Cryptography Reference
In-Depth Information
This time, however, we will not simply select a shift,
s
, to serve as the key.
We will select a keyword, such as ABCD. We write this keyword repeatedly
over the plaintext:
ABCDABCDABCDABCDABCDABCD...
MEETINGTODAYEVENINGATTHE...
Next, we add superimposed pairs:
A+M=M
B+E=F
C+E=G
D+T=W
A+I=I
...
(Like in the example above, we have to think of letters as numbers: A
=
0,
B
=
1
,...,
Z
=
25). That's already the ciphertext. So, with this keyword of
length 4, we have defined four different Caesar additions, which we will use
cyclically. We can already see from the first few characters that the 'EE' in
'MEETING' become 'FG': patterns are generally destroyed. And unless you
know the length of the keyword, you can't tell which same plaintext characters
correspond to which same ciphertext characters.
This encryption method is called the
Vigenere cipher
, which is not entirely
correct, because Vigenere described a more general method in 1585: he took
an arbitrary substitution of the alphabet and shifted it cyclically. This, too, is
merely a special case of the general polyalphabetic encryption. But let's go
back to our example.
How do you break this Vigenere cipher? It is basically simple. Assume we know
the key length, which is 4 in the above example. We pick out the ciphertext
characters at positions 1, 5, 9, 13,
...
, i.e., each 4 characters apart. This subset
of the ciphertext is Caesar-encrypted since, at these positions, there is always
the same character above the plaintext line. We determine the frequencies of
all characters in this subset and assume that the most frequent character is 'e'.
That produces a shift. Similarly, we proceed with the subset formed from the
ciphertext characters at positions 2, 6, 10, 14,
...
We may be able to retrieve
the plaintext. If we don't, we have to play a little — we might want to guess
