[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: VMs: Sukhotin's Algorithm

In addition :-)
if I remember correctly, hidden Markov processes 
can be applied rather well to identifying 
vowels and consonants.
Notably, they don't work for the VMs. There, 
instead, this technique highlights the standard
word patterns, where the appearance of certain
characters depends havily on the 'position in
the word'. 

I am speaking from memory, and the analysis in
question was not done by me, but a.o. by Jim Reeds
and the late Denis Mardle. 

This tendency of characters to appear in a 
certain order is, in my opinion, what causes
the low digraph entropy, and not the other way
around. I really don't know how to explain that
better, and by itself it is probably an untenable
statement. I just mean that the tendency of
character pairs to be unusually restricted in
the VMs is really just part of the scheme. There
is something else going on which causes it as
a side-effect, so to speak.

Cheers, Rene

--- Bruce Grant <bgrant@xxxxxxxxxxxxx> wrote:
> In addition to Jacque's article in Cryptologia,
> which explains the
> Sukhotin algorithm nicely, I ran across a couple
> other interesting
> articles there about other approaches to detecting
> vowels:
> 1.    An article from vol. 8 no. 1 (1984) by Roland
> Anderson, Finding
> Vowels in Simple Substitution Ciphers by Computer,
> uses a technique based
> on plotting the frequencies of interval lengths
> between occurrences of the
> same letter or group of letters. The idea is that
> such a curve will
> decrease smoothly as the interval sizes get larger
> for consonants, but
> will have a peak for vowels. (There was a second
> article, in 1986 I think,
> elaborating this approach.)
> 2.    An article by Caxton Foster (didn't write down
> the date) tries to
> identify vowels based on the claim that they tend to
> be more evenly
> distributed than consonants when a message is
> chopped into equal-sized
> blocks. (This is a modification of a technique used
> to guess the likely
> rectangle dimensions for a transposition cipher
> according to Foster.)
> 3.    Another article by Foster in vol. 16 no. 3
> compares the preceding
> method, Sukhotin's algorithm and a third method
> described in Helen Gaines
> _Cryptanalysis_ based on identifying as consonants
> letters which have few
> neighbors, then identifying as vowels letters which
> appear close to and
> evenly distributed around these "consonants" - he
> calls it the "consonant
> line" approach.
> Bruce
> To unsubscribe, send mail to majordomo@xxxxxxxxxxx
> with a body saying:
> unsubscribe vms-list

Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list