Re: MD5P (Variable length hash, kindof)
Available news archives: comp.lang.tcl - comp.lang.python - comp.security.firewalls - sci.crypt - comp.lang.php - comp.lang.javascript
Google
 
Web news.hping.org


sci.crypt archive

Re: MD5P (Variable length hash, kindof)

From: aie93 <aie9321@gmail.com>
Date: Wed Dec 21 2005 - 12:46:32 CET

Yes, the whole Idea is states[2] and states[3] have a whole lot of 0's.
 That's why it's a variable length hash.

But the comments you made where interesting. You said

> Modding the output of the hash function
> [without being a function of the message] will NEVER decrease the
> likelyhood of a collision. At best it only doesn't increase it.

But surly using the sum of the message is using a function of the it,
in a very basic sense.
Received on Fri Dec 23 20:11:22 2005