1994-07-04 - MD5 is 1=>1?

Header Data

From: Roger Bryner <bryner@atlas.chem.utah.edu>
To: N/A
Message Hash: bd1564761e6650285d74483f5dc62b87d31c03c785234ba154490096df804ce8
Message ID: <Pine.3.89.9407041124.A6205-0100000@atlas.chem.utah.edu>
Reply To: <199407041555.IAA01229@jobe.shell.portal.com>
UTC Datetime: 1994-07-04 17:03:19 UTC
Raw Date: Mon, 4 Jul 94 10:03:19 PDT

Raw message

From: Roger Bryner <bryner@atlas.chem.utah.edu>
Date: Mon, 4 Jul 94 10:03:19 PDT
Subject: MD5 is 1=>1?
In-Reply-To: <199407041555.IAA01229@jobe.shell.portal.com>
Message-ID: <Pine.3.89.9407041124.A6205-0100000@atlas.chem.utah.edu>
MIME-Version: 1.0
Content-Type: text/plain

On Mon, 4 Jul 1994 nobody@shell.portal.com wrote:
> Now if I have entered a phrase with EXACTLY 128 bits of entropy, 
> hypothetically, is that enough to have flushed the padding or 
> whatever out of the pipeline?
I have had this question also, has it been shown that the transformation 
of 128bit words through md5 is *theoretically* invertable, as if it is 
not, iterating it 1024 times could actually make you *LOOSE* entropy.
(say it was a random transformation, it would not contain each of the 128 
bit outputs, ie some inputs would map to the same output.)

I am not aware of any such result.