> initSeed() should be improved, by doing more than simply adding up the
> 32-bit or 64-bit "digits", at least in case of symbols. Any proposals?

I discovered these collisions while comparing the picolisp hash function
with the djb2 string hash function, that didn't have that problem.
May be a solution could be to use the djb2 hash function:

(de djb2 (S)
   (let A 5381
      (for C (chop S)
         (setq A (& 65535 (+ (* 33 A) (char C)))) ) ) )



-- 
UNSUBSCRIBE: mailto:[email protected]?subject=Unsubscribe

Reply via email to