Stephane,

>Going back to the code i saw the problem and fixed it. The problem was a little piece of code written a long time ago and was not caught by any of our tests because it append only at a very extrodinary time.

But it was a testable error in your code. Just like it was in my case, and as it is in Fabio's case. I think that is the point George is trying to get across. Better design methodology, better testing can result in this open dialog never being an issue from the end-users' perspective.

>I know this will append again, this is what appends when you work with a team of developer not all to the same skills and on a very large system. Even with all regression tests, Q&A, etc.. there is nothing I can do about it.

This boils down to a cost/benefit analysis. How much does it cost to develop adequate regression tests? But at the same time you have to invert the argument and ask how much time did you spend with that user trying to find out what was wrong, then digging into the code to find the place to fix it? And how does that time compare to what it would have taken to have built an adequate test before deployment.