Re: SHA1 hash safety

From: C. Scott Ananian <cscott@cscott.net>
Date: Apr 18, 2005, 17:04 UTC
Message-ID: <Pine.LNX.4.61.0504181252590.1039@cag.csail.mit.edu>
In-Reply-To: <20050418074323.GA29765@hexapodia.org>

On Mon, 18 Apr 2005, Andy Isaacson wrote:

Show 12 quoted lines

> If you had actual evidence of a collision, I'd love to see it - even if
> it's just the equivalent of
> % md5 foo
> d3b07384d113edec49eaa6238ad5ff00 foo
> % md5 bar
> d3b07384d113edec49eaa6238ad5ff00 bar
> % cmp foo bar
> foo bar differ: byte 25, line 1
> %
>
> But in the absence of actual evidence, we have to assume (just based on
> the probabilities) that there was some error in your testing.

I've already had a long correspondence with this poster. He claims that "this happened 7 years ago", involved a "commercial contract covered by Swiss Banking Law" (with caps!) and that, of course, he "certainly doesn't retain [his] client's documents", and even if he *did*, he wouldn't show them to *me*.

And then he was unable to comprehend that I couldn't accept his word alone as prima facie evidence that the laws of probability did not apply to him or his clients.

I've been a coder far too long to attribute to "The Mysterious Hand Of God" what can adequately be described by subtle programmer error.

The most reasonable explanation, given the (lack of) evidence, is that 
the programmer involved quickly took refuge in a (wildly improbable, but 
his clients'll never know) "MD5 collision" instead of buckling down and 
finding the bug in his code.
  --scott

ODOATH Ortega FBI SGUAT AEBARMAN India Peking ODACID operation RYBAT 
[Hello to all my fans in domestic surveillance] for Dummies KUCLUB
                          ( http://cscott.net/ )