Re: [math-fun] Characterizing polys which permute GF(p)

Could it be, that some other "nice" set of polynomials generates the full perm group, but while RCS's set involved p-1 in exponents, the nice set is defined with, say, only bounded degrees. Then you could do everything RCS does, but using that generator set instead, and then you'd overcome the Criticism. Dickson found all perm polynomials of degree<=5 in finite fields; when are they enough to generate? -- Warren D. Smith http://RangeVoting.org <-- add your endorsement (by clicking "endorse" as 1st step)