posibility of directory "hashing" of cddb

Hi there,

 I'v noticed that one of the first filesystem optimization techniques
for large directories (directory hasshing based on first/second/third/etc.
character) aren't benig employed for the freedb-{complete|update}
data.

Is there a reason for not doing it? (except perhaps that the code 
haven't been written ;^)

Greetz
Hendrik