Load Factor In Estimating Other Projects

AlistairCockburn expresses a concern about estimating other people's projects: "it doesn't work to map other people's logged programming hours to XP staff logged programming hours.

Recall that C3 developers report ideal time spent programming. This includes all time spent programming and some other time such as interrogating customers.

Now then, suppose we get statistics from another project. They report that they did 1,000 days of actual programming. We assume that what they report as time spent programming is the same as what our developers report when they say how long they spent programming, that is, ideal days. This assumption may not be valid, but it seems to be the way to bet.

We have measured that our programmers produce 1,000 days of what they call programming in 2,500 days. We assume that there are 1,000 days of what we call programming in the other project, because the other programmers spent 1,000 days doing what they call programming. Therefore we estimate that our programmers can do that other project in 2,500 days.

I state that the above can fall apart if what the other guys call programming isn't what I call programming. But I assert that it can't fall apart on the arithmetic.

What am I missing? --RonJeffries

---

One basic assumption that seems to be being made is in the nature of how XP speeds up the development process and/or makes it more efficient. Either:

• XP enables programmers to reuse their code better, and spend less time debugging the results of their changes, resulting in less IdealProgrammingTime per arbitrary unit of functionality.

..or..

• XP enables programmers to focus better on the actual programming, and spend less time producing non-deliverables etc, thereby reducing their LoadFactor relative to non-XP programmers.

..or some combination of the two. It seems to be the fact that something like this seems to be implied by XP that leads people to believe that there shouldn't be a direct correlation between XP project estimates and non-XP estimates. Thoughts? --DjMoon?