The Pure Crypto Project's Hash Function

[John Kelsey]

At 08:13 PM 5/3/03 -0400, Rich Salz wrote:
>Very simple:  known to be cryptographically secure.  SHA-1 is good.  Your
>invention is bad.  End of discussion (from me).

Actually, SHA1 isn't known to be good, it's just strongly suspected to be 
good.  Other than information-theoretic stuff (e.g., one-time pads are 
really known to be good), most stuff in cryptography is presumed good 
because nobody knows how to break it, or even how to realistically come 
close to breaking it.)

Of course, that doesn't mean that rolling your own hash function is a good 
idea.  Or that it makes any sense at all to build all your own primitives 
in order to design some kind of secure system.  It's like deciding you want 
to design a better word processor than Word, and so starting by trying to 
design your own microprocessor architecture.

>         /r$

--John Kelsey, kelsey.j at ix.netcom.com
PGP: FA48 3237 9AD5 30AC EEDD  BBC8 2A80 6948 4CAA F259



---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com