Re: advice sought on key/data histogram analysis of rijndael/128 and serpent

"Bryan Olson" <fakeaddress@nowhere.org> wrote in message
news:Fpx6f.5209$BZ5.1493@newssvr13.news.prodigy.com...
> lkcl wrote:
> > and the graphs STILL show correlations:
> >
> > http://lkcl.net/crypto/rjd.1.512000.64x64.rangelimited.jpeg
> > http://lkcl.net/crypto/rjd.1.512000.32x32.jpeg
>
> What things are you claiming to be correlated? From what you've
> posted I cannot fathom your method, and the graphs with three
> unlabeled dimensions are useless.

What he's trying to prove is that Rijndael where the output is used as the
key shows correlations between outputs. If he's correct (and the information
so far does not show the claim) then this could be very interesting. I'm
attempting to regenerate his results using different code, so far my results
indicate that these are just statistical anomalies that are expected with
any random output, and as such are not interesting. That doesn't mean there
isn't an anomaly in there, just that it is almost certainly at below 1/2^32
probability (my largest test so far).
                Joe
Received on Mon Oct 24 02:08:05 2005