[Cryptography] Now it's personal -- Belgian cryptographer MITM'd by GCHQ/NSA

[John Kelsey]

On Feb 2, 2014, at 11:41 PM, Richard Outerbridge <outer at sympatico.ca> wrote:
...
> In retrospect it seems that certain aspects of the then au currant NIST
> statistical randomness tests were badly flawed at the time that AES candidates
> were being evaluated.  Don’t know that it would have made much difference,
> just noting that they were badly skewed.  How could this have happened?

What was the flaw?  I remember the NIST statistics guys ran a bunch of tests that didn't make sense on the candidates (like encrypting random plaintexts and then doing statistics on the ciphertexts).  Was there some other problem alongside that?  That sure looked like a bunch of people doing statistics on stuff they didn't understand. and I don't think it had any impact, as I remember everyone calling it out as soon as the report came out.  

> __outer
> 
> /ps co-submitter DEAL.

--John