Personally, I agree that hash tables are a lot faster when the number of items is quite large (although they're about equally fast when that number is small), simply because my tests have shown so. I wasn't arguing this though; my point was that nobody should take any opinion (including mine of course) on such things for granted, because no matter whom it is coming from, it might be inaccurate. Either because anybody can make mistakes or simply because one can't know what mirc does internally. In the end, the only opinion you can rely on is $ticks's.
