Re: State machine hypothesis...

[Nick Pelling <incoming@xxxxxxxxxxxxxxxxx>]

Todays thought on the SMH:

(1) having random polyalphabetic tables would give a random 
entropy-near-to-100% feel to it, which is very much a modern conception of 
ciphertext. But if the tables were ranked to give a low-entropy output (as 
I suggest), it would give a much more steganographic feel to it, which 
would be effective at concealing the text in a different way.

(2) as the characters following EVA f and p have very similar stats, it 
could be that they either code the same letter, or they ping-pong between 
similar letters in different polyalphabets.

(3) Has anyone noticed the relatively few numbers of repeated characters in 
the text? (daiin aside, which may be another issue, and ee) It struck me 
this morning that (space) might sometimes code for "same as last symbol": 
so, "pepper" would be "pep.er". Has this been suggested before? People 
often use pairs of letters to help find cribs: perhaps these have been 
deliberately obscured here? This might also explain "8am.8am" --> 
"8amm8am". :-/

Cheers, .....Nick Pelling.....