Robert posits that the technical aspects of agile development can apply perfecty to open source development. I kind of rate this a "duh" -- this is things like continuous integration, ubiquits version control, test first, etc. I don't think anyone is going to disagree here. It is seeing if there is much that can be done with the agile management techniques that is intrigueing. Iteration planning, blog as substitute for scrum meeting, etc.
One of the fundamental pieces of most agile methods is that time is fixed and scope is not. Most open source projects don't fix time -- unless there is funding it is very tough to maintain any kind of consistent velocity, which makes fixing time pretty pointless. An approach Robert has talked about is fixing time-units or even (gasp!) fixing estimated scope -- and being actively willing to re-estimate effort levels for a story and adjust the iteration target to stay on a target number of points. I am phrasing this differently than Robert has, but it is how I describe what I have watched him do =)
Anyway, I am looking forward to seeing Mesquite written up as a case study -- it will make a fantastic one I expect.
Previous Topic
