Is it really that difficult?

[J.J. Maxx]

All the time you hear about codes that different countries use to transmit top secret information and I was wondering today how hard it is to make a code that is impossible to crack.

The thing is, I don't know the process that you (or the CIA) goes about cracking codes. For example, lets say you took a small sentence like;

Meet me at midnight

Okay, so you have to replace all the letters with other letters or multiple letters using a system that can translate the message. You could, for example, set all the letters as their sequential number in the alphabet.

M = 18
E = 5
T = 20
A = 1

And so on and so forth. But something like that would be so very easy to crack because there would be a pattern. Every time an 'a' was used there would be a '1'. So it would be as simple as solving an anagram. Most codes I thought of had this inherant problem.

But if you could add some sort of random factor to the code while still retaining the key to translate it then you would be getting closer to a really good code.

I will post my thoughts on making a code later I have some neat ideas. Anybody have any thoughts about this subject?

[AtmaWeapon]

Any code can be cracked. If there is a way to decode it, there is a way to crack it. It can become very difficult to crack, but it is still crackable.

[Fatty Lumpkin]

Having a one-time-use decoding key helps a lot. Use it once, throw it away. Then if someone gets a hold of one, you won't be in as bad shape. But the problem is distributing all the keys safely.

[MottZilla]

There is some encryption that some phone company or something uses for their service that is supposed to be really hard to crack. But these people that cracked part of it had like some super computer and brute force attacked it or something. So I don't think that's really figuring it out, just waiting till it comes to you.

Anyway encryption and "codes" are just ways of slowing down finding out the meaning of information while it being discovered would be a bad thing. Cause with enough time and effort, you really can't stop someone.

Anyway, it is a very interesting subject.

[fatcatfan]

RSA encryption uses a one-way function that *is* crackable but would require all computers in the world working in tandem longer than the age of the universe to crack it by brute force. beyond that, if you don't have the key, it's essentially undecipherable.

Quantum cryptography is *impossible* to crack. At least until we develop quantum computers. I'm not near my "Code Book" so I can't say more than that right now, but when I get home, I'll share more.

[Paradox]

there's a lot more stuff than number substitution to code.

there's the classic every-other-letter thing, so "I am god" becomes
I M O A G D

then there's a complicated grill format, where you take a 10x10 grid and cut some holes into it, so to encode something, just fill the grid with letters and another person with a grill can place it over the grid with the right letters showing. kind of like a word search, but with letters.

then there's the classic "TEXT TO FRENCH TO LEET TO ZELDA64GLD" code. Let me demonstrate.
I am God.
Je suis dieu.
J3 5u15 d13u.
3J su51 d31u. <--- practically indecipherable from the original text.

[TheGeepster]

Current 128 bit PGP takes a long time to crack using brute force. It uses really large prime numbers (There are formulas to find large primes using logarithms and such, but no good method exists to test a number for prime factors that work with any speed for numbers that large). I think it has 3 parts. It has a public key (large prime), and a private key (a second large prime), and the message (Encoded using the 2 keys).

[The Savior]

Originally posted by fatcatfan
RSA encryption uses a one-way function that *is* crackable but would require all computers in the world working in tandem longer than the age of the universe to crack it by brute force.

Isn't it possible that the first code that they try would work? Obviously it would be very unlikely, but possible, right?

[fatcatfan]

it is possible, but highly unlikely. RSA (which is used by PGP to encode only the *key* to a one-time cipher used to encode the remainder of the message) is based on extremely large prime numbers, and to decrypt RSA, you must factor the product of two extremely large prime numbers. There's no known mathematical shortcut for factoring, and a brute force method would have to start by correctly guessing one of the prime numbers in order to finish in just one iteration. So even assuming you have a large database of *all* prime numbers, there's no shortcuts.

[DarkDragon]

One-time pads are also impossible to crack. Suppose I and a friend had the same sequence of random letters, say, I H S B L Q. I could then encode "decode" by "adding" each of its letters to the random sequence, yielding MMVQPV. Unless someone takes a peek at either my or my friend's pad, this message is uncrackable, since any six-letter word is just as likely to be MMVQPV as any other six-letter word.

As fatcatfan has already explained, PGP is also extremely secure and, more importantly, much less cumbersome since you don't have to secretly distribute one-time pads. Of course, as soon as I discover a method of factoring large composite numbers, the NSA will make me rich. :sb:

[Xyvol]

During World War II the US used the language of the native Navajo people to make a code. I don't know all the details, but I believe it worked rather well for a time. True, it could be sloved if you learned the language, but not many people in other countries even knew of the Navajos. Something like this could work again. Make something cryptic to start with, then translate it to another language, the more unknown the better.

[inori]

Of course, there's the time-honored way to crack any encryption: send your spies to steal the key. :kawaii:

[J.J. Maxx]

Originally posted by Xyvol
Something like this could work again. Make something cryptic to start with, then translate it to another language, the more unknown the better.

Like Zelda64Gld? :D

[Menne]

Currently there is a project running which trys to crack the RC-72 Bit encryption algorithm with brute force.
The current speed is 54,746,397,783 Keys/sec, and still it would take around 685,345 more days to test all keys.
In the whole there are 4,722,366,482,869,645,213,696 possible keys, 378,407,101,476,634,624 are already tested.

The project is hostet here: http://stats.distributed.net/projects.php?project_id=8

Even if it would be possible to test all key with a super computer in one month with, it most likely to late, because nowadays a war can be over in one month.

[Strider]

Originally posted by fatcatfan
Quantum cryptography is *impossible* to crack. At least until we develop quantum computers.

It should be inpossible even after quantum computers are made, becuse once your read the quantum stream it will be altered, so if someone spy the real one will notice that the stream have been altered.