Expository Essay – The Chaocipher #2

This was an interesting paper to read because it covered a cipher that is not as well known as others. Additionally, it is extremely relevant, having only been fully explained within the last year.
The explanation of how the cipher’s alphabets and encryption worked was very thorough. This allowed me to understand it without having to go back and reread a sentence a dozen times, making the paper a smooth read.
Though this may not be known, I’m curious about why the cipher was not widely used. Were there flaws to it? Was it perceived as inferior to other encryption techniques of the time?
This is yet another example where the inventor of the encryption technique, John Byrne, went without recognition for his accomplishment. The inventors of public key encryption were not attributed as inventors until decades afterward. This paper confirms the idea that cryptographers do not always receive the recognition they deserve.
Good job!

My name’s Moshe Rubin, and I’ve been involved in the recent disclosure of Chaocipher. I’ll try to answer Aubrey’s and Max’s questions.

Aubrey: I’m curious about why the cipher was not widely used. Were there flaws to it? Was it perceived as inferior to other encryption techniques of the time?

As implemented mechanically back in the early 20th century (remember, there were no computers then ) Chaocipher was susceptible to garbles and transmission errors, resulting in a potentially very high rate of error propagation (because the keying sequence feeds off of the previous plain- and ciphertext letters). This meant that, if either the sender or the receiver dropped, added, or modified even a single letter, or if a letter was changed during transmission, this would prevent the legitimate receiver from decrypted the message from that point onwards (we know that the demonstration Byrne gave to William F. Friedman in 1922 was plagued with just such mechanical errors and corruptions). The danger was (and still is) that the sender might correct the error and retransmit the message, thus providing enemy cryptanalysts with two nearly identical underlying plaintexts, which is a goldmine of entry points to solve the messages. One of the cornerstones of selecting a cipher system is that a corrupted plaintext or ciphertext letter should not prevent decryption.

In 1942, Byrne reapproached William F. Friedman regarding Chaocipher. While Friedman correctly asked him to submit a regulation number of cipher messages, Byrne refused to do so, thus preventing his system from being professionally evaluated.

Aubrey: This is yet another example where the inventor of the encryption technique, John Byrne, went without recognition for his accomplishment.

The reason Byrne was never recognized for this Chaocipher is because he never disclosed the general method of operation. Kerckhoffs’s_Principle (http://en.wikipedia.org/wiki/Kerckhoffs's_Principle) has been a well-accepted axiom for over a hundred years that the security of a crypto-system cannot lie in the system being kept secret (aka “security by obscurity”), but only in the secrecy of its key. Byrne never disclosed his general system to the public (he did disclose it to military persons, but they were prohibited from publishing the method).

Until Chaocipher was disclosed in July 2010 the method was unknown to the public. Thus, he could not have been recognized by the public. It is hoped that his invention will now be well analyzed and evaluated.

Max: I was surprised to learn that this cipher remained unbroken until this year, whereas ciphers that seem much more complex (i.e. the Enigma) were broken relatively quickly.

Again, the inner workings of Enigma were basically known to foreign governments (although the Germans introduced complications and modifications, many which required espionage to discover). Thus, the Poles and the British could begin to devise methods to solve Enigma messages because they know how the Enigma worked.

In the case of Chaocipher, the system couldn’t be cracked until the general system was known! Numerous non-governmental amateur cryptanalysts tried unsuccessfully to deduce the general method. Now that the general system is known we are in a good position to compare Chaocipher to other WWII systems.

Max: You mentioned that the cipher was broken in 2010 when Byrne’s daughter in law donated Byrne’s notes to the National Cryptological Museum, but I was confused as to who was responsible for actually breaking the code. Was it the museum or an individual?

If by breaking the cipher you mean deducing the general system and deciphering the hidden plaintext messages in Byrne’s exhibits, then Chaocipher was never broken. I played a role in Pat Byrne’s magnanimous decision to donate all the Chaocipher papers to the National Cryptologic Museum. In return I was send scans of some of Byrne’s Chaocipher enciphering worksheets. From them I was able to deduce how Chaocipher worked, and I published a paper in July 2010 presenting the general system on the web. So no, Chaocipher was not broken, it was disclosed.

In my opinion, the chances of anyone deducing the general system from what amounts to one, long plaintext/ciphertext exhibit was near zero. To break almost any complex machine system you need hundreds, if not thousands, of encrypted messages. The cryptanalysts them looks for patterns between messages, juxtaposing and sliding one against the other to look for an “in”. Given only one exhibit, there simply wasn’t enough material to deduce the general system.

Max: I was also a little confused about what you meant by the Chaocipher “making it a quick way to encipher plaintext in a secure fashion for those who lack the supercomputers necessary for more complex, modern ciphers.”

I believe Tanner’s meaning is that one could conceivably use Chaocipher today and get good security (as far as we know today) without needing expensive hardware and software. All one needs is a simple computer or programmable calculator and a program (Chaocipher can be coded in 10-20 lines of code).

Max: I found it interesting that Byrne, like so many others, was motivated by money, and resorted to the promise of money to publicize his cipher.

Byrne would have been happy to make money from Chaocipher (who would) but that was not the driving factor. he was even willing to license it to the US military for $1.00 (one US dollar) a year, so long as he would retain the rights to market it commercially. Money does not seem to have been the driving force. He did offer a $5,000 reward to anyone that cracked Chaocipher back in 1953, but if anything Byrne would have had to shell out the money, not collect it .

I hope that answers some of the questions you both posted.

Moshe Rubin
moshe.rubin@gmail.com