1997-01-14 - RE: encryption program

Header Data

From: ed.falk@Eng.Sun.COM (Ed Falk)
To: Mullen.Patrick@mail.ndhm.gtegsc.com
Message Hash: c4f997d97562fd04a24672341d4a45112a6bbcf7b27d2471bd71d3801563cd9e
Message ID: <199701141808.KAA25979@peregrine.eng.sun.com>
Reply To: N/A
UTC Datetime: 1997-01-14 18:13:51 UTC
Raw Date: Tue, 14 Jan 1997 10:13:51 -0800 (PST)

Raw message

From: ed.falk@Eng.Sun.COM (Ed Falk)
Date: Tue, 14 Jan 1997 10:13:51 -0800 (PST)
To: Mullen.Patrick@mail.ndhm.gtegsc.com
Subject: RE: encryption program
Message-ID: <199701141808.KAA25979@peregrine.eng.sun.com>
MIME-Version: 1.0
Content-Type: text/plain


> What is a Kappa test?

Kappa test is described in Kahn's "Codebreakers".

Basicly, you take two streams of random text and line them up one over
the other.  Look for letter matches between top and bottom.  The percentage
of matches is called Kappa.

If the text is truly random, you should see matches 1/26 of the time.
However, if the text is english, you should see matches about 6.6% of
the time instead.

Likewise, if the text is encrypted with a monoalphabetic cipher, you
should *also* see matches 6.6% of the time.

Likewise, if the text comes from two different messages encrypted with
the same polyalphabetic cipher (and they're lined up properly) you
again see Kappa = 6.6%.

Finally, if the key repeats, and you've guessed the length of the
repitition correctly, and you line up the ciphertext with itself
accordingly, Kappa will again be 6.6%.


>  Is there software I can run to perform such a test?

I've been writing my own as I go along; it's more educational to do
it that way.  As been mentioned earlier in this list, there's an
archive of crypto software at

	ftp://ftp.ox.ac.uk/pub/crypto/cryptanalysis

I've been getting ideas from there, but it's more fun to write your own.





Thread