Re: C-equivalence aware hash function

Polarity preservation is the easy step as I demonstrated earlier.

The row and column sorting wouldn't be a problem if you only had to do
one of them, as Rose showed they arn't communitive. I'm still not sure
how to solve FG's problem.

JLC

Mike Amling <nospam@icic.gov.ir> wrote:
> To canonicalize, you also have to disambiguate the polarity reversal
> criteria for matrices with an even number of rows or an even number of
> columns. Both of the matrices

> 1 -1
> 1 -1

> and

> 1 -1
> -1 1

> are canonical under Dr. Wagner's specification in that they have sorted
> columns and rows, and do not invite polarity inversion. But they are
> C-equivalent.

> --Mike Amling

--