Call it professional bias, but I like to think about it as a big VM. If for some reason, things go awry, the (said) programmer will hit the pause button, then diligently go to the last valid snapshot, fix the bug and then get things going again. If this isn't desired, fixing the current mess is also an option.
The time required for fixing bugs isn't really an issue in this case, he can take as long as he likes.
The time required for fixing bugs isn't really an issue in this case, he can take as long as he likes.