Firms’ involvement in Open Source projects: A trade-off between software structural quality and popularity

Open Source (OS) was born as a pragmatic alternative to the ideology of Free Software and it is now increasingly seen by companies as a new approach to developing and making business upon software. Whereas the role of firms is clear for commercial OS projects, it still needs investigation for projects based on communities. This paper analyses the impact of firms’ participation on popularity and internal software design quality for 643 SourceForge.net projects. Results show that firms’ involvement improves the ranking of OS projects, but, on the other hand, puts corporate constraints to OS developing practices, thus leading to lower structural software design quality.  

Understanding knowledge sharing activities in free/open source software projects: An empirical study

Free/Open Source Software (F/OSS) projects are people-oriented and knowledge intensive software development environments. Many researchers focused on mailing lists to study coding activities of software developers. How expert software developers interact with each other and with non-developers in the use of community products have received little attention. This paper discusses the altruistic sharing of knowledge between knowledge providers and knowledge seekers in the Developer and User mailing lists of the Debian project. We analyze the posting and replying activities of the participants by counting the number of email messages they posted to the lists and the number of replies they made to questions others posted. We found out that participants interact and share their knowledge a lot, their positing activity is fairly highly correlated with their replying activity, the characteristics of posting and replying activities are different for different kinds of lists, and the knowledge sharing activity of self-organizing Free/Open Source communities could best be explained in terms of what we called “Fractal Cubic Distribution” rather than the power-law distribution mostly reported in the literature. The paper also proposes what could be researched in knowledge sharing activities in F/OSS projects mailing list and for what purpose. The research findings add to our understanding of knowledge sharing activities in F/OSS projects.  

面向开源生态的软件数据挖掘技术研究综述

全球开源软件生态中孕育的大众化软件生产模式正快速形成一种新型的软件生产力，在软件开发和应用各个环节发挥了巨大作用.大众化软件生产的数据规模日趋庞大、协同范围不断扩展、管理模式高度精简，这些全球化特征使其在软件复用、协同开发、知识管理等环节面临诸多挑战，迫切需要新的理论指导和工具支持.本文首先界定了大众化软件生产活动的分布范围、基本过程和数据形态，然后从软件复用、协同开发、知识管理三个核心环节对开源社区数据挖掘技术的研究工作进行了归类与分析，最后总结了该领域研究工作存在的问题和未来发展趋势.  

基于弱连接理论的GitHub网络的分形特征分析

研究者发现许多真实网络中存在分形特征,并广泛认为网络的异配性导致了分形性。因此,具有同配性的社交网络的分形特征很少被研究。针对开源合作平台中存在许多大型项目的开发者之间未必有实际的合作关系的现象,引入了边的权重来移除弱连接的边。通过重整化群分析,发现GitHub网络在移除弱连接的边之后,其网络结构从小世界变化为分形网络。此外,对网络的Pearson相关系数和邻居相关度进行分析后,发现网络具有很强的同配性,验证了之前对分形起源的理论分析。