Well Trey Hutcheson brings us some interesting news regarding List across threads.
I’ve seen a lot of concern over performance during very premature stages of development. Worse I’ve actually seen code changes (very complicated ones) made to avoid that possible performance hit. So I find this discovery kind of funny because at the end of the day it came due to premature optimization. From what details Trey provides it dosen’t appear that the original code was causing any real issues (especially since he switched back to it). So by switching his implementation to an implementation that he thought would be faster for the sake of it he actually caused a big performance hit.
The good news is that Trey noticed the big hit the switch had made and did actual metrics to show there was a problem. As a smart man taught me.. show me the performance problem and I’ll fix it. Until then let me code with testing and maintainability as my primary focus.
And if you do happen to read this Trey don’t consider it a hit on you.. we all fall victim to the evil spirit of premature optimization at points.