PDA

View Full Version : Normal form for Cubic


Newtime
Aug8-11, 05:32 PM
Is there an algorithmic way to put a given elliptic curve into Weierstrass normal form? If not, what's the general procedure?

micromass
Aug8-11, 08:15 PM
Is there an algorithmic way to put a given elliptic curve into Weierstrass normal form? If not, what's the general procedure?

Yes. Say that C is a cubic. We start by picking a rational point O on the cubic. Then we take the tangent line of C at the rational point O. This will intersect the cubic in another rational point. We take the X-axis to be the tangent line at that other rational point. And we let the Y-axis to be any line through O.

By changing coordinates, you get an equation of the form

xy^2+(ax+b)y=cx^2+d+e

Multiply by x and change into the variable u=xy. This will get you something of the form

u^2+(ax+b)u=cx^3+dx^2+ex

Change the variable again by setting v^2=u^2+(ax+b).

Now change the variables once more to obtain that c=1.