[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Letters coded by groups of symbols...?
Hi everyone,
I've been trying to determine in what way the VMS' low-entropy symbols
might group together to form something closer to a normal-entropy stream.
The underlying assumption is that the apparent impenetrability of the VMS'
code owes more to obfuscation than to its cipher structure - something like
Vigenere's "Alphabeth Northmanique" [1].
The following is only a preliminary analysis, but I believe it shows some
promise already. I should also point out that I assume that spaces in the
text are not necessarily coincident with plaintext word boundaries.
There appear to be three classes of gallows usage:-
(a) o + gallows
ot ok op of octh ockh ocph ocfh
(b) qo + gallows
qot qok qop qof qocth qockh qocph
(c) gallows everywhere else
The (c) gallows don't seem to fit any obvious pattern I've tried on them,
so my working hypothesis is that they are instances of some kind of
replacement code or cipher, under which they get replaced by a
(non-gallows) character (or character-sequence).
Then, most of the text forms itself into groups very naturally:-
al aiii ain aiin ar air am
ol oiii oir oiir or oir
ee eo ea ey es
ch cho <-- "normal picnic tables"
Sh Sho <-- "embellished picnic tables"
However, there are some frequent singletons which seem to have their own
rules:-
d y s m
And finally, there are some rarer singletons which don't quite fit the
pattern:-
q o
Here are my observations having tried this scheme out on eight pages (all
Language A botanicals):-
- The first 8-10 characters of each page often don't fit
- The last 1-3 characters of many lines often don't quite fit
- EVA "d" and EVA "y" seem to work as if they were Latin shorthand
- Many of the groups could well be duplicates of each other
- This gives an underlying alphabet size of between 36 and 43
- Doubled letters appear in many places
Cheers, .....Nick Pelling.....
[1] http://www.chass.utoronto.ca/~wulfric/rentexte/vigenere/chif_f338r1.jpg