06.01.2015 Views

Cryptology - Unofficial St. Mary's College of California Web Site

Cryptology - Unofficial St. Mary's College of California Web Site

Cryptology - Unofficial St. Mary's College of California Web Site

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

100 CHAPTER 6. DECRYPTING MONOALPHABETIC CIPHERS<br />

the same letter. That is the goal <strong>of</strong> our next chapter. We end this chapter with a<br />

couple <strong>of</strong> ways that can make monoalphabetic ciphers harder to decrypt. (These<br />

tricks can actually be used with most <strong>of</strong> the ciphers we will see.)<br />

Homophones. When a substitution alphabet has multiple substitutions for a<br />

given letter these substitutes are called homophones. For instance, we may<br />

decide that every other e will be replaced by a z before the message is enciphered.<br />

Since z is so uncommon, our partner should be able to figure out we’ve replaced<br />

some e’s by z’s without prior warning. But to the enemy who intercepts the<br />

ciphertext, the tall 12% peak <strong>of</strong> e’s will be split into two far less visible 6%<br />

bumps. This would make the frequency analysis <strong>of</strong> our adversary a bit harder.<br />

Homophones are more commonly used when one is sending the ciphertext in<br />

numerical form. For example, consider the replacement list in a cipher <strong>of</strong> Henri<br />

IV <strong>of</strong> Navarre <strong>of</strong> France, c. 1600. [Pratt, page 64.]<br />

plaintext a b c d e f g h i l n o p r s t u w x y z<br />

ciphertextA 31 26 27 28 31 29 3 33 12 14 44 15 16 17 9 20 21 22 23 24 25<br />

ciphertextB 34 35 36 37 38 39 30 41 42 43 67 18 46 47 19 50 51 52 76 54 55<br />

ciphertextC 37 59 60 61 62 40 64 65 66 85 45 69 70 49 73 74 75 77 78<br />

ciphertextD 80 81 82 68 83 72 84 86 87<br />

(Most <strong>of</strong> the unassigned numbers stood for common words: 10 = le, 39 = mon.)<br />

To encipher a letter choose one <strong>of</strong> the numbers beneath it. So tres might<br />

be enciphered as 20-17-31-9 or as 50-70-38-49. These multiple substitutes<br />

flatten the frequency chart, making it much harder for our adversary to decide<br />

which number(s) represent which letter.<br />

To make this effective we would probably want several homophones for each<br />

letter, and then somehow force ourselves to pick the homophones at random.<br />

We might have six homophones for each letter and then to encipher a letter<br />

we would roll a die and if the die comes up with a 4, then pick the fourth<br />

homophone as the cipherletter. Unfortunately, homophones make enciphering<br />

and deciphering much slower.<br />

Further, if the message is thousands <strong>of</strong> letters long frequency analysis will<br />

still win out. A handwritten page will have approximately 500 characters on it<br />

(about 25 characters per line and 20 lines per page) and a typewritten page can<br />

easily consist <strong>of</strong> 3000 characters (70 characters per line and 45 lines per page).<br />

For our die-homophone example, the cipher-numerals standing for uvwxyz still<br />

would very seldom be used, while those standing for e would be popular, and<br />

<strong>of</strong>ten occur paired with those standing for t and h. So while homophones are<br />

helpful, they cannot make monoalphabetic ciphers secure.<br />

Nulls. Extra meaningless symbols that added to a text only to confuse the<br />

enemy analysts are known as nulls. One could spread a number <strong>of</strong> unpopular<br />

letters randomly throughout the plaintext. Itz yisx xnot jtooquk qdifficult<br />

wtowq jread ax meksskage cbontgainuing nzullys, but it is harder to cryptanalize<br />

it. However, it is a bother to add (and later, remove) such letters, and they make

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!