OK Glurk, so when you use the simple substitution cipher like this, you actually take the alphabet and scramble it and put them under the organized alphabet right?
Meaning every letter in the alphabet is only used once.

ABCDEFGHIJKLMNOPQRSTUVWXYZ <-alphabet
GTVFRCDEXSWZAQBYHNUMKIOPLJ <-key

So what if you chose to use part of a text as your key?
Then you would probably have several of the alphabet letters to have the same substitution letter as their key right?

Like this feks:

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
A L L G O O D C H I L D R E N G O E S T O H E A V E (the last letter in the sentence "N" is not used because one should only use 25 letters text like the alphabet right?)

So here "B" and "C" are both substituted by the same letter; "L"
and
"E" and "F" are both substituted by the same letter; "O"
and
"I" and "V" are both substituted by the same letter; "H"

So is this the way you do it if you want to use part of a text as a key?
And if that is correct what do you call this kind of Cipher/key?

Seems to me zkdecrypto is incapable of doing anything with my 405, which makes me wonder if it can do much at all with the 340, which has a higher multiplicity than the 408.

ZKDecrypto has solved a multitude of example ciphers with the same multiplicity as the 340. All of them, in fact. Yours, however, 405 in length with 99 uniques us far more difficult.

Try again. Make a 340 length cipher in English with the same multiplicity. Oranges and oranges, and Apples and apples and all.

-glurk

OK Glurk, so when you use the simple substitution cipher like this, you actually take the alphabet and scramble it and put them under the organized alphabet right?
Meaning every letter in the alphabet is only used once.

ABCDEFGHIJKLMNOPQRSTUVWXYZ <-alphabet
GTVFRCDEXSWZAQBYHNUMKIOPLJ <-key

So what if you chose to use part of a text as your key?
Then you would probably have several of the alphabet letters to have the same substitution letter as their key right?

Like this feks:

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
A L L G O O D C H I L D R E N G O E S T O H E A V E (the last letter in the sentence "N" is not used because one should only use 25 letters text like the alphabet right?)

So here "B" and "C" are both substituted by the same letter; "L"
and
"E" and "F" are both substituted by the same letter; "O"
and
"I" and "V" are both substituted by the same letter; "H"

So is this the way you do it if you want to use part of a text as a key?
And if that is correct what do you call this kind of Cipher/key?

There is a way pf doing that – it is called a Vigenere – please look here:

http://en.wikipedia.org/wiki/Vigen%C3%A8re_cipher

That was NOT used in Zodiac’s ciphers. The Wikipedia article is pretty good, and I just do not have time to explain it, although I do understand it completely. I’m doing my best to post here – people HATE me for some reason, I guess ciphers make people angry. I’ve never understood that. I think they are fun.

-glurk

If anyone takes the time to solve the 405 (which was never my goal, actually), please kindly PM me.

Thanks.

OK Glurk, so when you use the simple substitution cipher like this, you actually take the alphabet and scramble it and put them under the organized alphabet right?
Meaning every letter in the alphabet is only used once.

ABCDEFGHIJKLMNOPQRSTUVWXYZ <-alphabet
GTVFRCDEXSWZAQBYHNUMKIOPLJ <-key

So what if you chose to use part of a text as your key?
Then you would probably have several of the alphabet letters to have the same substitution letter as their key right?

Like this feks:

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
A L L G O O D C H I L D R E N G O E S T O H E A V E (the last letter in the sentence "N" is not used because one should only use 25 letters text like the alphabet right?)

So here "B" and "C" are both substituted by the same letter; "L"
and
"E" and "F" are both substituted by the same letter; "O"
and
"I" and "V" are both substituted by the same letter; "H"

So is this the way you do it if you want to use part of a text as a key?
And if that is correct what do you call this kind of Cipher/key?

There is a way pf doing that – it is called a Vigenere – please look here:

http://en.wikipedia.org/wiki/Vigen%C3%A8re_cipher

That was NOT used in Zodiac’s ciphers. The Wikipedia article is pretty good, and I just do not have time to explain it, although I do understand it completely. I’m doing my best to post here – people HATE me for some reason, I guess ciphers make people angry. I’ve never understood that. I think they are fun.

-glurk

Thanks Glurk, I´ll take a look at that

And please know that, as far as I’m concerned/ know, you are one of the most respected persons in the Zodiac community, so where you got that idea that people hate you I have absolutely NO idea

Seems to me zkdecrypto is incapable of doing anything with my 405, which makes me wonder if it can do much at all with the 340, which has a higher multiplicity than the 408.

One has to wonder if zkdecrypto was built to solve only the 408 and then kludged to barely handle the 340.

Either that or The Zodiac deliberately fubar’ed his masterpiece.

Do you understand how many man hours have been spent on these type of questions?

First off, zkdecrypto (and my own, unreleased, cipher-solving application CipherExplorer) were not built to solve the 408. They are built to solve homophonic substitution ciphers. Homophonic substitution ciphers exist on a continuum of cryptograph strength from simple substitution to one-symbol-per-symbol-instance (as glurk points out). The former is easy to solve, the latter is literally impossible.

