Advances in Complex Systems

Large collaborative software engineering projects are interesting examples for evolving complex systems. The complexity of these systems unfolds both in evolving software structures, as well as in the social dynamics and organization of development teams. Due to the adoption of Open Source practices and the increasing use of online support infrastructures, large-scale data sets covering both the social and technical dimension of collaborative software engineering processes are increasingly becoming available. In the analysis of these data, a growing number of studies employ a network perspective, using methods and abstractions from network science to generate insights about software engineering processes. Featuring a collection of inspiring works in this area, with this topical issue, we intend to give an overview of state-of-the-art research. We hope that this collection of articles will stimulate downstream applications of network-based data mining techniques in empirical software engineering.

It is an intrinsic property of real-world software to evolve, which is usually accompanied by the increase of software complexity and deterioration of software quality. So successful software has to be reconditioned from time to time. Though many refactoring approaches have been proposed, only a few of them are performed at the package level. In this paper, we present a novel approach to refactor the package structure of object-oriented (OO) software. It uses weighted bipartite software networks to represent classes, packages, and their dependencies; it proposes a guidance community detection algorithm (GUIDA) to obtain the optimized package structure; and it finally provides a list of classes as refactoring candidates by comparing the optimized package structure with the real package structure. Through a set of experiments we have shown that the proposed approach is able to identify a majority of classes that experts recognize as refactoring candidates, and the benefits of our approach are illustrated in comparison with other two approaches.

Large software projects are among most sophisticated human-made systems consisting of a network of interdependent parts. Past studies of software systems from the perspective of complex networks have already led to notable discoveries with different applications. Nevertheless, our comprehension of the structure of software networks remains to be only partial. Here we investigate correlations or mixing between linked nodes and show that software networks reveal dichotomous node degree mixing similar to that recently observed in biological networks. We further show that software networks also reveal characteristic clustering profiles and mixing. Hence, node mixing in software networks significantly differs from that in, e.g., the Internet or social networks. We explain the observed mixing through the presence of groups of nodes with common linking pattern. More precisely, besides densely linked groups known as communities, software networks also consist of disconnected groups denoted modules, core/periphery structures and other. Moreover, groups coincide with the intrinsic properties of the underlying software projects, which promotes practical applications in software engineering.

In this paper, we apply concepts from Distributed Leadership, a theory suggesting that leadership is shared among members of an organization, to frame models of contribution that we uncover in five relatively successful open source software (OSS) projects hosted on GitHub. In this qualitative, comparative case study, we show how these projects make use of GitHub features such as pull requests (PRs). We find that projects in which member PRs are more frequently merged with the codebase experience more sustained participation. We also find that projects with higher success rates among contributors and higher contributor retention tend to have more distributed (non-centralized) practices for reviewing and processing PRs. The relationships between organizational form and GitHub practices are enabled and made visible as a result of GitHub's novel interface. Our results demonstrate specific dimensions along which these projects differ and explicate a framework that warrants testing in future studies of OSS, particularly GitHub.

In a large software project, the number of classes, and the dependencies between them, generally increase as software evolves. The size and scale of the system often makes it difficult to easily identify the important components in a particular software product. To address this problem, we model software as a network, where the classes are the vertices in the network and the dependencies are the edges, and apply K-core decomposition to identify a core subset of vertices as potentially important classes. We study three open source Java projects over a 10-year period and demonstrate, using different metrics, that the K-core decomposition of the network can help us identify the key classes of the corresponding software. Specifically, we show that the vertices with the highest core number represent the important classes and demonstrate that the core-numbers of classes with similar functionalities evolve at similar trends.

In open source software (OSS) projects, participants initially communicate with others and then may become developers if they are deemed worthy by the community. Recent studies indicate that the abundance of established social links of a participant is the strongest predictor to his/her promotion. Having reliable rankings of the candidates is key to recruiting and maintaining a successful operation of an OSS project. This paper adopts degree-based, PageRank, and Hits ranking algorithms to rank developer candidates in OSS projects based on their social links. We construct several types of social networks based on the communications between the participants in Apache OSS projects, then train and test the ranking algorithms in these networks. We find that, for all the ranking algorithms under study, the rankings of emergent developers in temporal networks are higher than those in cumulative ones, indicating that the more recent communications of a developer in a project are more important to predict his/her first commit in the project. By comparison, the simple degree-based and the PageRank ranking algorithms in temporal undirected weighted networks behave better than the others in identifying emergent developers based on four performance indicators, and are thus recommended to be applied in the future.

We develop a model of innovation communities which allows us to address in a systematic way the influence of users and developers as well as communication between and within these groups. Based on this model, we derive a formal approach to quantify communication flows, community activity and community turnover. These measures are calculated using the data of 100 open source software projects. Our empirical analysis shows that: (i) Users play indeed a predominant role in communication, which points towards the vivid role of an active user community; (ii) communication is highly concentrated, which points towards the importance of active individuals and (iii) community turnover exhibits only little correlation with community segregation, which may allow to benefit from high turnover rates while keeping negative effects small. We argue that insight from this extensive analysis not only complements existing case studies, it also provides a reference frame to put these singular results into perspective when aiming at generalizations.

Established open source software (OSS) projects can grow in size if new developers join, but also the number of OSS projects can grow if developers choose to found new projects. We discuss to what extent an established model for firm growth can be applied to the dynamics of OSS projects. Our analysis is based on a large-scale data set from SourceForge (SF) consisting of monthly data for 10 years, for up to 360,000 OSS projects and up to 340,000 developers. Over this time period, we find an exponential growth both in the number of projects and developers, with a remarkable increase of single-developer projects after 2009. We analyze the monthly entry and exit rates for both projects and developers, the growth rate of established projects and the monthly project size distribution. To derive a prediction for the latter, we use modeling assumptions of how newly entering developers choose to either found a new project or to join existing ones. Our model applies only to collaborative projects that are deemed to grow in size by attracting new developers. We verify, by a thorough statistical analysis, that the Yule–Simon distribution is a valid candidate for the size distribution of collaborative projects except for certain time periods where the modeling assumptions no longer hold. We detect and empirically test the reason for this limitation, i.e., the fact that an increasing number of established developers found additional new projects after 2009.