Cryptography Reference
In-Depth Information
another frequent character instead of 'e'. In practice, it is better to compare
the distribution of all characters in the ciphertext with the distribution in the
language; that almost certainly produces the shift.
And how do we determine the key length? By experimenting with different
lengths and analyzing the frequency distributions in the subsets mentioned! We
will see in Chapter 3 that there are much more reliable methods.
You can see that this approach is easy to program. I actually thought I'd found
such a program on the Internet:
solvevig.c
by Mark Riordan, written in 1991.
Unfortunately, when I had a closer look I found that the program merely tries
a given list of keys and that the cipher tests the character frequency to check
whether or not it might be English text. I find this incomprehensible. Apart
from the fact that there are other tests for revealing text (including digrams
or words, and many more), we know how it's done. Trying many keys is
really the very last resort in cryptanalysis. You are likely to experience a
similar disappointment with some cryptanalytic programs available for free.
Chapter 3 will discuss various possibilities of how to break Vigenere ciphers
in detail. Chapter 3 will also introduce a program you can find on our Web
site.
The cryptanalyst has to proceed more elaborately even with the general polyal-
phabetic substitution. Rather than doing Caesar ciphers, he does general sub-
stitutions. The method's principle remains unchanged: the substitutions are
applied one after the other to the single characters of the plaintext, and once
you have used up the last one, you start over again with the first. The number
of substitutions used is referred to as the
period
of the method.
We can think of a large variety of rules to be used to form each of the sub-
stitutions — they won't change the cryptanalysis much. In principle, it is not
much different from that of the Vigenere method, except that we have to find
the substitution for each subset. Knowing the rules by which the substitutions
are formed from one single keyword can strongly reduce the number of mis-
trials. We will come back to the cryptanalysis of polyalphabetic substitutions
in Chapter 3.
The designer of the algorithm defends himself against this attack by making a
huge number of possible substitutions available, i.e., enabling as long a period
as possible. The set of the ciphertext to be tested will (hopefully) not be enough
for a static analysis. Out of these considerations emerged the rotor machines and
particularly the Enigma. We will see this in a moment. First, let's summarize
the most important things:
