Date: Wed, 8 Apr 2009 23:03:49 -0400 From: Watanabe Dai Dear all, I looked at Wu et. al's report [1] on LUX and I believe that their observation can obviously be extended to a collision attack and a second preimage attack which would be labeled orange in the SHA-3 zoo according to their computational complexities. The following is the brief sketch of the attacks. == Wu et. al's observation Let H=(h0, h1, ..., h7) be the hash value and hi=(ai, bi, ci, di) be its byte expression. Then the following relation holds for 0