Codes are a completely different form of cryptography than the ciphers previously discussed. Unlike ciphers, plaintext is not 'encoded' in the traditional sense. Instead, plaintext is broken into phrases that can be expressed as predetermined combinations of numbers or letters. The phrases and their equivalent code groups are listed in large dictionary type codebooks. Phrases and words are randomly assigned (generally) four or five characters.

The Hudson Code, in use by American forces during part of 1918, is a fairly standard of a code system. In the Hudson codebook, '1655' meant 'enemy machine gun fire,' '5639' meant 'sub,' and '1673' meant troops. For more popular terms, like 'tank,' two or more code groups were provided, in the 'tank' case, '3287' and '3408.' These additional code groups act like the variants of advanced monoalphabetic substitutions to dim the usefulness of frequency distributions that could be performed on code groups to identify common words like 'the' and 'and.' In some respects, code books are huge substitution ciphers with a ciphertext alphabet that has a unique symbol for every possible word and character.

NOTE: The Hudson Code specifies that '1685' means 'Question Mark'. The only reason I mention this is that the students who made this site almost called the site '1685' because we weren't completely pleased with our plain and drab title of "Cryptology: The Science of Concealment". We all thought that 1685 'sounded good'...but two of us also thought that it might be too confusing for a site that is supposed to make concepts clear (the third still thought/thinks that 1685 is a better title.) Just a little FYI from behind the scenes.

Using a code book system requires a large infrastructure and good deal of time. Communication is a major aspect of the military and smart enemies will monitor radio traffic by opposing forces. Analyzing enough messages sent in code will eventually lead to partial understanding of at least some of the common code groups. Perhaps an even greater risk is that the enemy will capture (or buy from a disgruntled officer) a code book. To help prevent the enemy from translating a code, the U.S. military employed new code books to its forces sometimes as often as just 3 days, and sometimes as many as 22 days.

Reissuing code books is no easy task, either. Some books contains over 30,000 phrases, words and letters and their equivalent codes. (Each code book had an encoding section where the phrases were listed alphabetically and a decoding section where the numbers were listed in order, so 30,000 entries actually means 60,000 for the code books). The code books were printed under the watch of several officers to prevent break-ins, and each book had its own ID so that it could be tracked. When new code books arrived, old one were returned and accounted for. During WWI, the American code book facility printed 80,000 books and pamphlets to secure the communications of America (imagine how many trees we're saving by using paperless digital cryptography.)

While the logistics of distributing and accounting for code books is a large job, creating the code book is a sizable challenge, too.