Folding keys etc...

[Nick Pelling <incoming@xxxxxxxxxxxxxxxxx>]

Hi Steve,

I've had a go at making folding keys, as per your website...

	http://www.diac.com/~ekwall2/voynich/

.....so as to understand how the basic mechanism would translate into purely 
cryptographic terms.

As I understand it, each gallows character (plus their four c--c 
struck-through versions) has its own separate translation table for the 
characters that follow on from it, which you believe are encoded 
left-to-right, then top-to-bottom.

So: the decoding algorithm would then look like:-

(1) if the first character on a page (or paragraph?) isn't a gallows character
	then revert to a pre-agreed translation set   [ie set a default set]
(2) read in an input symbol (left-to-right, then top-to-bottom)
(3) translate it to an output symbol using to the current translation set
(4) Note: these can map an input symbol to either single or multiple 
characters.
(5) if that input symbol also happened to be a gallows character
	then switch to a different translation set  [ie refold your folding key]
(6) loop to step #2

4 gallows characters plus 4 c--c'ed gallows characters ==> 8 translation 
sets... but we don't have any, so will need to guess/infer/predict them.

Have I understood this correctly? And do believe there were many folding 
keys or just one?

If so: the obvious test for this would be to examine the statistics for 
each block of text following each of the 8 gallows characters up to (and 
including!) the next gallows character along.

If these 8 sets of statistics show similar frequency distributions, we 
could build arbitrary mappings between them, to try to normalise them into 
a single unified alphabet: and see where that leads.

.....or has this kind of code already been analysed to death in the 1940s? :-/

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