Talk:Twofish
From Wikipedia, the free encyclopedia
[edit] Information
There appears to be a great deal of information regarding the AES selection and Twofish (and all the other nominees) here at [1]. I'm going to incorporate parts of this into the article, including the info about supposed breaks, but be forewarned, I'm not a crypto geek. Sword 17:45, 4 Nov 2004 (UTC)
[edit] Murphy & Robshaw attack
Twofish's S-boxes are key dependent but do not change with each encryption. The attack referred to is [2] (ps) but it's not very relevant (it only breaks five of Twofish's 16 rounds) and should either form part of a longer section on cryptanalysis of Twofish or be dropped. I've removed it for now. ciphergoth 13:01, 2004 Nov 16 (UTC)
- For reference, S. Murphy and M. Robshaw report that knowing that the S-boxes change with each encryption, that one can craft a differential attack by choosing a plaintext and seeing what the output is. This has only been suggested against a reduced-round (near 5) Twofish. is the removed text. — Matt 13:16, 16 Nov 2004 (UTC)
[edit] Diagram ERROR
The one-bit left rotate in the column containing K2 and K6 should be a one-bit RIGHT rotate. The other rotate (in the rightmost column) should indeed be a left rotate. Compare with figure 1 on page 6 of the twofish paper.
It also might be worth describing the symbols for addition modulo 232 and XOR. —The preceding unsigned comment was added by 192.35.100.1 (talk • contribs) 07:27, 8 June 2006 (UTC)