An Hypergraph Based Approach for Unlocking Power Relations in Rhizomatic Open Source Software Communities

This study, first, points the rhizomatic nature of open source software (OSS) communities. Then, it proposes an hypergraph based research methodology both to capture and to analyze dynamics of OSS communities. An exemplary hypergraph theoretical analysis of power antecedence within rhizomatic network of an OSS community is illustrated. I. RHIZOMATIC STRUCTURE The concept of Rhizome is first presented as a model relevant to human society by Deleuze and Guattari [1]. Rhizome acts as a web-like network structure of nonhierarchically connected independent nodes, borrowing its name from the structures of plants such as bamboo and other grasses. This Rhizome metaphor is widely used in social network studies mostly to demonstrate non-authoritarian, diverse, and self-sufficient relations [2]. The Internet also has often been described as a form of Rhizome: A space composed of flows that transcend boundaries and forge new connections between places and events with no central facility that organizes the communication [3]. Although some studies [4] on OSS community implicitly attributes rhizomatic characteristics to its organizational and social aspects, a more explicit reference to its rhizomatic nature, similar to these studies [2,3] on social networking and internetworking, is necessary. OSS systems and communities exhibit very nature of Rhizome like structural characteristics. Their decentralized network structures are consisting of loose connections of individuals and/or other OSS systems and communities [5], all of which, akin to rhizomic nodes, are functionally selfsufficient but not isolated, cooperating but not controlled. Self-sufficiency removes the individual rhizome node from dependence on the standard set of outside suppliers. Loose network connections, such as those in rhizome structures, does not eliminate partnership-based exchange that exists as a voluntary activity. On the other hand, a broader interaction between networks of contributors occurs that allows spontaneous developments, which are regenerative and spreading with a potential of continuous transition. II. HYPERGRAPH METHOD The Rhizome like structure of OSS systems and communities form complex but non-hierarchical networks. Hypergraphs are promising instruments to explore and analyze this complex network that interrelates OSS systems, communities, projects; individual contributers, users, advocates; and all other set of activities taking place within the nexus. In general, hypergraphs are adopted as a suitable model for social networks with non-dyadic relationships [6], where occasionally more than two actors take part within the same event. See Fig.1. Fig. 1. A sample hypergraph. Note that vertices, consisting of actors, are taken as individual Rhizome nodes and nets are events or affiliations. III. UNLOCKING POWER Data collected from mail listings, homepage and documentation of OSCAR clustering project an adjacency matrix representation of hypergraph is constructed. In order to analyze power relations 'clique overlap centrality', 'closeness', and 'betweenness' of each actor and subgroups derived through clustering performed on actors. Findings are supported through triangularization by discursive content analysis of posted messages and semi-structured interviews collected from secondary sources. The findings point out that nature of power is both normative and expert based. As such, it discovers a jointly developed power within subgroups where a co-active power, which is not coercive, is epitomized. There are evidences of memetic control through standards, style, and behavior, which are constructed collectively.