Wikimedia Blog/Drafts/A refreshed technical collaboration strategy for Wikimedia

Title ideasEdit

  • Onboarding new developers for the Wikimedia movement
  • How Technical Collaboration is bringing new developers into the Wikimedia movement
  • A new strategy for bringing on new Wikimedia developers
  • Helping new Wikimedia developers with a refreshed strategy
  • HELP WANTED: New Developers and Mentors
  • New developers and old mentors: a tale of two cities
  • ...


  • What is being done to improve the experience of new developers in one of the largest free software communities
  • How the Wikimedia Foundation is supporting new developers in our technical community
  • ...


Logic model for the Onboarding New Developers program.
File:Example public domain image.jpg
Caption caption caption goes here. Image by Author Name, public domain/CC0.

Wikipedia is probably the most popular free content creation project in history. Most software developers out there know about Wikipedia and many use it regularly. However, fewer developers (and in fact not many Wikipedians) know that Wikimedia as a whole is one of the biggest and most active free software development projects that has ever existed. Indeed, with more than 500 active developers, more than 300 code repositories, and about 15.000 code contributions on a regular month, Wikimedia is in the same league than the Linux kernel, Mozilla, Debian, GNOME, KDE, and more.

One would think that the Wikimedia technical community must be thriving, based on these numbers and Wikipedia's popularity. In fact, that is not the case. While we have a technical community of professionals and volunteers that crunches a lot of work, and while every week there are a number of bugfixes, updates and announcements that is hard to follow, a closer look shows that the long term sustainability of our developer community might be in trouble.

The population of the Wikimedia technical community is stalled, if not decreasing. Meanwhile, many Wikimedia projects, communities and affiliates are craving for developer help. The causes of this problem are many and complex. Several efforts have been made over the years to address them, without much success. This is why the Technical Collaboration team at the Wikimedia Foundation is focusing on one goal: onboarding new developers. We want to reach out to more and more diverse volunteer developers. We want to offer learning and reputation paths to them and their mentors, and we want to connect them with their closest wiki communities. And here is the thing, we will only succeed with the help of technical mentors and Wikimedia affiliates.

The arrival of new developers is not offsetting the usual attrition that online communities see plus the not unusual conversion of key volunteer developers into professionals hired to work for the Wikimedia Foundation. Conscious of this problem, our team is focusing almost obsessively on onboarding new developers, aiming to support better the people and activities involved.

The first area that we started addressing was our outreach programs, namely Wikimedia's participation in Google Summer of Code and Outreachy. It was clear that our success lining up new developers, projects and mentors didn't bring the expected rates of retention of those new developers after their internships were completed. Nowadays we aim to work with mentors sharing an interest for retaining new developers from the start. We pay more attention to the learning paths and the social connections of these developers during their internships, and we support them finding new challenges and roles after their internships end.

Wikimedia hackathons and other technical events have been another area where we have changed our approach in order to focus on new developers' outreach and retention. One could say that our developer events were successful, and for those participating they clearly were and still are. However, here as well we saw that we are failing at retaining newcomers. In the last editions of the Wikimedia Hackathon and the Hackathon at Wikimania, we have put more attention to supporting new developers specifically, also working with mentors from the start. We have promoted smaller regional hackathons to reach out to more developers, and we have adapted our scholarship processes connecting better the different events, so top newcomers in a local event have a better chance to end up joining our global events.

Outreach programs and developer events were obvious places to start due to the relevance that these activities have in our regular work. However, it was clear that in order to change the declining trend we had to do something new, something different. These are the other main pieces of our annual plan:

  • An explicit focus on diversity. Diversity is an intrinsic strength in creative communities, even more when those communities aim to develop software tools for every single human being. In addition to this, we face a very tough competition when we aim to recruit and retain mainstream developer profiles cherished and sought by professional recruiters everywhere. Meanwhile, we might find a good match for mutual aid among developers related to minorities or marginalized majorities. We want to improve our outreach and support to identify these groups, invite them to join us, and support them.
  • Quantitative and qualitative research. Most of our current knowledge and assumptions are not based on systematic research. We have decided to cut through all the metrics we have and focus on these key progress indicators: number of volunteer developers, number of new volunteer developers, and number of new developers active after one year. We run satisfaction surveys after our big hackathons, and that has helped us improving those events for existing participants, but that was all when it came to qualitative research. We are starting to survey all newcomers who contribute a first code patch, and we are also about to survey those developers who seem to have left. We want to learn more about initial motivations and first obstacles, and also about the factors that make developers leave. We are going to compile the data, findings and lessons learned in a quarterly report.
  • Featured projects and their mentors. Implicitly, we have been aiming to connect potential new developers with any of the hundreds of Wikimedia projects... when the big majority of them are not a good destination for volunteers. Many projects are inactive, others are so active that jumping to them is rather complex, others don't have mentors available or apt documentation. The size of the frustration can be deduced from the amount of patches waiting for code review. After trying to fight strong currents for so long, we have decide to jump away from the main stream, select a reasonable amount of projects ready to welcome newcomers, and work closely with their mentors.
  • Multilingual documentation and support. The selection of featured projects provides accurate information about the documentation for newcomers that we need to work on. While traditionally we have sent new developers to read How to become a MediaWiki hacker as a first step, in fact new developers won't need to setup any MediaWiki environment if their goal is to contribute to tools, bots, gadgets, mobile apps... Probably their Android, Python, JavaScript skills will get them almost to the point where they can start fixing an easy bug after reading just some sensible Wikimedia specific instructions. With a more specific set of documentation to curate, we can attend translation needs better. With a more specific scope covered by active mentors, we can also organize better developer support. In this context, we are are refreshing our developer documentation for newcomers, we plan to refresh the homepage accordingly, and we plan to offer one support channel for new developers easy to find and maintain.

By connecting all these pieces, we aim to attract more developers from diverse backgrounds, and to offer learning and reputation paths that motivate them to stick around. For many of us, joining the Wikimedia movement was a life-changing experience. We want to help new developers (and their mentors!) to walk their own paths in Wikimedia, to gain experience and contacts in our unique community of communities. We want to offer them opportunities to become local heroes fixing technical problems and creating missing features for the Wikimedia communities living in their regions or speaking their languages. We want to offer them opportunities to meet peers across borders and boundaries, working on volunteer or funded projects and traveling to developer events.

We want... to bring the Wikimedia technical community to the levels that one would expect from one of the biggest and most active free software projects, from probably the most popular free content creation project. The chances to succeed depend heavily on current Wikimedia developers (volunteers or professionals) willing to share some of their experience and motivation mentoring newcomers. It also depends heavily on Wikimedia chapters and other affiliates willing to scratch their own technical itches working with us, co-organizing local or thematic developer activities with our help. The first experiments have been very positive (and fun) so far. Join us for more!

Quim Gil, Senior Manager, Technical Collaboration @ Wikimedia Foundation