YouReachedTheEnd

What is wrong with this picture?

Certainly, finishing a task doesn't tell you anything very useful about the total cost to your customer of your choosing INTERCAL for the implementation of their risk system, but it tells you a great deal about how hard or easy it made that task you just did. After that iteration where everything seemed unreasonably difficult, maybe you'd get to thinking that perhaps you'd be better off writing the decision tree management code (the story coming up next) in Befunge instead. Or maybe you'd think back to that project you worked on where the front-end was built in TCL, the pricing engine in Lisp, and the...hey! Wait a minute...maybe...but no. Of course. You can't know that you could do a better job more quickly with those other tools. No. Best get back to pounding out those punched cards, after all, this isn't the end, so you don't have the information.

And you can't guess right at the beginning.

But another idea that XP has is that development activities should be a bit fractal, a bit self-similar, somewhat without characteristic scale.

Part of the joy of XP is getting to the end all the time. The end of writing this test case. The end of this refactoring, the end of this task, of this story, of this iteration, this project.

And just as every sale is a purchace, every end is a beginning.

Maybe there is a group of developers who have no knowledge upon which to base their decisions at the beginning of a project. Fresh computer science graduates, perhaps (and that's what you've got, it really isn't going to matter what language they choose).

Meanwhile, part of the reason that good people are expensive is that your new project is their next project.

You can't look into the future to know the outcome of the decision you make now. You don't need to, because the future will be right along in just a few minutes. -- KeithB ³³agent³Mozilla/4.0 (compatible; MSIE 5.12; Mac_PowerPC)³YouReachedTheEnd