ExtremeProgramming or more precisely extreme programmers have cultivated quite a few myths over the time. Maybe it's useful to bring them back to their real status of assertions, which might be truthful from time to time depending on the situation. They might be truthful in the majority of cases one can encounter in practice, however this should not promote them to the level of God sent truth.
Here is a list of some:
Two eXtremos (PhlIp, BenKovitz) do not believe any of these myths. Engineering is about tradeoffs, and the above items are all absolutes.
One of the MythsAboutXp is that eXtremos believe in myths like these.
Furthermore, if the above myths "might be truthful in the majority of cases one can encounter in practice", then they are clearly not myths at all.
Some experimental results about test first has been published.
http://domino.mpi-sb.mpg.de/internet/news.nsf/0/f35061940b316ea7c1256b870050b767/$FILE/tichy.pdf
More independent evaluations and analysis can be found here:
The biggest myth is that ExtremeProgramming projects usually succeed. In practice, they fail just as often as with other methods.