2023:Program/Submissions/Surpassing RfCs: how to improve the governance of large software changes - N3JLEG

From Wikimania

Title: Surpassing RfCs: how to improve the governance of large software changes


Tisza Gergő

I am a long-time Wikipedian and MediaWiki developer. Originally from Hungary, currently working in the USA (for the Wikimedia Foundation - only representing myself at Wikimania though). I'm interested in improving technical infrastructure, supporting volunteer developers, broadening and improving community self-governance, implementing movement strategy, and thinking about the future. See my user page for more info.

Pretalx link

Etherpad link


Start time:

End time:

Type: Roundtable / open discussion

Track: Governance

Submission state: submitted

Duration: 90 minutes

Do not record: false

Presentation language: en

Abstract & description[edit source]

Abstract[edit source]

Our decision-making processes for large-scale changes to wiki software perform poorly: they create unpredictability and discord, don't incorporate expertise and data well, and don't live up to our standards of inclusivity and shared power. Can we do better?

Description[edit source]

PLEASE NOTE: This session has been accepted as a 90-minute roundtable discussion, but then rescheduled into a 30-minute slot. It's not really feasible to have a roundtable discussion in 30 minutes so it will be a presentation with Q&A instead. If you are looking for something more interactive, there is a roundtable discussion with a similar topic: Discuss improvements to our collective decision-making processes for software releases

Large-scale user interface changes tend to be managed by some combination of the RfC processes of large wikis, and the WMF acting as unilateral decisionmaker. RfCs (the way they are practiced in the Wikimedia world) are a poor decision-making process for matters requiring immersion in details and expertise that most experienced editors do not possess; and the WMF monopolizing power doesn't live up to the Wikimedia movement's standards of participatory governance, erodes trust, and is vulnerable to bad organizational incentives. In other words, the status quo is broken.

Community members should have meaningful influence in the decision-making process (more than just "being consulted") so the decisions align with the movement's values and don't ignore the hard-won experience of editors in the trenches; but interested amateurs shouldn't crowd out expertise, decisions should be driven by data and experimentation, not prejudices, and change aversion should be understood and accounted for. Moreover, decisions should be predictable and avoid wasting of donor money, while giving the Wikimedia community a meaningful influence over how it is spent.

The session will consist of a short presentation on the history of regulating software changes, the problems with the present state, and the solutions proposed in the movement strategy, followed by an open discussion. My hope is that this and follow-up discussions lead to a proposal for a new governance structure, which can be included in the Movement Charter drafting process.

(About me: I have run both global and local consensus-building processes as a volunteer developer and a Wikipedia admin for over fifteen years; I worked for the WMF on one of the most controversial software changes; I was a chapter board member during the time when the WMF erased the financial independence of chapters; I worked on movement strategy. I feel I am familiar with WMF-community power struggles from both sides, and with the benefits and drawbacks of RfCs.)

Further details[edit source]

Qn. How does your session relate to the event themes: Diversity, Collaboration Future?

I see software governance as one of the major collaboration failures in the Wikimedia movement today, and as the kind of complex and possibly emotionally intense topic where high-bandwidth face-to-face discussions can be beneficial. Software governance is the topic of the Technology Council movement strategy initiative, which I hope this discussion could move forward.

Qn. What is the experience level needed for the audience for your session?

The session is for an experienced audience

Qn. What is the most appropriate format for this session?

  • Tick Onsite in Singapore
  • Empty Remote online participation, livestreamed
  • Empty Remote from a satellite event
  • Empty Hybrid with some participants in Singapore and others dialing in remotely
  • Empty Pre-recorded and available on demand