in reply to
help! red-black binary tree problem
One of my first perl modules ever was an attempt at a family of binary search tree modules. The few programming classes I took taught a lot about trees and they seemed really neat, especially splay trees which could self-optimize.
I discovered this: The performance stinks. Unless you're dealing with a bazillion records, just use a hash. Even having to sort a hash each time the performance will blow a tree written in Perl out of the water. One of those cases where C is significant in calculating algorithm efficiency.