Fact of the matter is, the usual functional designs made by a.o. business analysts just describe the desirded situation, hardly ever the, and never all, undesired ones

This leaves developers fairly free in deciding which situations to regard as error, and which to handle as exception. And that’s why we are so used to get illegible messages (exceptions) when we make a silly mistake

At the end, there’s the testers. They basically can’t possibly have a clue about all unwanted situations as those just aren’t (all) described. So they do their best. And sometimes, their best is not “the” best when they go at length to find bugs with hardly any impact that will only occur if you fall asleep on your keyboard while hitting those few keys in exactly the right order <– stretching it to the max here but trying to make a point

I thought and hoped offshoring would make a difference, forcing designs to cover all wanted and unwanted situations. Then again, that would take the creativity out of developing and testing, wouldn't it? Nonetheless, it hasn't happened. All kewl stuff developed over the last decade was just aimed at faster development, doing more projects in less time

Will the perfect application ever exist? It probably will, with a lifespan that lasts until the next release

Someone’s already picking which bugs to fix (and which not), by some logic or other (probably consisting wholly or in part of “we ran out of time and resources”). At best, if the QA team actually knows more about the business than this person, this technique might ensure that the more important bugs get fixed–but it’ll be at the expense of those bugs formerly thought to be more important.

“Fix better bugs,” maybe, but not “more.”