Improve and Secure Your Cipher: Expert Tips and Feedback

[anarkest]

i have made a new cipher and was hoping you guys could comment on it, tell me how i could improve it, and what's wrong with it. its a very simple mathematical cipher, and i feel it is very secure. please try to crack it. Heres a link to it https://docs.google.com/document/pub?id=1ZaoMnAKtkswUh3k8ldKta-XeOwXMlhMlQnHAeo7A3uU

thanx!

 

[CRGreathouse]

I don't think the key improves security. It should not be too difficult to look at the fractional parts and decide what numbers are possible keys (depending on how many digits are given). And depending on how many digits you give, you may also clip letters from the end of long words.

 

[PrakashPhy]

what if the division gives irrational number?

 

[Office_Shredder]

what if the division gives irrational number?

What if dividing two integers gives an irrational number?

I dare say you probably don't care what the message is anymore.

The simplest attack on this is obviously the fact that you need a key.

A lot of work might be able to be done to determine what the key is even if somebody doesn't know. For example, if I see that one of your encrypted words is .015 I know that your key must be relatively large, so starts with a letter near the end of the alphabet. I'm sure that some number theory could be used to drive home at possible keys.

There's no real reason to break up your numbers into blocks of 8 besides this cell phone thing. Usually characters are broken up into blocks so that when transmitted you don't give away what the word lengths are. You don't have such a problem here, and it could make it harder to decrypt if you use different length codewords, because then you'll be transmitting a wider range of numbers. This would mean that the intercepting party has no idea how long the codeword is which would probably be a good thingOne good thing is to not send over the whole decimal, which is how you presented it. Otherwise I could simply find what the decimal is in smallest form and then just scale it upwards until the numerator and denominator both fit the pattern necessary for a codeword. I tried that on two of the decimal numbers you had and it doesn't work because you don't get the full number

 

[bapowell]

My initial thought is that the security of modern ciphers is such that knowledge of the algorithm, which is typically publicly available, does not jeopardize the security of the cipher. In your case, if I know the algorithm, a simple brute force attack on the key space should crack the cipher, since I need only to multiply each cipher text word with different keys. Now, the security of your cipher certainly improves as you make the key longer. However, now you must also make the world length longer. In the limit that the key becomes equal in length to the plain text message, you have the well known one-time pad, which is indeed provably secure (in the event that your key is truly random and used only once).

 

[Mark44]

what if the division gives irrational number?

That can't possibly happen. Both numbers involved in the division are integers, so the quotient will necessarily be rational. By definition, a rational number is the quotient of two integers.