I have gotten tired of making telescopes, at least for awhile, and looking about for something to exercise my mind I stumbled across cryptography. The old stuff, pencil and paper, not secure internet correspondance. Though it could be I suppose.

Anyway, question! The hitch in the substitution ciphers and the Vigenere ciphers seems to be associated with language structure, spelling etc. If I used a computer to limit the letter frequencies and used 'yoda speak' to compose the message, could I do a pencil code that could not be cracked?

What would be the minimum length of enciphered data that could be cracked?

David Davis
Toledo, OR 97391

Sure. Create a randomized one time pad, or use a book as your key. Without the key, it's unbreakable, though with a book key, knowing that it's a book might allow someone to eventually do so. Small pads can be used as well, though the shorter the key the more likey it can be broken. Avoid simple subsitution codes as they are generally easy to break.

__________________
Howling from the Shadows

It must be fun to lead a life completely unburdened by reality. --- JayUtah

You can't reason an irrational person out of an irrational belief. --- Noclevername

Apollo: The History and the Hoax
Enter the World of Athran

In the late 60s in San Fransisco there was a serial killer who went by the name Zodiac. His first killings were followed by a three part letter mailed to three different newspapers. Each was a code. Technically, it was a substitution cypher, but for the more common letters, like E and S, he used a rotatating set of symbols. The letter E had 11 different symbols. The letters were sent tothe FBI and naval intelligence and remained unbroken. The uncoded letter that cam with the coded ones said that if the coded letters were not published there would be more killings. Two newspapers published them on the first day, the thind waited until the next day to do it. A man or his wife that liked puzzles sat down with the three papers that Sunday morning when they had finally seen all three parts. Whichever liked puzzles started working on it and the spouse who had no interest in puzzles up until then joined in. They had craked it within 3 hours. the FBI and Naval Intel confirmed that their solution was most likely correct.

A very short message with a rotating set of symbols may be nearly impossible to crack, but I don't think any cypher created by a human would be totally impossible. Someone somewhere will have an "a-ha" moment and the whole thing will open up.

Personally, I've always been partial to doing multiple cyphers. Scramble the letters, using odd words with a minimum of double letters and common strings, then using a box cypher to finish it off.

__________________
A Nerd can figure out how long it will take the original Enterprise traveling at warp 6.5 to travel from Regulus to Antares.
A Geek will think he can use that to pick up a girl in a bar.
A Dork knows he can't pick up the girl with it, but will hang around for hours anyway, just in case she asks.
She might. You never know.

If you're planning on using a key of finite length (not the Phantomwolf mentioned), I would bet that a message ten times the length of the key would be easy to break. Years ago, a few people took me up on the challenge--piece of cake.

I started messing with the "Rail Fence", I think that is what it is called. Where every other letter is moved to a lower string, then the lower string is moved to the end of the top string. This would break up the basic word patterns, 'qu', 'ee' etc. but it is easy to solve.

I got the idea to move the lower string a fixed number of spaces and reinsert it. This made me think of doing this more than once. The key would be some repetative number and that made me think of a phone number.

So, using a phone number as the key, i.e. 555-1212, split the message off into some number of staggered strings, two is the classic. Now move the bottom string five spaces left or right and reinsert it into the message string. Do it again using the phone number as the spacing guide.

David Davis
Toledo, OR 97391

One of the principles of cryptology is that all analysis assumes that the "enemy" knows the system.

Any system is safe, if the enemy doesn't care to expend the energy to crack it. If you assume a large amount of energy, resources, and time, then no system is truly safe. Even one time keys have their weaknesses: key distribution, key security, key misuse, etc. If it's just you and a friend, and your messages are fairly innocent and not attractive to anyone else, the caesar sub would work. Da Vinci just wrote backwards.

The broad rule of thumb when encrypting is this.

Assuming your enemy knows the way the encrytion works, the data will be useless to him by the time he breaks the cypher.

That means that if your data is only sensitive for 24 hours, a basic transposition or subsitution cypher might work fine. If it is sensitive for 10 years, go for an RSA cypher, if it's sensitive for 50+ years, one time pad it.

__________________
Howling from the Shadows

It must be fun to lead a life completely unburdened by reality. --- JayUtah

You can't reason an irrational person out of an irrational belief. --- Noclevername

Apollo: The History and the Hoax
Enter the World of Athran

That was also a hitch with the German's Enigma usage in WW2. Their rock-solid belief that it was unbreakable led operators to have standard message formats - eg. weather reports always started by saying "Wetter" (that's German for "Weather", not that it was always raining ). When the breakers realised this, they simply searched for which arrangement of the dials would give this result. Hey presto: the code keys for that day !

Later, the Germans introduced code abbreviations to break up the language (eg. a weather report could be denoted by "AB", raining could be "XY", so what's sent out would be the encyptment of "ABXY"). This both removed language sequences, and also shortened the messages. A lot tougher to break, but they managed it.

And, If ya' Wanna Fake a Moon-Landing, Forget The Codes Completely ...

Do it All By Messenger, and Then KILL, Every One of 'em!

Cue Really Evil Laugh ...

__________________
If you Ignore YOUR Rights, they Will go away.

Vigener can be tackled statistically as well, even without knowledge of the language involved, but for that I'd expect to need at least 10-20 times longer message than the key.
I wrote a program to eat vigeners once which didn't use knowledge of any language, only statistics.
It's subsecond crackable by a computer when it's plain language that's encoded.

