Talk:Autokey cipher

This article has no information on how the cypher was broken. &#9999; Sverdrup 02:28, 21 Nov 2004 (UTC)

So if the message is incorporated in the key, why do you need to decode the cyphertext? You've already got the message. {Answer: The key+message is just a temporary copy used by the sender. The recipient doesn't have it, only the original key and the cyphertext.}

{T}o answer your question: if you have cipher text that says WMPM MX XAF YHBRYOCA and you know the keyword of KILT, then you can decode the message as such:

CIPHERTEXT:_WMPM MX XAF YHBRYOCA

KEY:________KILT

{ Go across the top of the table to the key letter (K), search down to the cyphertext letter (W), then find the plain text letter at the left edge (M). After four iterations, you have: }

PLAINTEXT:__MEET

Now that some { = key length } of the plaintext is found, it is appended to the end of the key:

CIPHERTEXT:_WMPM MX XAF YHBRYOCA

KEY:________KILT ME ETA TTHEFOUN

PLAINTEXT:__MEET AT THE FOUNTAIN

so, the reciever of this message needs to have that original key, which a potential "eve" would not have. the reason to use this is that the methods to break the vignere cipher can't be applied beccause the key is as long as the message, so the length of the key can't be found by measuring the distance between repeated phrases.

--Wakingrufus 19:38, 10 December 2005 (UTC) {edit by gatmo for additional detail - feel free to rephrase}

"This text-autokey cipher was hailed as "le chiffre indéchiffrable", and was indeed undeciphered for over 200 years, until Charles Babbage discovered a means of breaking the cipher." isn't this what happened to the vigenere cipher? how was this cipher actually broken?--Wakingrufus 19:38, 10 December 2005 (UTC)

I've added an example of cryptanalysis. Don't know exactly how Babbage first cracked Autokey but probably along the same lines. Frankd 12:21, 12 June 2006 (UTC)

Frankd: Thanks. That example was very helpful in understanding the basic frame of mind a cryptanalyst has to be in to break a code.

Self-synchronizing
I am a little bit confused: is Vigenere autokey cipher self-synchronizing? --Necago 16:38, 9 June 2007 (UTC)


 * Yes, the Vigenere autokey cipher is a self-synchronizing stream cipher . How can we make this article less confusing? --DavidCary (talk) 15:32, 26 March 2016 (UTC)

Why does the article say, "In modern cryptography, self-synchronizing stream ciphers are autokey ciphers."? Shouldn't it read, "In modern cryptography, autokey ciphers are self-synchronizing stream ciphers."?

Also, in the Wikipedia article on stream ciphers, it states that self-synchronizing ciphers use several of the previous N ciphertext digits to compute the keystream. However, this article states that "An autokey cipher (also known as the autoclave cipher) is a cipher that incorporates the message (the plaintext) into the key." If an autokey cipher uses the plaintext and a self-synchronizing stream cipher uses the ciphertext, how can an autokey stream cipher be classified as a self-synchronizing steam cipher? Either the definition of self-synchronizing steam cipher needs to be changed to state that "self-synchronizing ciphers use several of the previous N ciphertext OR PLAINTEXT digits to compute the keystream" or the statement that autokey ciphers are self-synchronizing stream ciphers needs to be removed.

The issue seems to be that self-synchronizing is both the name of a class of ciphers and also the description of a cipher property. Autokey ciphers have the property of being self-synchronizing (able to self-recover after getting out of sync) but don't fit the Wikipedia description of a self-synchronizing cipher because they use plaintext rather than ciphertext. Maybe the statement should read, "Like ciphertext autokey (CTAK) ciphers, autokey ciphers have the advantage that the receiver will automatically resynchronize with the key if digits are dropped or added to the message stream.


 * Whoops, my mistake. The discussion at "Why is Vigenère Autokey self-synchronizing?" has convinced me that the traditional Vigenere autokey cipher with an n-character primer key does not self-recover after getting out of sync -- an error during transmission changing one ciphertext character will propagate during decoding into every n-th following plaintext letter. An error inserting or deleting a character is even worse. So it is not a self-synchronizing cipher. Sorry, Necago! --DavidCary (talk) 23:17, 25 February 2021 (UTC)

Brute-forcing the key
For all you cryptanalysts out there, you should know that if you have computer, and your given just the ciphertext, you can brute force the key by attacking each key letter individually. So let's say I've been given a long autokey cipher. Given a keylength, or a number out of a list of keylengths to try, each letter of the key can be found by deciphering each letter of the cipher that that key decrypts. I can try the letter 'a', see how close the frequency analysis of the decrypted characters is to normal frequency analysis, try 'b', see if the frequency analysis is closer than 'a', and so on. For a computer, this takes a few seconds, and is easier than the cryptanalysis shown in the article. Hope this helps any future programmer, Jessemv (talk) 00:26, 24 November 2009 (UTC)
 * This is likely original research unless someone else has had this idea in the past, I think there is perhaps a more efficient and reliable solution to the method of brute forcing an Autokey Cipher. Instead of doing frequency analysis comparisons on the plaintext that a particular key character decrypts to, do the frequency analysis over all the characters that have been found using the detected key. For example, if I try for a key length of 3, and for the first character I run through A-Z and find that 'c' makes the best plaintext frequency analysis, I then repeat this to find the next character in the key, but using the 'c' character as well. Eventually I might find that 'ca' makes the best frequency analysis, and the final character 't' makes it even better. Key = 'cat' Jessemv (talk) 00:36, 25 November 2009 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 2 one external links on Autokey cipher. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20131202225102/http://asecuritysite.com/security/Coding/autokey?word=attackatdawn,queenly to http://asecuritysite.com/security/Coding/autokey?word=attackatdawn%2Cqueenly
 * Added archive https://web.archive.org/web/20131203104209/http://asecuritysite.com/security/Coding/autokey?word=meetatthefountain%2Ckilt to http://asecuritysite.com/security/Coding/autokey?word=meetatthefountain%2Ckilt

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 05:49, 22 October 2016 (UTC)

Example is wrong or misleading
The article says that "A key-autokey cipher uses previous members of the keystream to determine the next element in the keystream" but from the example only the current byte of the autokey. Essentially it's a weak form of a book cypher where the key is (mostly) the plain text

Am I misunderstanding?