.png){kind=icon}
.png){kind=icon}
Return
Co-change patterns: A large scale empirical study
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jss.2019.03.014.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Co-Change Clustering is a modularity assessment technique that reveals how often changes are localized in modules and whether a change propagation represents design problems. This technique is centered on co-change clusters, which are highly inter-related source code files considering co-change relations. In this paper, we conduct a series of empirical analysis in a large corpus of 133 popular software projects on GitHub. We describe six co-change patterns by projecting them over the directory structure. We mine 1802 co-change clusters and 1719 co-change clusters (95%) are covered by the six co-change patterns. In this study, we aim to answer two central questions: (i) Are co-change patterns detected in different programming languages? (ii) How do different co-change patterns relate to rippling, activity density, ownership, and team diversity on clusters? We conclude that Encapsulated and Well-Confined clusters (Wrapped) implement well-defined and confined concerns. Octopus clusters are proportionally numerous regarding to other patterns. They relate significantly with ripple effect, activity, ownership, and diversity in development teams. Although Crosscutting are scattered over directories, they implement well-defined concerns. Despite they present higher activity compared to Wrapped clusters, it is not necessarily easy to get rid of them, suggesting that support tools may play a crucial role. (C) 2019 Elsevier Inc. All rights reserved.
Keywords:
Modularity
Co-change clusters
Co-change patterns
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
4.1
Papers:
5.4K
Citations:
8.4K