This is precisely why I created the cipher generator; so I and others could easily create examples of known cryptographic strength in order to test cipher-solving applications and gain insight into their abilities.

The fact that you created a cipher that is difficult to solve simply because you didn’t know what you were doing is illustrative of nothing, although it does seem to destroy your original point of this thread.

Again, as glurk points out, a multitude of ciphers that are equivalent to the apparent strength of the 340 have been solved by zkdecrypto (and CipherExplorer, for that matter). This is why many of us concluded, circa 2007, that the 340 is not a straight-forward HSC, but rather a HSC with some additional, yet-unknown, complexity.

I wonder (to my self, probably) if Mathematica, MatLab, MathCad or Maple could handle the decryption of homophonic ciphers with multiplicity > 340?

Do any of them have a programmable cryptographic toolbox? Surely one could be built….

Don’t you think that if these applications were the key to solving the 340 that somebody would have done it? There have been masters theses written about the 340. There have been computer science classes taught with the intention of solving the 340. The result thus far: nothing.

You seem to neither understand nor appreciate zkdecrytpo. In terms of solving HSCs similar to the 340, it is the best available application and a very valuable asset to the Zodiac community. The inability of zkdecrypto to solve the 340 is not an indictment on zkdecrypto, but rather a clue into the nature of the problem.

Glurk

ABCDEFGHIJKLMNOPQRSTUVWXYZ <-alphabet
GTVFRCDEXSWZAQBYHNUMKIOPLJ <-key
@
$
%

In this case, each alphabet letter each has 1 substitute, except for "A" which has 4.
"A" = G,@,$,%

That’s called a homophonic substitution cipher. That’s what Zodiac used in the 408.

Thanks for the simplistic explanations Glurk.

My question Glurk: is there a rough breaking point of multiple substitutions for the average code breaking system
Eg zkdecrypto when they fail too decode and the whole thing turns to superfudge (poo).

Well this thread has not only not followed the original intent, its original joy has been thoroughly demolished (no surprise there, considering the human audience) by insecure people looking to make some kind of point, although I think the point is, contrary to their goal, that zkdecrypto is a rather limited program designed to handle very low multiplicity only. Is it even super-scalar? I doubt it. Looks to be at the bottom rung of Flynn’s taxonomy, AFAICT.

So I feel compelled to apologize to certain angry people if I failed to craft a cipher simple enough for the modest abilities of a certain program, when in reality ciphers are meant to maintain, from a third-party intermediary, plaintext secrecy, which was done here.

If someone possessed my key, they could decrypt the message therein.

So instead of a failure, I view this as a success. You cannot "crack" the cipher with your little program, even with a provided hint. I have thus proven the efficacy of basic homophonic ciphers with increased multiplcity.

I simply had no idea I was artificially constrained by zkdecrypto’s weaknesses. I also expected people to have more innate cryptographic brains than merely loading up a program and waiting for it to do all the work. Guess you need a "make plaintext" button to go with the "make art" button of digital artists. lol.

Bet Doranchak could solve it, if he wanted. That guy’s pretty smart. I always greatly enjoyed his thoughtful, educational and illuminating posts on Mike’s forum. I was there from 2008 and on. Learned about hill-climbing, multiplicity and genetic algorithms from him. Got the whole thing about The Zodiac doing "something" to the 340, rendering it unsolvable by simple automated tools. Yep. It’s not a straightforward homophonic cipher, blah blah blah. Some Indian or Chinese kid with merely pencil stub and paper will solve it one day while taking a dump on the john, like the Hardens solved the 408 with nothing more than a few intelligent assumptions and a lethargic, laconic weekend.

Reminds me of that old canard about a hundred-thousand monkeys slamming away on a hundred-thousand typewriters for a hundred-thousand years, eventually producing the entire works of Shakespeare, except in this case, sans the actual eventual "works". To whit, you cannot replicate brilliance with programmatic automation.

If I ever decide to make another handcrafted cipher, for fun, I will be sure, though, to construct one with an artificially-low cipher ceiling so as to render zkdecrypto capable of "cracking" it.

"Mankind’s psychological weaknesses are the single source of all its evil.".

jroberson-

Not only is your post insulting to myself and the other ZKD programmers, and the time we spent on it, it is also misleading and incorrect.

… I think the point is, contrary to their goal, that zkdecrypto is a rather limited program designed to handle very low multiplicity only.

It was not designed for "low multiplicity" only. Multiplicity is not an "artificial limit," it is a very real limit, above which ciphers become literally impossible to solve, lacking the key.

And as far as I know, ZKD remains the only publicly available interactive program to solve homophonic ciphers. Maybe there are others, but I’ve never found one.

If someone possessed my key, they could decrypt the message therein.

Assuming your cipher was correctly constructed (it probably is, don’t know,) this is true. If someone possessed my key, they could drive away with my car. That’s kind of a given.

I have thus proven the efficacy of basic homophonic ciphers with increased multiplcity.

