Monday, July 14, 2008
An interesting article on User Stories estimation techniques
I've read with interest an article on InfoQ by Jay Fields covering some tips on estimation.
In particular, some points raised my interest:
1. The concept of throwing an estimation, in which every developer should give his own estimation together with others so that no one is influenced by others. This concept is put in practice with a sort of rock/paper/scissor style game that it looks really, really amusing. I must try it.
2. The concept of estimations that get stale. It is very important for me, and I also written about it in a previous post, describing my usual practice to add an expiration date to every estimate. As time passes, estimates are less and less realistic until they are completely useless and meaningless. In this article, estimates get stale after 12 weeks, and it seems a good time limit for an average project.
3. The overall focus on people involvement. This concept applies in small rules like: no laptops; small team size (no more than 6); business people involved; everybody expected to estimate and read user stories. This is very positive, because estimation meeting do not have to be focused only on one domain expert, a team manager and a leader developer with all other people IMessaging around or just watching outside the window. The estimation, like development, should be the result of a whole team effort.