Recap of OuterConf2013, the Outercurve Open Source Software Conference
5/16/2013 1:14 PM
Last week marked the first Outercurve Open Source Software Conference. It was a great first conference. The Outercurve Foundation has a growing set of project communities, and we wanted to provide project developers models for success in their communities.
There was a great speaker line-up that included excellent experienced community leaders from outside the Outercurve Foundation (Jono Bacon, Ross Gardler, Donnie Berkholz, Kohsuke Kawaguchi) and project leaders from some of Outercurve’s larger communities (e.g. Bertrand Le Roy from Orchard, and Phil Haack, Jeff Handley, Howard Dierking from NuGet). Jono Bacon was our keynote for the first day. We were very fortunate to land Scott Guthrie to keynote our second day. Scott was the initial driver at Microsoft that put the wheels in motion that resulted in the creation of the Outercurve Foundation. He gave a great talk on how his team continues to work in the open source community at large to improve the experience for customers between Microsoft technology and open source communities.
A couple of broad themes were present over the talks. One of things that folks often wrestle with in communities is trying to determine what conversations are loud, opinionated but genuinely constructive and which ones become destructive to the community. To that end we created Outercurve Fight Club and followed it up with Donnie Berkholz’s talk on “A**holes are Killing your Community.”
The fight club started as a discussion between Ross Gardler and I on how much we hate panel discussions at conferences. In general they are content-free (or you sit through an hour of contrived discussion for a couple of small gems of observations), too positioned, and can often become competing product/service pitches. By the time the moderator takes 5-10 mins to introduce the topic, and each panelist gets “a few minutes to introduce themselves” the panel is half over. Ross observed, “You don’t want a panel -- you want Fight Club.” And I knew just the people. The Outercurve CoApp, NuGet and WiX projects sit in complementary slightly overlapping spaces. Each has very strong opinionated leadership. And each project leader has a wealth of depth and experience in the nuances and relationships between security, dependency tracking, installation versus upgrade versus uninstallation, authentication and verification. And the Fight Club was born.
Garrett Serack (CoApp), Jeff Handley (NuGet) and Rob Mensching (WiX) drove a set of discussions about the relative strengths and weaknesses of their solutions and the other solutions, and the history of why certain decisions were needed, and the sensitivity to their respective communities AND product decisions around the use of the projects at Microsoft. They were loud. They were lively. They were deep. And at the end of the hour, they had collectively demonstrated what constructive critical discussion looks like. It would have taken [possibly tedious] hours of reading on a development email reflector to have seen what was shown in Fight Club.
After a brief pause, Donnie Berkholz gave his perspective on destructive conversations based on the metrics they tracked in the Gentoo community. Many thought that the fall of Gentoo popularity coincided with the rise of Ubuntu, but Donnie was able to more accurately demonstrate that the Gentoo community continued to grow until certain destructive developers began discussions in community and the collateral damage that it caused. They lost 20% of their community, and the sad part was even after the bad actors were managed out of the community, it never bounced back. One of the things they do in Gentoo to track potential trolls in the community is to log bugs in the bug tracker. Bugs in community are project bugs too.
Another broad theme through the talks was the need to make it easy to join the community and easy to contribute. Jono as opening keynote talked about how the Ubuntu community continues to make it easier and easier to understand how to get involved, and to ensure a developer understands how to contribute and can quickly succeed. The intent is to capture drive-by contributors and turn them into long term contributors in the community. Kohsuke followed this up with a discussion on how the Jenkins community relentlessly pursues ease and efficiency in encouraging users to become contributors and then to make it equally easy to contribute. He believes that communicating in code is more efficient that communicating in English and can improve the efficiency of the contribution pipeline. Kohsuke observed that every contributors starts as a potential user visiting your project’s website, so that is where it all begins. Likewise, Donnie reminded folks that people lurking on your forums and email lists and IRC are the people thinking about joining your community.
Ross Gardler led the discussion around governance. He started with the observation that while the license provides the legal framework to a project, a project governance model provides the social framework for the community. A lot of folks wrestle with the governance question in their communities. Is it better to be a meritocracy similar to the Apache Software Foundation or Eclipse Foundation projects, or is a benevolent dictatorship the way to go to quickly move forward? After reminding folks that a meritocracy is based on useful activity and very much NOT a democracy he shared a number of keen insights. He wrapped with the observations that a benevolent dictatorship is really a meritocracy of one, and the only true difference between meritocracies and benevolent dictatorships is how the project breaks deadlocks when decisions are needed.
It was great to see the observations presented by Bertrand Le Roy about the Orchard community and Phil Haack, Jeff Handley, and Howard Dierking about the NuGet community. Each of these communities are wrestling with growth and success in their own way. Bertrand pointed out that while there are very few Orchard installations on the Internet relatively speaking (~37,000), they are often very high traffic sites. He observed wryly that a community needs to learn to be happy with their numbers, i.e. focus on the community and not numbers like downloads, or installations, etc. The NuGet presentation was equally interesting. At this point in history, NuGet is a part of the shipping Visual Studio product line. That brings certain stresses on the product. The more important community aspect of the growing NuGet community is that it is now part of many users’ continuous integration builds, i.e. the NuGet gallery needs to run 7x24. That brings its own set of success stresses.
Lastly, I gave the first of what I hope to be a series of presentations on an Open Source Software Project Community Framework. It’s a set of patterns and practices that successful growing open source projects demonstrate. The life blood of a FOSS project comes from its contribution flow. This is tied to the number of users. There are a set of measurable activities that projects do to grow their user base to grow their contribution flow to their success. Look for more on this.
I’ve started a Flickr group to collect together photos from the event. If you have some, do please contribute. We recorded the event and I’m presently editing and preparing video for uploads, and I’ll keep folks abreast of updates as they happen. It was a great event. Thank you to all who participated, and I look forward to seeing you all next year!