Measuring Project Velocity

See LoadFactorArithmetic et al.

Ron- what your wrote is what I wrote that was clobbered by what you wrote. kb

[Draft] LoadFactor relates a developer's estimate to real time. We (C3) tell developer, "Think how long this task would take you and your partner to test and code if you did it with no interruptions."

We track individual tasks by asking the developer: "How long did you originally estimate? How long did you actually work on it?"

Thus the developer gets to hear herself say "I thought this would take me one day. It took me one and a half." This provides feedback to the developer on how accurate her estimates are.

Note that so far we have not addressed LoadFactor, just developer estimation feedback.

At the end of the iteration, we see how many tasks were completed. We take the ratio of the elapsed days to the (original) estimates for those tasks. That's LoadFactor.

A measure of ProjectVelocity which C3 does not yet use would be the total of the estimated days for the tasks completed in the iteration.

Either LoadFactor or ProjectVelocity provide feedback to the project on how fast the overall pace of development is in terms of developer estimates. --rj

There are four measures relating to velocity:

The next iteration's velocity is the sum of the estimates from the previous iteration (not the actual IdealProgrammingTime). The LoadFactor is the ratio of actual IdealProgrammingTime vs actual CalendarTime?. But if LoadFactor is deprecated, why should we track the actual IdealProgrammingTime or actual CalendarTime?? Clearly, tracking the IdealProgrammingTime is useful to improve your estimates, but this is irrelevant if you are using last iteration's velocity. If you are estimating the same way, it does not matter how off your estimates were as long as they are off by a linear factor (in other words, a task estimated at 2 takes half as long as a task estimated at 4). This is because you know the ratio of whatever unit you are using to estimate and the amount of calendar time it takes to complete (assuming you estimate the same way).

So why track any actual time besides iteration length? -- Justin Francis

CategoryExtremeProgramming

EditText of this page (last edited October 9, 2005) or FindPage with title or text search