From: Jim Gillogly <jim@acm.org>
Date: Wed, 21 Oct 1998 02:26:07 +0800
To: cypherpunks@cyberpass.net
Subject: Re: Homophonic substitution [could you help.....]
Message-ID: <362CCCAC.94253462@acm.org>
MIME-Version: 1.0
Content-Type: text/plain



Bernardo B. Terrado writes:
> It's about Homophonic  substitution and 
> map E to 17,19,23,47,64
> map A to 8,20,25,49
> map R to 1,29,65
> map T to 16,31,85
> but otherwise the ith letter maps to the 3ith letter
> MANY A SLIP TWIXT THE CUP AND THE LIP
> will become
> 3608397220543324451666246931852117066045253909162147332445
>
> My question is what/how did they represent the other letters like L (etc.)
> coz I've tried to map them and yet I still can't understand
> I even wrote A to Z then map them to 1 to 99, I still can't figure it out.

That's the "otherwise" rule right after the four "maps".  If the
2-digit number is divisible by 3 (like the first "36"), divide
it by 3 and count that many letters through the alphabet starting
with A.  The 12th letter (0-origin) is M, so 36 corresponds to M
and L corresponds to 33.

That particular example doesn't use the ciphertext space very efficiently:
only 41% of the available 2-digit numbers are used.  If you must use a
homophonic, I'd suggest a 100-letter pangram, which gives a reasonable
distribution of letters, full coverage of the alphabet, and some chance
of remembering the thing without carry incriminating documents.

-- 
	Jim Gillogly
	Sterday, 29 Winterfilth S.R. 1998, 17:38
	12.19.5.11.2, 13 Ik 15 Yax, Sixth Lord of Night