Ask Your Question
2

Symbolic calculations in finite field extension

asked 2016-08-04 10:12:00 -0500

mksanders gravatar image

Hello,

Please consider the following code snippet:


P_GF2.<X> = PolynomialRing(GF(2))
GF23.<x> = P_GF2.quotient_ring(X^3+X+1)

def f(val):
  return val**3

This works as expected, when val is something like $1+x+x^2$. What I wanted to do is to calculate the value of $f$, but using a generic element of $GF(2^3)$, e.g. $a_2x^2+a_1x+a_0$. The idea is to have the result expressed in terms of the $a_i$. Is this possible in Sagemath?

I have tried to do it using symbolic variables, but they always belong to the Symbolic Ring, which (as far as I can tell) does not mix with other rings. Because this example is small, I was able to do the computations by hand; the value of having SAGE doing it is of course, to apply it to cases that are infeasible to do without a computer.

Thank you very much in advance.

edit retag flag offensive close merge delete

1 answer

Sort by » oldest newest most voted
5

answered 2016-08-04 22:21:49 -0500

tmonteil gravatar image

updated 2016-08-06 15:13:19 -0500

Instead of symbolic variables, you can define the ai as polynomial indeterminates, as follows:

sage: P_GF2.<X,A0,A1,A2> = PolynomialRing(GF(2))
sage: GF23.<x,a0,a1,a2> = P_GF2.quotient_ring(X^3+X+1)
sage: def f(val):
....:       return val**3
sage: f(a0*+a1*x+a2*x)
x*a0^3*a1^3 + a0^3*a1^3 + x*a0^2*a1^2*a2 + a0^2*a1^2*a2 + x*a0*a1*a2^2 + a0*a1*a2^2 + x*a2^3 + a2^3

By the way, note that the field GF(2^3) is well defined in Sage:

sage: F.<x> = GF(2^3)
sage: F
Finite Field in x of size 2^3
sage: x.minpoly()
x^3 + x + 1
sage: R.<a0,a1,a2> = PolynomialRing(F)
sage: f(a0*+a1*x+a2*x)
(x + 1)*a0^3*a1^3 + (x + 1)*a0^2*a1^2*a2 + (x + 1)*a0*a1*a2^2 + (x + 1)*a2^3

EDIT

Note that since you work on a field of characteristic 2, you necessarily have ai + ai = 0:

sage: a0+a0
0

If you want your variables to also satisfy ai^2 = ai, you can just add this condition when you define the quotient ring:

sage: sage: GF23.<x,a0,a1,a2> = P_GF2.quotient_ring([X^3+X+1, A0^2+A0, A1^2+A1, A2^2+A2])
sage: GF23
Quotient of Multivariate Polynomial Ring in X, A0, A1, A2 over Finite Field of size 2 by the ideal (X^3 + X + 1, A0^2 + A0, A1^2 + A1, A2^2 + A2)
sage: a0^2
a0
sage: a0^3+a0
0
sage: f(a0*+a1*x+a2*x)
x*a0*a1 + a0*a1 + x*a2 + a2
edit flag offensive delete link more

Comments

This is most helpful, but there is a detail still missing: what I wanted is a way for a0, a1 and a2 to be treated as generic elements of GF(2). Meaning for example, that $a_i^n=a_i$ and $a_i+a_i = 0$. Without this, the resulting expression is needlessly complicated. Can it be achieved?

mksanders gravatar imagemksanders ( 2016-08-06 05:39:12 -0500 )edit

I updated my answer.

tmonteil gravatar imagetmonteil ( 2016-08-06 14:46:55 -0500 )edit

I accepted this answer because, well, it fully answers my question. But now I must ask for what would truly be the cherry on top of the cake: is there any way to group the result by 'x'? I.e. to make the result of your example look like x*(a0*a1+a2) + a0*a1 + a2? I tried these methods, but they don't work. The closest I can get is sorted(r.monomials()), where r is the produced result. Again, thank you in advance.

mksanders gravatar imagemksanders ( 2016-08-07 05:41:55 -0500 )edit

Should I have asked another question for this?

mksanders gravatar imagemksanders ( 2016-08-10 02:09:36 -0500 )edit

@mksanders, sure, why don't you ask a follow-up question. For reference, you can state in the new question that it is a follow-up to http://ask.sagemath.org/question/3433... .

slelievre gravatar imageslelievre ( 2016-08-23 03:04:40 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower

Stats

Asked: 2016-08-04 10:07:17 -0500

Seen: 79 times

Last updated: Aug 06 '16