Information Technology Reference
In-Depth Information
new cipher alphabet would be the key needed to decode the mes
sage, as in the following example:
Plain Alphabet:
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
Cipher Alphabet:
R V I N T O Q F Z P A X H B K D C J W U M E Y G L S
Here, each letter in a message can be replaced by the corre
sponding letter in the cipher alphabet, so that “FUN” is encoded by
“OMB.” This approach resolves the first major deficiency in the
Caesar Cipher, because a knowledge of part of the code (O stands
for Y) does not appreciably help decipher other parts of the code.
The second deficiency is often addressed by following one of
two approaches. In the first approach, a different code is used for
each subsequent letter in the message. Thus, to encipher “FUN,” the
letter “F” is enciphered using one cipher alphabet, “U” using a sec
ond, and “N” using a third. Certainly, if this approach is used, with
a different coding scheme for each letter, and if the pattern of cod
ing schemes is changed for each message, then the messages may be
unbreakable. However, as a practical matter, using different cipher
alphabets for each letter in a message is unwieldy. Both the sender
and the receiver must agree on the sequence of codes to be used, and
management of many different codes can be difficult.
Thus, in practice, it is not uncommon to use one basic cipher al
phabet, but then to use different shifts for subsequent letters. For
example, the sender and receiver might agree that the first letter of
the message would be coded by the cipher alphabet, shifted by 3 let
ters (as in the original Caesar Cipher), the next letter shifted by 1
letter, the next by 4, then by 1, 5, 9, 2, 6, 5, 3, 5, and so on (where
this sequence of shifts may be remembered as the digits of the math
ematical number pi). To simplify the logistics further, the pattern of
shifts might be repeated after a certain number of letters. For exam
ple, once the first eight letters are coded by shifts of 3, 1, 4, 1, 5, 9,
2, 6, then the next eight letters also are coded by the same pattern
of shifts.
Codes produced in these ways are much better than the simple
Caesar Ciphers. The simplifications needed to manage their logis
tics (like shifting to the numbers in pi), however, also can open po
tential weaknesses that may be exploited by those trying to break
the code to obtain the underlying data. Statistical methods used for
single alphabet codes often may be extended to these multiple ci
pher approaches, at least when patterns repeat or when they can be
predicted.
