Membership herding and network stability in the open source community: The Ising perspective

The aim of this paper is twofold: (1) to conceptually understand membership dynamics in the open source software (OSS) community, and (2) to explore how different network characteristics (i.e., network size and connectivity) influence the stability of an OSS network. Through the lens of Ising theory, which is widely accepted in physics, we investigate basic patterns of interaction and present fresh conceptual insight into dynamic and reciprocal relations among OSS community members. We also perform computer simulations based on empirical data collected from two actual OSS communities. Key findings include: (1) membership herding is highly present when external influences (e.g., the availability of other OSS projects) are weak, but decreases significantly when external influences increase, (2) propensity for membership herding is most likely to be seen in a large network with random connectivity, and (3) for large networks, when external influences are weak, random connectivity will result in higher network strength than scale-free connectivity (as external influences increase, however, the reverse phenomenon is observed). In addition, scale-free connectivity appears to be less volatile than random connectivity in response to an increase in the strength of external influences. We conclude with several implications that may be of significance to OSS stakeholders in particular, and to a broader range of online communities in general.

本文有两个目的：（1）从概念上理解开源软件（OSS）社区中的成员动态，（2）探究不同的网络特征（即网络规模和连通性）如何影响OSS网络的稳定性。通过在物理学中被广泛接受的伊辛理论，我们研究了相互作用的基本模式，并对OSS社区成员之间的动态和相互关系提出了新的概念性见解。我们还根据从两个实际OSS社区收集的经验数据进行了计算机模拟。主要发现包括：（1）当外部影响（例如其他OSS项目的可用性）较弱时，成员的从众行为非常明显，但当外部影响增加时，这种行为会显著减少；（2）在具有随机连通性的大型网络中最有可能出现成员从众倾向；（3）对于大型网络，当外部影响较弱时，随机连通性将导致比无标度连通性更高的网络强度（然而，随着外部影响增加，会观察到相反的现象）。此外，在应对外部影响强度增加时，无标度连通性似乎比随机连通性的波动性更小。最后，我们得出了一些对OSS利益相关者尤其重要，对更广泛的在线社区也普遍具有重要意义的启示。