Re: Q: SAC
- From: Mok-Kong Shen <mok-kong.shen@xxxxxxxxxxx>
- Date: Mon, 05 Oct 2009 10:58:59 +0200
Sure... the number of ones is fine after the transposition, the only
problem could be the bijectivity.
It passed a small test I wrote. It found no solution for even n. So
I'm happy with it for now.
I don't quite understand your problem. The bijectivity is in this
case the same as the non-singularity of the matrix, which however
is particularly easy to check with gaussian elimination, since the
matrix is boolean.
Neither I can now, but it worked at the time of my last posting.
Obviously there're redirecting to https now (for whatever reason). Try
look at "Specification Version 2.0 (15 Sep. 2009)", pages 14 and 23
"Supporting document (15 Sep. 2009)", pages 6-7, 10-11
I'm layman too, but I'd say it's hardly simpler than LDMax and LPMax
and much less important. Look at
In case you find it complicated, just try to understand "Table 4.
Linear Approximation Table" and "Table 7. Difference Distribution
Table" - computing them is no harder than evaluating avalanche.
Doing the computation of the avalanche measure for the function
like the one I found -- I mean when the function is given -- is
actually very straightforward and fast. Thus I don't see why this
apparently useful information about any such functions used in
crypto should be ignored.
M. K. Shen
- Re: Q: SAC
- From: Maaartin
- Re: Q: SAC