Extreme Planning Simplified

C3 is starting to experiment with an even simpler approach to IterationPlanning.

One of our motivations has to do with discipline: we have generally failed at tracking well enough to measure load factor. RonJeffries is constitutionally incapable of doing it (good to know there's something he doesn't excel at), the guy who is supposed to do it has similar difficulties. Doctor, it hurts when I try to Track.

In his recent talks, when he's not DeclaringTheDeathOfEntireProgrammingLanguages, KentBeck says that a key to PlanningGame is to know your velocity: the pace at which you can generate functionality.

C3 will continue to estimate cards in ideal days. At the end of the iteration, we'll count up how many ideal days' work we got done. This will give us a vector of the number of ideal days work we've historically been able to accomplish. This will let us tell Business how many cards we can do. We can also derive load factor by the obvious math using days worked, number of employees, cans of soda consumed, etc.

In CommitmentSchedule, my current guess is that we'll just keep doing it like we always have: estimate in ideal weeks, multiply by three. No reason to change.

We'll get as good or better a measurement of inefficiency from the graph of days' work accomplished as we can get from LoadFactor. So on that side I think we'll be OK.

With one less reason for Tracker (TrackerRole) to go around, the chance of people going off track will go up. We'll address that by being a little more inquisitive during the stand-up meeting, and I'll be more proactive in sitting down with people to see how they're doing.

So ... some discussion, please. What's wrong with this idea? Thanks! -- RonJeffries

"In his recent talks, when he's not declaring the death of entire programming languages, KentBeck says..." What an extremist!

Makes sense. More feedback.

The estimation technique has a few flaws: gut feel isn't very objective and the unit of measurement confuses some people. I'm hopeless at it too. Does anybody around here do FunctionPointAnalysis or one of its derivatives? -- MattRickard

Practice makes perfect. We can estimate 10 people's work for three weeks in half a day. -- RonJeffries