Last week held the first day of Code for America’s 2012 fellowship. We are more diverse in personal and professional backgrounds (illustrator, designer, developer, urban planner, lawyer-hacker, former freelancers, community organizer, entrepreneur, wilderness youth counselor, Googler, Apple-er, Esri-er, puma trainer), yet more focused in mission, than I could have imagined. Spending the year with these people is going to be equal parts hilarious and powerful.

In a morning devoid of paperwork and red tape, Tim O’Reilly (former board member and now a sort of fairy godfather to CfA), led us in a discussion about what good open software looks like. The fascinating tension and balance between breadth and depth will, I’m sure, surface repeatedly. Should we invest time into a robust platform, or a quick and discrete feature improvement? Run an analysis to support an important policy citizens won’t see until 2016, or make it painless for any neighborhood to get a block party permit next summer?

I suspect the ideal alternates between the two zoom levels, with one informing the other. What do you think?