I l @ ve RuBoard |
The Release Plan document that is generated from all of these exercises includes
The budget and timeline estimates are given in man hours, and a man week is assumed to be 25 hours long. (You can pick your own number to define a man week. Ours comes from our experience that an average person working 40 hours a week can be expected to do 25 hours of actual customer work. Some do more and some do less.) This benchmark is our expected individual velocity. Velocity is an important concept in XP. [2] Some measure it by how many function points ”individual pieces of functionality ”are done in an iteration, but function points are difficult in Web projects because not everything is a function. [3] On Web projects time is the only common denominator between the various disciplines. We will deal with velocity in detail later, but for now understand it as the work effort you commit to. To every iteration you are committing a certain amount of scope. Using the number of hours you are in the office as a guide to how much work can be accomplished is a quick way to get in trouble. Velocity is all about being realistic about what can be done in an iteration and determining how large a team you will need.
We give the customer multiple scenarios in the Release Plan for various team sizes and for soft dates for launching the site to a review group and to the public. For Web projects here are some helpful metrics for estimation:
One big advantage of involving the customer in release planning is that you will have less explaining to do once the plan is presented to the customer and the customer will have a sense of ownership of the document. This can have an enormous impact on the project success. |
I l @ ve RuBoard |