Various attempts have been made to improve the monoalphabetic substitution cipher:
The finest example of this was used by King Louis XIV of France, who used a very complicated variation on the system, by having not one symbol for each letter, but one symbol for each syllable (and a few for individual letters), and having certain symbols that had the meaning 'delete the last symbol'. This was quite sufficient to keep his messages hidden from cryptanalysis current at the time, but was successfully deciphered a couple of hundred years later, still using frequency analysis. It may have kept secrets hidden for 200 years, but if a man could work it out in 1900, then a modern computer could solve it quick enough for it to be insufficient to guarantee secrecy if we still used it nowadays.
Invented thanks to a few people, and credited finally to Vigenère, a polyalphabetic substitution cipher was invented in the 16th century. This works essentially the same way as the previous system, but uses the following square:
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
1 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 A
2 C D E F G H I J K L M N O P Q R S T U V W X Y Z A B
3 D E F G H I J K L M N O P Q R S T U V W X Y Z A B C
4 E F G H I J K L M N O P Q R S T U V W X Y Z A B C D
5 F G H I J K L M N O P Q R S T U V W X Y Z A B C D E
6 G H I J K L M N O P Q R S T U V W X Y Z A B C D E F
7 H I J K L M N O P Q R S T U V W X Y Z A B C D E F G
8 I J K L M N O P Q R S T U V W X Y Z A B C D E F G H
9 J K L M N O P Q R S T U V W X Y Z A B C D E F G H I
10 K L M N O P Q R S T U V W X Y Z A B C D E F G H I J
11 L M N O P Q R S T U V W X Y Z A B C D E F G H I J K
12 M N O P Q R S T U V W X Y Z A B C D E F G H I J K L
13 N O P Q R S T U V W X Y Z A B C D E F G H I J K L M
14 O P Q R S T U V W X Y Z A B C D E F G H I J K L M N
15 P Q R S T U V W X Y Z A B C D E F G H I J K L M N O
16 Q R S T U V W X Y Z A B C D E F G H I J K L M N O P
17 R S T U V W X Y Z A B C D E F G H I J K L M N O P Q
18 S T U V W X Y Z A B C D E F G H I J K L M N O P Q R
19 T U V W X Y Z A B C D E F G H I J K L M N O P Q R S
20 U V W X Y Z A B C D E F G H I J K L M N O P Q R S T
21 V W X Y Z A B C D E F G H I J K L M N O P Q R S T U
22 W X Y Z A B C D E F G H I J K L M N O P Q R S T U V
23 X Y Z A B C D E F G H I J K L M N O P Q R S T U V W
24 Y Z A B C D E F G H I J K L M N O P Q R S T U V W X
25 Z 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
26 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
To code a phrase, you use a different line for each letter in turn. To choose the line to use, choose a keyword. For example:
Keyphrase = CARLING; phrase to be sent = "meet at the pub at eight o'clock" CARLINGCARLINGCARLINGCARL meetatthepubateightoclock
To encode the first m, look at the letter above it - C. Use the 'C' row (the second row) of the square to code the m - so m=O. To encode the next letter, e, use the 'A' row (the last row) since A is above e. So e=E. Continuing to do this gives:
CARLINGCARLINGCARLINGCARL meetatthepubateightoclock OEVEIGZJEGFJNZGIXSBBINOTV
Suddenly the frequency analysis that was used before seems to be useless - t has been coded as E, G, Z, Z, and B, depending on which letter was above it at the time. Picking the most common code-letter and assigning it to be e as was done before is impossible, since the most common code-letters are E and G - neither of these can represent e since they both represent three different letters.
This code was pronounced invincible, but it was later broken. The addition of a clever initial step meant that frequency analysis could once again be used to break the code.
