@jay Sry took a bit to convert the video – here you go. It’s just one example out of many. Actually it is possible to define how many words in which sections shall be found as well as how long those word should actually be. 

In this example I defined for the first strings to find at least one word, for all the other strings one word each. The results were THOSE TOTAL STEEL TOOLS TOWNS. Other results may come up, of course.

The video shows the first result being found. Letting run the program for a while, dozens of such show up. Also you can see the first letter is defined for the + symbol, followed by others. The longer string with the double letter is from line 17/18 of the cipher.

The current ‘setting’ covers 181 symbols or 53% of the total cipher.

 

Regarding your question:

Yes I stick with the cipher to be a homophone substitution, like the Z408 (without transposition etc.). I do not comment on other ‘solutions’ or trials to find such as long as I can’t come up with a complete cleartext. Normally, I’d expect about 8-10 spelling errors in a homophone cipher of such length. Not many other encryption methods known/credible to cope with such a type of cipher, if at all. It’s a matter of logic: The Z340, as a homophone substitution, is way more complicated than the Z408. First of all, no need to use transposition/anagramming etc. Also, the type of cipher – more than 26 homophones – complies with such type of encryption (so did prove the 408). Also, transposition usually uses alphabetical letters. Imagine: Transposition, Anagramming, Line shifting, Homophones – in the end you can write into the cipher whatever you like to, as long as you keep on shifting/explaining. I had warned about such method as someone can indeed start writing words into the z340 cipher structure, then start to use transposition/line shifting/anagramming to complete the rest – mostly with short words. That’s what has been done and found (but lacks of longer words and logical text combinations such as ‘car light’, ‘blood drop’ etc.). Whatsoever, not my business.

QT

@quicktrader, sorry for not getting back to you earlier! Thanks for sharing your video and methodology. It’s always interesting to see what other cipher enthusiasts are up to.

I think you and I agree that there’s more to z340 (and maybe other ciphers) than has been discovered so far but I think we have different ideas about how z340 was made and how best to approach it. From my perspective, we have a very credible solution to the z340 already so any additional information I might find will have to incorporate the existing solution somehow.

“It’s a matter of logic: The Z340, as a homophone substitution, is way more complicated than the Z408. First of all, no need to use transposition/anagramming etc.”

Although a simpler cipher, the z408 can be cracked by automated solvers in a matter of seconds these days. It would be interesting to generate a large number of random homophonic substitution ciphers with the same characteristics as z340 and checking if they could also be cracked automatically (I suspect they could).

“Imagine: Transposition, Anagramming, Line shifting, Homophones – in the end you can write into the cipher whatever you like to, as long as you keep on shifting/explaining.”

I definitely agree that transposition is often abused by people who’re driven by wishful thinking, sensationalism or even book sales. However, I also think transposition can be a completely valid technique. To distinguish “good transposition” from “bad transposition”, I typically look for a simple and consistent pattern or algorithm. For example, I don’t think anagramming is very convincing whereas columnar or periodic transposition can be. It’s hard to put into words but I know it when I see it 🙂

Small update regarding the progress of the FCCP (homophonic substitution) cipher solver:

We do know that the 340 cipher shows up with

26^64 = 9.106685769537214956799814036095e+33 possibilites to configure all homophones. Needle in a – very large – haystack.

The initial set-up of

IoFBc
dNp
kdW

as well as the symbols

L, l, C, z

might be a possibility to ‘attack’ the cipher (technical issue: Oranchak’s great z typewriter links get removed automatically?).

Thus, by preselecting some most frequent n-grams, there now is a total of 65 quintgrams and 2×300 trigrams running in the program (besides the other symbols mentioned before).

In other words, the program simulates a total of 12 letters or

26^12 = 9.5428956661682176e+16

variations – all to be checked against a pretty large dictionary (~4,500). As we have to admit, this still is almost nothing compared to the full strength of the cipher.

 

Another thing which has been added/improved: From now on, the level of (basic) set-up is called ZEROH and from there on ALPHA, BETA etc. for each ‘success’ (=words found at specific locations in the cipher). Also, for each of those levels, the amount as well as the length of the words to be found can be defined accurately.

 

Interestingly the program is getting better in producing ‘readable’ phrases. Please still keep in mind that both, the program and also the method did not exist before. Of course it is still a long way to go – if possible at all – but here you go with another screenshot of what the program is actually capable of, so far.

 

QT

 

 

Another Update…

Attacks on such a cipher as the 340 are nice if you find multiple words. However this doesn’t help unless you know those words hadn’t occurred accidentially. Because of this, I tried to get longer (partial) strings from the cipher which is, knowingly, extremely difficult (one homophone different to the previous one etc..).

However, so far, I managed to set up a 38-letter-string (!) out of which a total of 21 letters are part of the solving process. The hard part of this was to find two strings where to be able to search not only for short words but words of length >5 at least. Two of such, in a string using regularly similar homophones, could lead to a hint. In addition, the 38-letter-string produced – after lots of thoughts how to do so – two “neighbor” strings which are on different locations of the cipher, however could give some confirmation if the 38-letter-string is correct or not.

The results of the program now looks like this with the previous letters as the intial “set-up”, then the 38-letter string (green part not searched for words yet) as well as the two neighbor strings in yellow and green again.

QT