Hi all,
I've ever heard about a lecture taught in some universities about derandomization. What is derandomization exactly? is it "the opposite" of cryptography? I mean finding a rule in a "seemingly random" bits of information?
Posted on 2005-05-06 23:59:49 by Pinczakko
It's to do with how most random generators implement feedback.
Without discussing how we obtain a Seed value, imagine that via some obscure formula we transform the Seed value into a "random" output value, while also , via a generally much simpler formula, applying the resulting "random" output value to the Seed itself, so that the next "random" we generate is based on the previous input and output values according to straight math.

It stands to reason that if we presume to have cracked the algo the "randomgen" uses to manipulate its own Seed, then we have a much easier time cracking the "random" algo itself, and thus undoing the whole charade.

Since "random generators" are being applied in cryptography, people interested in that field are poking fun at them. The algorithms in standard random generators were never designed for security and the scrutiny and resulting criticism is unfair imho.
Posted on 2005-05-07 12:15:27 by Homer
Just a note - the RC4 "stream cypher" is an example of a randomgen being applied in cryptography.
Posted on 2005-05-07 12:18:47 by Homer