A most interesting approach, especially within the context of a wiki. From the outset, it would be in contradiction to much of what was discussed in the previous Microsoft topic.
The one thing that I did pick up through out the Wiki was LISTEN TO YOUR CUSTOMER!!!!
One poster was wondering how XP might work with big iron, and Cobol… So from my standpoint, we are talking about 747 reservation systems… But no, their definition of big iron within this category is mainframes… although a Cray would certianly be considered big iron as well.
One especially intereting definition was pair promiscuously Where in pairing is done rapidly. Its sort of the antithese of the dark room mentality.
It will be interesting to see how much of this shakes out in time. My inertia is based upon Big Design Up Front However, in many cases, such a big design is based upon incorrect data.
In other words, the big design upfront driven by the system archetict (which is usually me) is based upon flawsd intelligence. Its not that I don’t listen to the customer, I do. I drive them a little crazy sometimes with questions and options, in order to create the big design up front.
The fatal flaw, is that in the early days of a project, no one really knows…. The sad part is that often times, its not known until Alpha test, and more realistically, much gets sifted through in Beta…. There is nothing like chucking the architecture at the last minute. Especially when thousands of dollars have been spent to get the hardware to that stage.
As such, perhaps XP has some validity in the world of embedded systems. I guess the real key is to have flexible enough hardware such that XP can be used to focus the features in the early days. The next few years will be most interesting.