Jarlve, do you want me and glurk to move glurk’s message and my answer to the new thread, or should we finish that message here and move to the new thread? Because I just think that it is weird to start the new thread with discussion of a message that isn’t in your new suite of messages. But we can do that if glurk moves his message first. Whatever you guys want to do. I had to do some re-tooling of my spreadsheet, it only works with symbols numbered 1-63. Just had to make a conversion so that 792 = 1, etc., and that slowed me down a bit. I’ll post my best guess as to the cycles tonight in the designated thread. At quick glance, I see some perfect cycles, at least one with three symbols. Wow, Jarlve, that was a lot of work making that new suite of messages. That’s plenty of bones to chew on for a while. Thanks!!! Smokie.

Jarlve, do you want me and glurk to move glurk’s message and my answer to the new thread, or should we finish that message here and move to the new thread? Because I just think that it is weird to start the new thread with discussion of a message that isn’t in your new suite of messages. But we can do that if glurk moves his message first. Whatever you guys want to do. I had to do some re-tooling of my spreadsheet, it only works with symbols numbered 1-63. Just had to make a conversion so that 792 = 1, etc., and that slowed me down a bit. I’ll post my best guess as to the cycles tonight in the designated thread. At quick glance, I see some perfect cycles, at least one with three symbols. Wow, Jarlve, that was a lot of work making that new suite of messages. That’s plenty of bones to chew on for a while. Thanks!!! Smokie.

It’s all fine.

Does anyone actually have the source code to the "REMOVE_HOMOPHONES" program? I know it was written in Pascal, but I think that converting it to C (or anything better than Pascal) might not be too difficult, and might be useful.

I wrote my own version of the King/Bahler algorithm in Java, but I don’t think it will be very useful to you, because I made so many tweaks to it to account for the imperfectness of the cycles in the Zodiac ciphers, and to play with different ideas I had at the time. I’m not even sure that it still works.

Let me know if you still want to see it.

glurk,

Here are the results, color coded. It’s a little bit ugly, but hope you can see some of the cycles.

The yellow, red, dark brown and grey on the right are likely cycles. The light blue on the far left is a cycle, but short. I wonder if it is truly a cycle or not.

At the top, the crooked brown pencil mark shows several symbols with many different combinations of perfect cycling, but which cannot all cycle together perfectly. Analysis of that whole group is difficult and shows false perfect cycles.

Maybe 106 and 858; 357, 243 and 764; and 623, 758 and 784. I would have to really mess around with these much more extensively by hand, probably to not get anything resolved. Symbol 34 cycles into more than one other cycle. Maybe 34, 243 and 764 is a cycle.

So that is the problem with the system and also what I have found with the 340. There is a similar group of symbols in the 340 that have overlapping but high scoring cycles. It’s sort of comforting in a way, because I have suspected that Zodiac had some complicated cycle system that wasn’t a message at all. But the same thing happened here, so that is a benefit of having used my system on this particular message. I think that the symbol count has a lot to do with it, and also the shortness of the message. If the message was say, 400 symbols, then this would probably get sorted out better.

If you have any questions or want to see other cycles or parts of the spreadsheet, let me know.

Now I will make a clean break to the other thread. Thanks!

Smokie

Does anyone actually have the source code to the "REMOVE_HOMOPHONES" program? I know it was written in Pascal, but I think that converting it to C (or anything better than Pascal) might not be too difficult, and might be useful.

I wrote my own version of the King/Bahler algorithm in Java, but I don’t think it will be very useful to you, because I made so many tweaks to it to account for the imperfectness of the cycles in the Zodiac ciphers, and to play with different ideas I had at the time. I’m not even sure that it still works.

Let me know if you still want to see it.

Per glurk’s request, here is my (sloppy) implementation of REMOVE_HOMOPHONES:

http://zodiackillerciphers.com/HomophonesKingBahler.txt

The "testFindStrings2" method runs the algorithm against Z408. The bulk of the work is done in these methods: findStrings and reduceStrings. Here is the sample output produced for Z408:

http://zodiackillerciphers.com/Homophon … output.txt

Excellent input, did you test it on the 408? Or the 340?

QT