Those using monoalphabetic substitution codes (seriously obsolete and insecure for at least the past 250 years) have been told in countless books, and by the American Cryptogram Association that the way to create a substitution key is to follow this procedure:

1. Start with a seed: "I am Captain Kirk".
2. Sanitize Seed.
   1. Remove spaces and non-alphabetic characters and convert to all UPPER CASE.
   2. Eliminate repeated letters.
3. Generate Key.
   1. Fill in cyphertext letters sequentially beginning with A, then B, etc.
   2. Fill in the remaining cyphertext letters in alphabetical order.

The shortcomings of this method are obvious. Plaintext letters early in the alphabet tend to be represented by common letters, and plaintext letters near the end of the alphabet tend to be represented by less common leters, often themselves (U - Z in this illustration). Not a pretty picture!

So here's an alternative.

1. Start with a seed "I am Captain Kirk".
2. Remove spaces and non-alphabetic characters and convert to all UPPER CASE.
3. The seed letters represent "jumps" per their alphabetic value:
   A = 1, B = 2, C = 3, D = 4, E = 5, F = 6, G = 7, H = 8, I = 9, ... , V = 22, W = 23, X = 24, Y = 25, Z = 26.
4. In our example, the first seed letter is "I", which has the value 9.
5. From the first letter of the alphabet, (A) go forward 9 letters to "J".
6. "a" is now represented by "J".
7. There are now only 25 plaintext letters left to be represented: b - z, and only 25 ciphertext letters left to represent them A - I and K - Z.
8. The second seed letter is "A", which means move forward one letter from "J" to "K".
9. "b" is now represented by "K".
10. Continue this procedure, repeating the use of the seed as needed until the entire alphabet is covered. At each step one plaintext letter will be eliminated as will one CIPHERTEXT letter. Eventually all 26 letters will be eliminated. The 26 letters of seed that will be used are: "IAMCAPTAINKIRKIAMCAPTAINKI".
11. The last (26^th) letter of the seed doesn't actually do anything, because by the time it is used only one plaintext and one ciphertext character remain, and they will inevitably be paired regardless of the identity of the last seed character.
12. The "Recycle Key" button takes the characters of the Ciphertext and inputs them, in reverse order, into the seed input, which can then be used to create a new Key.
13. There is at least one "weak key" pair when keys are recycled:
    Key = "A" and Key = "AZYXWVUTSRQPONMLKJIHGFEDCB".
    The second of the two results from recycling the first and remains the same when it is recycled itself.

Sometimes it is necessary or desirable to create a seed from a much longer string of text, commonly called a "Passphrase". The reason why this is a good idea is that people are very bad at thinking up random sequences. A long phrase or sentence, however, gives a lot of possibilities. The question is how to turn a passphrase into a seed, as per the previous section. Creating a small string of characters that contains much of the "randomness" of a long phrase is called "hashing". Proper hashing requires that at least the following conditions be met:

1. The hash has a standard length output that does not depend on the length of the input, or at least this is true given that the passphrase has at least some small minimum length.
2. Changing, adding, or subtracting one character of the input (passphrase) to the hash will cause a large and unpredictable change in most of the output characters.
3. There is no known method to reverse the hash to recover the passphrase.

[sp]abcdefghijklmnopqrstuvwxyz0123456789.,?'!/()&:;=+-_"$@*%[CR]