Re: hash function

From: Aldar Chan (I_love_Nora_at_Nora.and.Aldar.net)
Date: 02/22/05 

Date: Tue, 22 Feb 2005 15:57:34 GMT

"Navin Kumar" <navin.kumar@gmail.com> wrote in message
news:1109052623.807997.7200@g14g2000cwa.googlegroups.com...
> the result for any public b. This is vastly superior to a construct
> like: h(x || "public hello world") for iterative O(1)-memory hashing.
> Iterative O(1) hashing is horrible for what you want, because, if it
> can store only n states, and the tail length(b) occupies k of these
> states, you're needlessly making the influence of x merely lg(n-k) bits
> of the result. As k grows larger, your solution becomes weaker.
> Instead, you want a solution where x has an invariant influence.
>
I have the same feeling about what you called interative O(1) hashing
although it is so widely used.

Quantcast