[OTR-dev] Forward secrecy/deniability for long messages with low overhead

[Sergio Lerner]

Ileana : Thanks for the paper.

On 22/02/2013 06:05 p.m., Ileana wrote:
> Your argument to use sha-512 versus sha1 would in theory be better for
> your application...and certainly altering successive inputs to the hash
> (which would certainly increase the entropy), but without a mathematical
> analysis of your application, I would not use it myself.  encipherment
> with xor to plaintext assumes one side of the xor is statistically
> random, and that finding a pattern is a difficult problem.
>
> You need to demonstrate mathematically that the input to the
> sha(whatever) function does not create a predetermination or reduced
> domain of the output either directly or showing it equivalent to
> existing derivative (i.e.  you show us a paper!)
>
Well, the idea came to my mind while I was typing yesterday, so it's 1
day old. I do not type that fast.
I promise I will write a paper, which I will finish in one or two years,
when my little kid becomes 4 years old and I have more free time. :-)
Maybe you would like to co-author the paper with me...