Mozilla & the Rebel Alliance

Network View with Contributor Segments

With more than 2 million contributions from 22,821 non-employee contributors between 2017-2019, Mozilla has a Rebel Alliance that’s vibrant, active, and impactful.

This network view shows contributors between 2017-2019 as defined by their segment, or their total lifetime amount of contributions (casual, regular, core). The node (circles) represent contributors, and the edges (lines) represent contributions from contributors to different Mozilla projects. Contributors are attributed to the project to which they most contribute. The graph shows where contribution across projects occurs, which can be filtered by contributor segment to highlight their different pathways (see pull down menu/filter). This analysis includes only non-employee contributors who made more than 5 contributions during this time period -- in other words, no ‘visitors’.

In the center of the network, we find an extremely productive group of core and regular contributors who are active on multiple projects. Non-employee core contributors contribute to an average of 4.3 projects, and non-employee regular participants contribute to an average of 2.5 projects.

Some of the big projects (larger labels) are placed at the periphery because they have a large community of non-employee contributors who only participate in that specific project (e.g. Firefox, Thunderbird, MDN). Projects that are placed closer together share many contributors.

In general, the more connections in a network, the stronger it is. By this view, contributor communities in Mozilla’s Rebel Alliance are generally well connected and robust.

The full report dives more deeply into the ‘health’ of ten Mozilla sponsored open source projects. This graph shows a short selection of insights related to these contribution areas - just select a project. Note that where we highlight attraction and retention numbers, they are specific to that particular project and do not indicate that those who no longer participate in that project aren’t continuing to participate in other contribution areas.

Data sources and definitions

This network was created by gathering and analyzing activity data from the following platforms and communitites: Git and GitHub, Bugzilla, Kitsune (SUMO), Pontoon (Localization), Kuma (Mozilla Developer Network) and Add-ons.

For this upcoming report, we developed platform specific definitions based on lifetime contribution activity (see table below). This method enables us to create more granular and meaningful segments that take into account the different natures of each platform and contribution types (code, translations, questions, articles, etc.).

Contributors were defined as casual, regular or core if they made a certain threshold of contributions over their lifetime of contribution to that particular platform. The segment definitions were developed based on all contributions to the platform, also including contributions from employees.

The Rebel Alliance colored by project activity

This network is based on the same data sources as the interactive graph above. The nodes represent contributors and the edges (lines) represent contributions from contributors to different Mozilla projects. Each contributor is coloured by the project, which most of their contributions go to. In this way, we can talk about project communities.

In the centre of the network, we see the same extremely active group who contribute on several different projects, however as the purple colour represent – a lot of these people have most of their contributions to Firefox.

We can also identify some overlapping groups – such as Firefox and Gecko. They both have their own community of people not contributing to any other projects, but they also have contributors scattered around the network contributing to various different project. As such, both projects are fairly robust and benefit from the diversity of knowledge and experience that the contributors get from taking part in Mozilla's Rebel Alliance.

Mozilla’s rebel alliance helps us influence the industry, share expertise, gain more diverse views, align upstream dependencies and encourage downstream adoption.

This network graph shows to which other GitHub projects Mozilla participants also contribute to, giving us an indication of the ecosystem of open source projects around our contributor communities and the relative strength of the connections through overlaps in contribution.

The yellow cluster represents projects which tend to be run by individuals rather than organizations. Here we also find a few Rust specific projects, but we also find Linux specific projects which border to the blue cluster.

The green cluster encompases a wide range of web technologies and web development tools, such as node.js, react, angular, vue.js, npm, babel, webpack and bootstrap. But we also find tools such as Visual Studio Code, Atom and Electron.

The blue cluster holds programming languages such as Go and Python, but also infrastructure projects such as Kubernetes and Ansible. On the edges of this cluster, we also find Homebrew and Travis, as well as ecosystems of repositories connected to operating systems.

The pink cluster to the right includes web standardization work, including repositories from the W3C and the WHATWG community. The top of this cluster also holds different 3D and VR/MR repositories, such as three.js and gITF.

Finally, the purple cluster at the top is more or less an ecosystem of repositories around ember.js, although we also find Rails in this cluster.

Methodology

The research for this report was executed in partnership with our data provider, Bitergia, a software development analytics company that specializes in open source analytics, with whom we also worked on the 2017 communities and contributors report. This year, we expanded the collaboration to include Analyse & Tal, an employee-owned, co-operative data analytics and visualization firm.

This research followed Mozilla’s privacy policy and lean data practices, meaning we only collected data that were required to deliver valuable insights, we protected these data through strong security, and we informed our communities of the work through direct messages and our privacy policy. No private, personally identifiable information was shared with our partners.

We worked with product teams and community managers across the organization to be sure we were looking at the right set of repositories and linking contribution from Pontoon (localization) and Kitsune (support) to the right products and projects. We did not include repositories we knew were dead or otherwise invalid (e.g. forks).

However, there are surely inaccuracies that need to be remedied in the future. In some areas, there were repos that were hard to categorize in the timeframe we had for this report. The GitHub repos and Bugzilla categorization we used for this report can be found here, and we welcome all feedback at rebelalliancereport@mozilla.com.

A note on the Infrastructure and Web Properties contribution categories. These are both somewhat catch-alls for GitHub repos and Bugzilla issues that are related either to IT infrastructure, such as cloud-based services for Firefox, or content and code for some of Mozilla’s websites. These categories should be better vetted in future reports to ensure a more accurate representation of Mozilla contribution. For example, non-employee contribution to some of the cloud-based services for Firefox might be better categorized as contributions to Firefox.