Was this ever in doubt? Is this what you set out to prove? Seems contrary to your original post.

It’s easy to show, here is a cipher: $@*%#!
I could create a key to make those 6 symbols stand for any six-letter word. Without the key, it cannot be solved. Not by the NSA, Sherlock Holmes, or Einstein…

I simply had no idea I was artificially constrained by zkdecrypto’s weaknesses….If I ever decide to make another handcrafted cipher, for fun, I will be sure, though, to construct one with an artificially-low cipher ceiling so as to render zkdecrypto capable of "cracking" it.

See above, re: "Articficial Limits."

If you are a programmer, or know any programmers that can write a program that does the impossible, let us all know. Maybe we could solve the 340 then and all go home.

-glurk

My question Glurk: is there a rough breaking point of multiple substitutions for the average code breaking system. Eg zkdecrypto when they fail too decode and the whole thing turns to superfudge (poo).

There is a rough breaking point. And it IS rough. It’s a gray area. See this chart:

Several ciphers are indicated in the chart. Z1 is Zodiac’s solved 408 cipher. Z2 is the unsolved 340. The round red dot is jroberson’s cipher from this thread. As a general rule, anything above the dashed line is effectively unsolvable without the key, hints or other outside information.

-glurk

EDIT: I did not create the actual chart itself, but edited it. Credit goes to cryptographer John King, who – as far as I am aware – also coined the term "Multiplicity."

It’s easy to show, here is a cipher: $@*%#!
I could create a key to make those 6 symbols stand for any six-letter word. Without the key, it cannot be solved. Not by the NSA, Sherlock Holmes, or Einstein…

Could you make it a four letter word? I could surely solve that one.

Instead of stirring eachother up we could discuss something more fruitful. Like bringing forward ideas on how to improve our solvers ability to tackle higher multiplicity ciphers.

One straightforward idea is to include even longer ngrams together with a larger corpus. Another idea is to include another measurement that somehow is able to differentiate further between gibberish and meaningful text.

Jarlve-

I agree with you 100%. I don’t know why cipher discussions turn into contention and animosity, but they often do. And I’ve been around the "Zodiac Community" since 2006 or so, and have seen it happen time and again. Makes me sad.

The only possible idea I have is for there to be a semi-private "crypto-only" section of the board to work on these things, but I don’t think there would be much activity, to be honest.

But I agree with you, and I personally apologize to everyone here for any animosity I have caused, which was never my intent. I just felt the need to defend my work, after spending 7 or so years on it.

I try to help on the crypto end. It’s not for ‘everybody’ and I know that. I’m going to stick around, regardless.

-glurk

Hey guys… this thread is indeed very interesting, even though swords have been crossed by and by.

That chart you uploaded is a great help to the understanding of us cyber-dunces, glurk. Thanks. Same goes for the simple explanations/reminders of the different cipher/encoding types.

I have done a 408 for you guys to look at. The multiplicity is lower than 0.17, so it’s near the "difficulty borderline" on the aforementioned chart. More difficult than 408/340, but less so than jrob’s.

And for a reason I thought logical (to me anyway), I have also altered the character/symbol frequency complared to Z’s 408. (Sorry if I’m not using the correct terminology here). Like I say, I have a reason, so it’s not unreasonable, i think. And I haven’t assigned outlandish numbers of symbols to each character.

Like Z, I have split it into 3 parts, and am not letting on about the correct order. But will do if you so require. There may be a bit of of Z-type spelling/phaseology here and there.

So, here’s my 3-parted 408 winging it’s way to the Glurk Chronicle, Jarive Times Herald and Doranchak Daily News…

q%2@FMGH&WK5Qh/%+
aWL4lc2JbH&#sTWT~
XoB9zxQ0!H<6dlEqR
FJ5oUYw4p6KUbam@7
kmOsxnHAaedd4@QW9
*2poTln%+NQXFcLbU
EO/FX+~qH&%Q¬5AOs
pDB>*W<K4eH9!^2CW

cR^QX</Fqr2d6Ub#D
=+7*U^n9z@WABp+X9
S&!hU#<bQ¬Ob&GDw7
xa0n4ClE/MU+9~oqQ
MPWpp5TOoUYbAnwHD
GTr^E/WlQJq%mN<#O
¬EzTG¬dOxU0FwQ065
bTk&Y*o2O+K4noxEY

hUQp@ao>c4R+%9H7s
xhQ^SMPEUKDl5Id<q
W%9O4JJA#T@!<zk&O
7/~21GH2zqoMH7#I¬
^FAGe*TX¬hRcUrnEa
xDlewa6W0X<zSW^P+
KS25OT&Ms&L¬EOJQr
K%0>T5c%s¬%2+FoHY

I (using ZKD) have already solved this.

THIS IS THE END OF THE BEGIN

Too easy. Amazing what my lousy kludge of a program can do, I guess.

(Sorry, jroberson, you got it wrong)

BE GOOD NOW.

-glurk

EDIT: I shouldn’t post the complete solution, should I?