I think that's about as close as you can get to a usable metric. Of course, it still has problems. The metric probably has to be calculated long before the nu

I think that's about as close as you can get to a usable metric. Of course, it still has problems. The metric probably has to be calculated long before the number of serious bugs is even known. On the other hand, good developers anticipate needs; there may be features implemented that the users haven't even realized they wanted yet. And of course, it's hard to come up with numerical weights for the features and bugs. But the worst problem with this metric is that it doesn't count the maintainability of the code. That's an even harder thing to measure, of course.