Interesting to see that people cite code churn as the main reason for _not_ adopting LP. I've argued in the past that LP is especially useful in highly volatile

Interesting to see that people cite code churn as the main reason for _not_ adopting LP. I've argued in the past that LP is especially useful in highly volatile environments: http://250bpm.com/blog:54 http://250bpm.com/blog:54 Tl;DR: If the environment is so volatile that the code is permanently broken reverse engineering doesn't work well. In such case the documentation may be your only recourse.