extended vigenere cipher problem

Hi all,

Firstly apologies for a somewhat cryptic(!) post, I'm terrible at explaining problem like this, but I hope I haven't made it too difficult to understand my meaning.

I'm just playing around writing a program for an extended vigenere cipher(1). Normally the vigenere cipher is mod 26 but mine is mod 127. This means you don't have to strip out non A-Z chars and it will encode spaces, punctuation, everything.

I've written this same program previously in ruby so I know the maths behind it is correct. However in this C version, decryption is not working correctly. Encrypting produces the correct output, however I can't track down why decrypting isn't working.

For example - expected output:

Code:

---

Plaintext:  This is a basic test message.
Codeword:  codewordcodewordcodewordcodew
Ciphertext: 8XNY^XYf^EE^PGGkYV^EXUXZ^X]XXWQLK&y
Plaintext:  This is a basic test message.

---

Actual output:

Code:

---

Plaintext:  This is a basic test message.
Codeword:  codewordcodewordcodewordcodew
Ciphertext: 8XNY^XYf^EE^PGGkYV^EXUXZ^X]XXWQLK&y
Plaintext:  Õéêô¡êô¡â¡ãâôêä¡õæôõ¡îæôôâèæ¯

---

Note that the pattern of the last line is exactly right. You can see where the spaces are compared to the original message, and the full stop at the end.

The encryption line is:

Code:

---

(c + Codeword::Instance()->getNext()) % 127

---

And the decription line is:

Code:

---

(c - Codeword::Instance()->getNext()) % 127

---

Codeword::Instance is a singleton class which will automatically return the next character in the codeword and queue the next one up. Everything looks fine in debugging as far as I can tell, but my mathematics isn't that strong.

Grateful for any assistance or pointers.

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

All fixed, the problem was that the C program was using the full extended charset. Once I changed to mod 255, it all works perfectly. Strange that it was different between Ruby and C, but never mind :)

Beware, vigenere is one of the easiest cyphers to crack. It's not secure at all.

Yeah, you should consider something safer, like rot13 ;)

Lol don't worry I'm not expecting it to hide my secrets. I wanted to see the comparison between the traditional vigenere cipher and the version I've done. Theoretically, with a decent length key, the frequency analysis attacks should be far less dangerous.

Well, that's true. At the extreme end of the key length spectrum you have the one-time pad, which of course is completely secure (up to the security of getting the key to the recipient, of course).