I sometimes wonder if I spend too much time making test data for scenarios that no one cares about. Is it the back of the fence or the shape of the nail heads?