very well put. Still, one totally different question is nagging me. Although I myself totally *believe* in acceptance tests as not only a good quality agent in the development process but also as an acelerator towards a good lead time as well as a good and solid release cycle, I have the following problem:

As I already said, I belive in that, I think that we have relatively few facts about ROI of acceptance tests. Given the situation that for any medium sized organization upwards this is a tremendous change, it might be that we need to collect for fact based ammunition to convince our environments?

Just think of my environment where we have a huge code base with only few automated acceptance tests and imagine being the first team including that in your definition of done, process or whatever. Not an easy and not a nice job, and one easy to convince your peers, upper management and team members of.

And it is even hard to argue that these are *small* issues ;-)

Cheers

Markus