Once all of the stories have been defined in sufficient detail to allow development, the code itself will be written, both for the actual functionality associated with each story, and for automated testing of that code – unit-testing with regression-testing capabilities baked in. If possible and practical, integration- and system-testing code will also be written with an eye toward providing the same automated, repeatable testing of new code from those perspectives. The end goal of each iteration will be a deployable and functional code-base that has been tested (and that can be retested on demand). It may not be complete or even usable during the early iterations, but it will be stable and predictable in terms of which capabilities it provides.

This part of the process will form the bulk of the next few chapters. Writing code is, after all, the key aspect of development.