If you’re interested in the Zodiac Killer’s ciphers and you have some time to spare, you may well find this thesis on automated cipher solving an interesting read. Apart from the other valuable content, the author, Michael James Banks, provides a concise overview of the various cipher systems that comprise classical cryptography (see page 8), including the type of cipher preferred by the Zodiac Killer: the homophonic substitution cipher.
Of particular interest, the author discusses the ciphers of the Zodiac Killer in a section devoted to unsolved cryptograms. Furthermore, he briefly describes the results of using the described cryptanalysis software in an attempt to decipher the killer’s 408 (part 1, part 2 and part 3) and the 340 ciphers. Sadly, no solution was found. Of course, the fact that the tool was unable to solve the 340 likely surprises no one since all known attempts to solve the cipher have failed. However, multiple software programs, including zkdecrypto and CipherExplorer (an unreleased cryptanalysis tool written by yours truly), have been able to solve the 408 quickly. Mr. Banks explains that there likely was a problem in the section of code that handles homophones – i.e. different cipher symbols that represent the same letter. Time constraints prevented him from fully resolving the problem.
Having dealt with a number of similar issues during the development of CipherExplorer, I found the section on Fitness Functions for Cryptanalysis (page 20) especially valuable.
The “Part 2” link above points to Part 1. Here is the correct link:
Right you are! It’s fixed now. Thanks for pointing that out!