At what price to community?

Dave has written a detailed post on the types of community that Tizen hopes to target. It’s great, and I recommend you read it through. For more background, you should also read my blog post on what’s really going on with MeeGo and Tizen.

There are many details still to come out about Tizen over the coming weeks and months, but I think one of the most interesting and valuable lessons we can learn from Tizen will be the effect that “platform resets” can have on a developer community.

A platform reset is what happens when a significant force in a given community decides to make a fundamental shift in project direction. It’s especially prevalent in projects that have few large corporate sponsors, and projects that are being sustained around otherwise-closed ecosystems.

A platform reset is what happened for example when Maemo and Moblin became MeeGo: a lot of the underlying direct and indirect engineering effort around Maemo was ostensibly thrown out, to be replaced by Moblin’s package system and by a lot of new integration infrastructure and tools that were developed or repurposed specifically for MeeGo. To a lesser degree, it’s also what happened when Ubuntu switched to the Unity desktop.

~

There are a number of effects that platform resets can have on a project. Sometimes, the clean slate enables developers to get involved, take more ownership, and leads to a huge leap forward in community. Sometimes it spurs innovation, and gives people a chance to re-imagine how things should work. Sometimes it is perceived as a negative, with developers wondering about the future of a project and their contributions to it. This often results in fragmented or duplicated communities – like Maemo and MeeGo.

A platform is necessarily a complex beast, and none more so than an open source mobile platform, and there are many points where people will want to try and get involved (as Dave illustrates). It’s critical to the long-term survival of the platform that these complex communities of developers are integrated and nurtured. They need to be empowered to contribute, they need to share the platform vision. They need to see the results of their efforts. The many well-documented community best practices such as open governance and meritocracy help to ensure this.

What, then, was the effect of the Tizen announcement on the various existing mobile platform communities? How well is the Tizen community forming? What was the effect on the MeeGo and Maemo communities?

~

There are many ways that a community can be assessed. The MeeGo Metrics wiki page provides links to many reports on MeeGo community. Unfortunately, not all of the source data is publicly available for all projects, so a simpler approach is needed.

One of the simplest ways is to monitor the traffic on public mailing lists. An excellent example of this are the Apache Software Foundation Mailing List statistics, which gives an at-a-glance idea of the activity across a multitude of projects. Fortunately, MeeGo, Maemo, and Tizen all have public mailing list archives, which we can use to compare activity.

There are lots of obvious caveats that need applying, for example some projects manage themselves through alternate channels such as IRC, some projects use web-based forums and wikis in preference to mailing lists. But in general, mailing list activity can be considered one reasonable indication of community health, and is often one of the few sources of raw data that is made public.

~

Here, then, are the rough and ready results of some preliminary analysis on the Tizen, MeeGo, and Maemo mailing lists. For each umbrella project, I analysed traffic across all available public list archives. Click through for a full-sized image of each, and see Mobile Linux Stats for the breakdown between individual lists and aggregates per project.

Traffic on the Tizen lists:Mail traffic on the Tizen mailing listsTraffic on the MeeGo lists:Mail traffic on the MeeGo mailing listsTraffic on the Maemo lists:Mail traffic on the Maemo mailing lists

It’s too early to tell what the Tizen announcement has done to these communities long-term, so I will be updating these charts on a regular basis over the coming months. Some interesting observations though:

  • The Maemo community, although theoretically replaced by MeeGo, has continued to be busy even after the introduction of MeeGo, and has even seen a similar almost identical rise in activity between July 2010 and July 2011.
  • In July 2011, activity notably dropped in both the MeeGo and Maemo communities. This is too early to coincide with the “Intel dropping MeeGo” or “Samsung acquiring MeeGo” rumours, so perhaps this is the logical result of the Nokia N9 shipping.
  • Compared to MeeGo and Maemo, the Tizen community is non-existent at this point.

I’d welcome suggestions for improvements or for other platforms and projects to monitor as part of this ongoing activity, or other analysis that could be performed on the data.

~

It’s particularly important to understand the effects of the Tizen ‘reset’ as we are about to enter into another era of platform fragmentation. After the extensive consolidation we saw in mobile initiatives in 2006-2008, we’ve now seen the death of ‘mobile linux’, and the fragmentation of Android with a proliferation of Android mobile platforms.

It’s possible for any one of the six vendor-driven Android-based mobile platforms predicted for 2012 to declare itself a “fork” of Google’s Android. Any of these platforms can implement the type of open governance and meritocracy that Google Android is missing. Several of these vendors have the skills, the understanding and the manpower to make it happen. And if it does, we will start to see significant pressure on the Android community – in fact, exactly the pressure that’s being exerted by Intel and Samsung’s fork of MeeGo through the creation of Tizen. Maemo, MeeGo and Tizen give us some perfect case studies for understanding the pressures and likely outcomes of Android divergence in the future.

This entry was posted in Research and tagged , , , , , , , , , , , , . Bookmark the permalink.