzing Class Stability Through C&K and Evolution Metrics: An Empirical Study

  • Mustafa Ghaleb,
  • Mohamed Alasow,
  • Azzah AlGhamdi,
  • Mosab Hamdan,
  • Sajjad Mahmood

摘要

This study addresses the challenges introduced by the shift from procedural-oriented to object-oriented paradigms, focusing on the need to maintain consistent design while satisfying market demands, particularly in achieving software stability. While previous research has explored various factors affecting software stability, a gap remains in understanding the correlation between class stability and C&K and evolution-based metrics. To fill this gap, we conducted an empirical investigation using two open-source Java projects, Android and Eclipse, across three versions each. Our approach involved collecting C&K and evolution metrics using two different tools, calculating the stability metric for classes, and analyzing the data with SPSS to determine correlations. The results show a significant negative correlation between the class stability metric (CSM) and nine other metrics, confirming that C&K and evolution metrics are generally negatively correlated with CSM, although three C&K metrics exhibit weak correlations. These findings enhance our understanding of the relationship between various metrics and class stability, contributing to more stable and maintainable software systems.