.png){kind=icon}
返回
An empirical study on the teams structures in social coding using GitHub projects
.png){kind=icon}
OA
AI
DOI：10.1007/s10664-019-09700-1.png)
.png){kind=icon}
.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Social coding enables collaborative software development in virtual and distributed communities. Social coding platforms (e.g., GitHub) provide the pull request feature that allows developers to clone a project, make code changes, and request the project owners to review and integrate the code changes to the main stream of a project. The pull request feature has been widely adopted by a large number of GitHub projects, as it minimizes the risk of exposing the projects to the open communities. The efficiency of the pull requests review process depends both on technical (e.g., the code quality) and social (e.g., the connection of a contributor to the project maintainer) factors. However, it is still unclear which social factors have the most impact on the efficiency of the review process. To identify the social factors, we study the team structures formed by the developers within the projects that adopt the pull-based development model. We build the pull-based networks, where two developers are linked if one has integrated a pull request submitted by the other. We investigate the 7,850 most popular projects on GitHub that are developed in ten programming languages. We identify the network metrics that have a significant association with the speed of processing the pull requests. Specifically, maintaining a strong core of contributors and denser interactions among the developers is associated with faster response and processing of the pull requests. We further find that more than 90% of the studied projects follow 8 dominant team structures out of 18 possible team structures. In the larger projects, only a set of developers is granted review and integration privileges of the pull requests, reflecting a strict decision making process. The small to medium projects are characterized by a small number of core contributors who maintain repeated interactions, and are able to process the incoming pull requests more efficiently. The evolution of the team structures of projects over time reveals that only a low percentage of the projects witnesses a change towards team structures associated to faster pull requests processing (e.g., stronger centralization).
Keyword:
Pull request
Social coding
Team structure
Github
AI总结
对已上传原文的论文进行重点信息的提取，主要内容包括：简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
3.6
论文数:
2.0K
被引数:
5.3K
机构
引用论文
Machine Learning Investigation For Tri-Magnetized Sutterby Nanofluidic Model with Joule Heating In Agrivoltaics Technology.png)
Nano
NanoIF0
PREAI
分享收藏An Efficient Algorithm for Nonlinear Model Predictive Control of Large-Scale Systems Part I: Description of the Method (Ein effizienter Algorithmus für die nichtlineare prädiktive Regelung großer Systeme Teil I: Methodenbeschreibung)大型系统非线性模型预测控制的有效算法第一部分: 方法的描述 (Ein effizienter algorithms f ü r die nichtlineare pr ä diktive Regelung gro ß er Systeme Teil I: Methodenbeschreibung)auto
autoIF0
PREAI
分享收藏分享收藏
查看更多内容