Wikimedia Engineering/WMF Tech Days 2012/MTN

From MediaWiki.org
Jump to navigation Jump to search

DV: Does Wikipedia have a "carrying capacity" -- a number / range of editors that can keep the quality of Wikipedia stable? Perhaps it's naive to think that it's possible or even desirable to grow the editor population without limit. We should consider the possibility that the current size of the community is optimal.

EM: What little data we have about editor engagement and retention suggests that people are being attracted / driven away by rather quotidian factors -- technical issues, community politeness, etc. Why attribute those to deep "ecological" reasons? Even if there is a "carrying capacity", it's not likely that we're hitting it, what with so many platform warts.

DT: Limits on growth may only apply for a particular type of editor population. Changing the composition of the editor population (for example, by enabling new types of contributions, expanding access to mobile users, etc.) could have different scalability characteristics.

EM: Our goal should be to unblock motivated and capable contributors, rather than speculate about "macro" questions, which are abstract.

FF: We should adopt an expanded sense of "contributor" and make it possible (by building tools, etc.) for people with a broad range of competencies to contribute to Wikipedia in different ways.

EM: The "limits" you discover have much to do with how you formulate the problem.

DT: Facilitating different types of contribution / engagement runs into questions of what the goal of the platform is.

FF: Anonymity is a major blocker -- it's hard to reach out to anonymous contributors, which makes it hard to understand (and respond to) their needs.

EM: more personalized experience (specifically: notifications) might help us understand and communicate with anonymous editors.

DV: We've talked about the supply side, but we still don't know how many more we need. Would it be possible to come up with a way to estimate?

EM: Gender ratio, underrepresentation in general -- it's not just a question of population dynamic or population size.

DV: We do have goals stated; we could start by trying to formalize them.

RK: Would there by a point, after having tried many different approaches to moving the needle, at which we would conclude that there may actually be deeper limits? Would we then change our approach, and if so, how?

EM: But we already know that some things are successful -- there are things that we are doing right now that are in fact moving the needle.

FF: When we think of editors, we have to stop thinking about just textual contributors, and start thinking about the unexplored territory of possible contributors and contribution types.

EM: The first thing is to actually try! Reduce frustration, improve usability, improve performance -- so many of these things have actually never been done! Consider how much you need to know to be an effective contributor.

EM: What we're trying to do is help the person who would have made for an excellent editor if they had joined in 2001, but who is unlikely to become one now because of the complexity that has piled up over the past decade.

HF: To challenge the notion that everyone that could be an editor has already been exposed to Wikipedia, just consider this: How many people start college every year? 3.7 million in the US alone every year!

DT: the prevalence of reverts and the mutability of content makes it difficult for people to feel ownership over their contributions to the site. We should try shifting toward a model in which you can maintain ownership somehow over time, rather than have them be totally obliterated in time.

EM: The system should facilitate collaboration.

SW: But consider that being able to surrender your contributions to public ownership is central to how Wikipedia works and may be an intrinsically desirable characteristic for editors to have.

EM: The fact that Wikipedia was so obviously incomplete and inadequate in its early years actually facilitated collaboration and participation. The increase in quality and scope of coverage now provokes very different emotional responses -- protectionism, jealous guardianship, etc.. The measure of "value" has become considerably stricter and harder to meet.

DT: We should facilitate collaborative editing, but also provide a protected, less mob-like experience for bringing new users on board. We need to encapsulate contributions somehow so that they are stable and can provide supports for a stable online Wikipedia "identity" -- i.e. make it easy for people to point to past contributions as a way of communicating who they are and taking pride in their contributions.

MP: The kind of people who become successful Wikipedians are those that are attra



  • Building code/regulations are wanted by experienced contributors
  • Intractable problem? Smaller contributions are less valued than 5-10 years ago

-> Maybe create side working spaces for improving/creating content incrementally

Problem with separate workspaces can be that the quality expectations become even more unrealistic (e.g. article creation wizard)

Teahouse takeaways?

  • Numbers suggest small scale success
  • Lots of mentoring mechanism have shown similar small scale promise, but all are clunky
  • Not a single obvious way for the newly registered users to get help

How do we move the big number?

Hackathon project:

  • Realtime editor metrics for retention