June 17th I was at MIT to attend the 2010 Future of News and Civic Media Conference. Civic hackers galore, including the 2010 Knight News Challenge winners. I was there because of the “Data into Action” plenary, where Nick Grossman from OpenPlans was on the panel — among other things, announcing OpenBlock.
What’s OpenBlock? In a nutshell, we’ll be leading an effort to make the technology behind everyblock.com more accessible to news organizations that don’t have a huge web budget. So a local paper could put up maps of “hyperlocal news” – stuff you care about happening in your neighborhood, down to the block level. My role will be to lead the core infrastructure improvements and open-source community outreach. More on this soon — the action will be taking place at openblockproject.org.
Nick’s “Data into Action” slides are are here. Also on the panel were Ellen Miller from the Sunlight Foundation and Laurel Ruma, “Gov 2.0 Evangelist” at O’Reilly Media. Laurel’s slides are here (I can’t find Ellen’s slides, are they online?).
Read on for a complete conference braindump…
Name dropping for fun and profit: It’s pretty gratifying when watching a talk by your boss and from the audience comes a tweet like this about your gig … and then it’s even nicer to later learn it’s from Mark Surman, executive director of the Mozilla Foundation. Cool!
Hallway schmoozing: After the panel I checked in with Nick, who introduced me to Judy Klein, former mayor of Palo Alto; she’s now doing work on disaster planning for the silicon valley area. I told her she should check out GeoNode, an awesome project from our OpenGeo team. She told me about her work with InSTEDD.
Lunchtime: people wanted to talk to me about OpenBlock. Yay! It was great to meet the staff of the Boston Globe that we’ll be working with on that, particularly Joel Abrams. Also Amanda Hickman from DocumentCloud (a resource for journalists to post primary source documents, with a bunch of tools around that); she had some good input on wanting better address recognition than everyblock currently provides.
After lunch I attended a session titled “Leading a Free Software Project.”
Mostly the session chair talked: a guy named Mako who wrote a (free) online book about the subject and also recommended the Karl Fogel book (also free online) which I’d already read. Mako has worked on Debian and Ubuntu among other things.
Institutional independence is good. If there are paid employees doing work (like on Openblock), people are less likely to volunteer. Modularization can help: Provide clear ways to build on top or add one little feature. Eg. Firefox doesn’t have a lot of volunteer core contributors, but a huge extension-building community (Relevant to OpenBlock: we need lots of “scraper” scripts – more about this later.)
Reduce barriers to entry. Eg. your bug tracker should accept anonymous reports. There’s a transaction cost associated with contributing: effort vs. perceived payoff. The incentive to fix a bug for yourself is greater than the incentive to contribute the fix to the project. So, reduce drag.
Infrastructure, governance, rules, etc. should be set up only as needed; don’t solve problems you don’t have yet.
Q&A: What is the role of leadership? NOT resource allocation, that’s uncontrollable. There was a PostGreSQL study: new features mostly came from outside the community per se — often unusable as submitted (even occasionally arriving in the wrong programming language) — people closer to the core then took those features and rewrote / polished as needed. A “leader” can serve to explicitly highlight and recognize good contributions. You don’t have any sticks you can wield, so use carrots.
Q&A: Charlie Detar talked about his own OLWidget project as an example. It has perhaps 100 users, 3 occasional contributors; statistically that puts his project in the 90-something percentile. The median number of contributors to an open source project is one.
Q&A: How do you attract developers? Attract users first. Almost zero dev contributors are not users.
More hallway schmoozing. I met a guy named Ryan Thornburg from the UNC J-school, who was interested in possibly using OpenBlock for a class project next term. Unfortunately I had to tell him that it wouldn’t be ready for non-techy users to set up by then; we will be working to lower the bar as far as possible, but some problems are tough. For one, you may always need somebody with enough Python skills to write scraper scripts – scripts that can interpret the data you care about and feed it in to your system. Nick suggested we look into Yahoo Pipes as a possible tool to leverage here. I’m hoping we’ll amass a nice pile of community-contributed scrapers.
Then I talked to Jennifer 8 Lee for a while, who isn’t merely friendly – she’s one of those instant-hub people. Everybody seemed to know her and I swear she grabbed every person that walked by and paired them up. “Oh, have you met so-and-so here? You should work together on [insert something she happened to know both of them were working on independently]”. For example, some people that were interested in making FOIA requests easier. Not surprisingly, she knew a lot about OpenPlans already.
Then I caught up with Nick for a few minutes, and unfortunately had to rush off to catch a train home. So I’m missing the grant winner demos, the dinner, and the afterparty. Normally that would be no great loss; I’ve been to some good conferences and at least one incredibly dull one, but this was unlike anything I’ve seen so far. The small size helped, making it possible to meet a lot of people without feeling overwhelmed. But really it was the focus of the event that mattered. If you put a bunch of tech-savvy journalists and open-data techies in the same space, the energy level is going to be pretty high.