AA implementation algo should be mentioned

[Sean Kelly]

Frits van Bommel wrote:
> Frits van Bommel wrote:
> [about inserts]
>> IIRC the Phobos implementation uses a hash table with binary trees for 
>> collision handling, which would mean /average/ O(1) for "good" 
>> toHash() implementations, but worst-case O(log N) (for "bad" toHash()s).
> 
> Oh sorry, forgot about resizing. What does that make it, average 
> amortized O(1)? :P

I think so.


Sean