The controlled test turned out to not show the symptoms at all. So now the QA person is going to have to figure out what he did to get it into the state where he saw the problem. I suspect that the database wasn’t in the state he thought it was in, and the problem is going to devolve into another finger pointing exercise, this time between the other group (the guys who send us the playlists) and a third person in our group (the guy responsible for accepting those playlists and putting them in the database). That one won’t be nearly as much fun, because Tony (the guy responsible for accepting the playlists) doesn’t have anywhere near the ego that I do.