Documentation

TauCeti.AlgebraicGeometry.EllipticCurve.FormalGroup.Point.Range

The range of the formal parametrisation #

Point/AdicCompletion.lean maps the formal-group parameters of the maximal ideal of š’Ŗ_v injectively into the points of the curve over the completion K_v. This file computes the image of that map: a point is parametrised exactly when it is the point at infinity, or its x-coordinate has a pole. Together with the injectivity already proved, this is the content of Silverman AEC VII.2.2 — the identification of the formal group Ê(š”Ŗ) with the kernel of reduction E₁(K_v).

The image is described by the valuation of the x-coordinate, which is the condition that cuts out that kernel: a point of E(K_v) reduces to the point at infinity exactly when its x-coordinate has a pole. Neither reduction nor E₁ is named below, no reduction map on points entering the statements.

Main results #

References #

Provenance #

The parametrisation whose image is computed here, and its injectivity, come from the Stoll development pinned in Point/Basic.lean's provenance.

A nonzero parameter has a pole in the x-coordinate of its point. The closed form x * (t ^ 2 u(t)) = 1 of xCoord_formalPoint_mul_eq_one makes x inverse to an element of the maximal ideal, t lying there and u(t) being integral.

A point with a pole in its x-coordinate is parametrised, by the parameter -x / y. This is the converse of one_lt_valuation_xCoord_formalPoint, and the substantial half of Silverman AEC VII.2.2.

The pole of x is what puts both -x / y and -1 / y in the maximal ideal: v x < v y at such a point, so v (x / y) < 1, and 1 < v y gives v (1 / y) < 1.

@[simp]

The range of the formal parametrisation of an adic completion: the point at infinity together with the points whose x-coordinate has a pole. Since the left-hand side is the range of an additive homomorphism, this exhibits that set of points as a subgroup; classically it is the kernel of reduction E₁(K_v).

A point outside the range has integral coordinates. The parametrised points are exactly those with a pole, so a point of the complement has v x ≤ 1, and then the Weierstrass equation forces v y ≤ 1 as well. This is the other half of the dichotomy of Affine/ValuationIntegrality.lean, read against the range.