__________________
"God bless thee, my son; I will give thee the greatest jewel I have ...
"The end of our foundation is the knowledge of causes, and secret motions of things; and the enlarging of the bounds of human empire, to the effecting of all things possible."Trying to make sense of computers, The Error Log.

Having had some time to read and think a bit, one question I have is; how long does it take to break a cipher? I would suspect that a practised cryptoanalyst could crack one with a computer PDQ. With that in mind..

I would like to test some of you cryptos, if you are game. I will provide a cipher, via email for anyone to break, unless the moderator suggests I just post the blocks here. I will send it in blocks of around 15 words, maybe more, maybe less, for as long as it takes for participants to break it. No fancy encryption here, just paper and pencil stuff. I will use the same cipher technique all the way through and I will post it after the fact, if there is any interest.

My questions would be, how long did it take? How did you do it? If you use some proprietary tool that is hush hush, you will obviously not be able to answer that last one in detail, but maybe you could tell me, by computer or by hand. What was your aha! moment? When did you know you had the key and why?

David Davis
Toledo, OR 97391

I'm game. Fire away.

__________________
A Nerd can figure out how long it will take the original Enterprise traveling at warp 6.5 to travel from Regulus to Antares.
A Geek will think he can use that to pick up a girl in a bar.
A Dork knows he can't pick up the girl with it, but will hang around for hours anyway, just in case she asks.
She might. You never know.

Traditional problems include the type of technique too. Also assumed in commercial or governmental applications as well, that the enemy knows your technique, just not the password.

ovgmin eyhav bjsoect anj dksain hu nta three mc eno eehr widgtdling indlymessl crsaos sg kesd arrgcyin u rbcum zi norsemou roposprtion nc mc u rbcum zi hewol htwea rdbea rmfo sg andhswic dna c oehp yb njseoy nc tub nc mc gerbig hnta eht tna sv hetre li orfu istme


I expect this one to stand for about a nano second. How did you crack it? What were your cribs, if any? What are the rules of the game?

David Davis
Toledo, OR 97391

The first three words are obviously anagrams of "Moving heavy objects" I could see that just looking at it.

PS: I also see "mindlessly across my desk carrying a crumb of enormous proportions" and "whole wheat bread from" and "I hope he enjoys" and "bigger than the ant by three or four times" I think the technique should also be revealed. Otherwise, it's too easy to come up with an uncrackable code of that length--I could use a simple caesar sub, but scramble each passage with the same random number generator, like a super rail fence, which would be broke eventually but the first pass would be impossible.

As I expected, a practiced eye could unscramble the words easily. The technique to encypher the message ..

Divide the word in half. If the word is of an odd number of letters then split it for the smaller amount to be at the start of the word, i.e. heavy = he avy. Now put the front part at the back, this yields avyhe. Leave the first and last letters together, yh, and swap the others end to end, e yh av=eyhav. Remember when you reverse the process that the smaller of the two parts belongs at the front of the word and keep your first and last letters in the proper orientation. Or put another way, the first and last letters are at the split point for the word, eyhav, would split at the yh. So those are your first and last letters, leave them alone. Shuffled the other letters end to end first and then split the word. eyhav = av yh e = avy he = he avy

Small words, one and two letters were impossble to hide, so I reversed them and did a seven letter shift left. The receiver would know this and work accordingly, the cracker should be able to interpret the words needed based on the rest of the text and so detect the shift.

My idea being to make a simple set of rules to follow so that it could be discussed casually over coffee, then used in the real world. This one would obviously benefit from a second level of encryption. BUT for the scramble technique to work the word form must be maintained.

OK, the next one will be harder and I will add text until it is broken. I will give some hints as to how it is made, but my question is can it be broken and at what point.

David Davis
Toledo, OR 97391

The Beale ciphers are a classic. When I started to look into cryptography I found a lot of pages on the Beales. There are some good opinions on the structure of them and the solvability of them. I know the number 2 was cracked using the DoI, but would a "book code" be the only way to make that table? If Beale were long enough, could we deduce the Declaration of Independance from the encrypted text?

This next problem is a variation on that theme. A basic substitution cipher it should yield to letter frequency analysis eventually. This is part one as it were and additional parts will be posted till someone breaks it.

BTB: It is tedious work doing a cipher by hand.

81 93 21 19 32 37 44 31 91 47 25 36 112 68 87 712 38 51 99 611 54 712 810 96 89 94 49 810 65 112 12 37 57 312 91 46 811 612 812 75 29 12 54 47 211 111 68 53 712 58 312 99 212 29 76 97 54 37 48 58 22 810 26 24 93 56 911 73 87 89 19 20 210 45 110 710 212 112 810 57 78 91 18 66 810 86 47 19 410 24 22 47 46 78 93 44 62 59 66 97 88 812 211 93 38 33 711 89 84

David Davis
Toledo, OR 97391

No, the rules should be that you tell us exactly how it is made. Otherwise, it's too hard, and a waste of time.

You want the complete how to, including the table or the book or what ever is used to create the cipher? Well, then it is solved isn't it? Pardon my ignorance, I just don't understand what you want, exactly. Can you give me an example?

David Davis
Toledo, OR 97391