Projective completion of the characteristic-three Kummer count
DefinitionN25_ExplicitKummerThreeProjectivemazur-n25
The file promotes the eigenbasis change to an equivalence on normalized projective coordinates and defines the dense Segre and denominator-cleared Kummer solution types, together with their partitions into and parts. It identifies the dense zero fibre with one point, the cleared zero fibre with two points, and the nonzero parts through , while retaining the two boundary rulings and their corner. This projective bookkeeping converts the affine Kummer pairs into the points of the canonical curve over .
Definition code
import Definitions.Def_N25_CanonicalThreeNormalization
import Definitions.Def_N25_ExplicitKummerThreeF81
import Theorems.Thm_KeystoneLadder_SameP1Vec_trans
/-!
# Projective bookkeeping for the characteristic-three Kummer model
The Kummer certificate counts 87 affine pairs `(r,t)`. This file proves the
two projective corrections structurally:
* the cleared equation has two solutions with `r=0`, whereas the dense Segre
chart has only `(r,s)=(0,0)`, so the dense chart contains 86 points;
* the two boundary rulings and their corner each contain one curve point.
Consequently the normalized `P¹ × P¹` Segre model has 89 points over `F81`.
All comparisons away from `r=0` are explicit equivalences induced by
`t=r²s`; no two-dimensional finite search is used.
-/
namespace MazurProof.RationalPointsN25QuotientKummerThreeProjective
open MazurProof.RationalPointsN25QuotientF2
open MazurProof.RationalPointsN25QuotientWeil
open MazurProof.RationalPointsN25QuotientWeilThree
open MazurProof.RationalPointsN25QuotientKummerThree
open MazurProof.RationalPointsN25QuotientKummerThreeF81
open scoped BigOperators
noncomputable section
/-! ## Canonical normalized points over a field -/
/-- The executable `F81` chart coordinates are the same coordinates supplied
by its transported field structure. -/
theorem ternaryCoordinates_f81_eq (P : NormalizedProjective4 F81) :
ternaryCoordinates f81Operations P = normalizedCoordinatesThree P := by
cases P <;> simp [ternaryCoordinates, normalizedCoordinatesThree,
f81Operations_zero_eq, f81Operations_one_eq]
/-- The executable `F81` quadric evaluator is the actual field polynomial. -/
theorem canonicalQuadric25Ternary_f81_eq (P : Coordinates4 F81) :
canonicalQuadric25Ternary f81Operations P =
canonicalQuadric25Three P := by
unfold canonicalQuadric25Ternary canonicalQuadric25Three
rw [f81Operations_add_eq, f81Operations_add_eq,
f81Operations_neg_eq, f81Operations_mul_eq,
f81Operations_neg_eq, f81Operations_mul_eq,
f81Operations_add_eq, f81Operations_mul_eq,
f81Operations_add_eq, f81Operations_mul_eq,
f81Operations_mul_eq]
ring
/-- The executable `F81` cubic evaluator is the actual field polynomial. -/
theorem canonicalCubic25Ternary_f81_eq (P : Coordinates4 F81) :
canonicalCubic25Ternary f81Operations P =
canonicalCubic25Three P := by
unfold canonicalCubic25Ternary canonicalCubic25Three
repeat' rw [f81Operations_add_eq]
repeat' rw [f81Operations_neg_eq]
repeat' rw [f81Operations_mul_eq]
ring
/-- The old executable normalized-point predicate has exactly the semantic
meaning asserted by the field-valued canonical equations. -/
theorem isCanonicalNormalized25Ternary_f81_iff
(P : NormalizedProjective4 F81) :
IsCanonicalNormalized25Ternary f81Operations P ↔
IsCanonicalNormalizedThree P := by
unfold IsCanonicalNormalized25Ternary IsCanonicalNormalizedThree
rw [ternaryCoordinates_f81_eq]
dsimp only
rw [canonicalQuadric25Ternary_f81_eq,
canonicalCubic25Ternary_f81_eq, f81Operations_zero_eq]
/-! ## Projectivizing the eigenbasis -/
/-- A first-nonzero normalized projective representative is never the zero
homogeneous vector. -/
theorem normalizedCoordinatesThree_ne_zero
{K : Type*} [Field K] (P : NormalizedProjective4 K) :
normalizedCoordinatesThree P ≠ zeroCoordinatesThree := by
cases P <;> simp [normalizedCoordinatesThree, zeroCoordinatesThree,
Coordinates4.mk.injEq]
/-- Normalizing a nonzero scalar multiple of an already normalized vector
returns the original projective chart. -/
theorem normalize_scale_normalized
{K : Type*} [Field K] [DecidableEq K]
(P : NormalizedProjective4 K) {a : K} (ha : a ≠ 0) :
normalizeCoordinatesThree
(scaleCoordinatesThree a (normalizedCoordinatesThree P)) = P := by
cases P with
| xChart y z w =>
simp [normalizeCoordinatesThree, scaleCoordinatesThree,
normalizedCoordinatesThree, ha]
| yChart z w =>
simp [normalizeCoordinatesThree, scaleCoordinatesThree,
normalizedCoordinatesThree, ha]
| zChart w =>
simp [normalizeCoordinatesThree, scaleCoordinatesThree,
normalizedCoordinatesThree, ha]
| wChart =>
simp [normalizeCoordinatesThree, scaleCoordinatesThree,
normalizedCoordinatesThree]
/-- The inverse eigenbasis commutes with homogeneous scalar multiplication. -/
theorem canonicalCoordinatesFromEigen_scale
{K : Type*} [Field K] (e a : K) (U : Coordinates4 K) :
canonicalCoordinatesFromEigen25Three e
(scaleCoordinatesThree a U).x (scaleCoordinatesThree a U).y
(scaleCoordinatesThree a U).z (scaleCoordinatesThree a U).w =
scaleCoordinatesThree a
(canonicalCoordinatesFromEigen25Three e U.x U.y U.z U.w) := by
rcases U with ⟨u1, u2, u3, u4⟩
simp only [scaleCoordinatesThree, canonicalCoordinatesFromEigen25Three,
Coordinates4.mk.injEq]
constructor
· ring
constructor
· ring
constructor <;> ring
/-- The forward eigenbasis commutes with homogeneous scalar multiplication. -/
theorem canonicalEigenCoordinates_scale
{K : Type*} [Field K] (e a : K) (P : Coordinates4 K) :
canonicalEigenCoordinates25Three e (scaleCoordinatesThree a P) =
scaleCoordinatesThree a (canonicalEigenCoordinates25Three e P) := by
rcases P with ⟨x, y, z, w⟩
simp only [scaleCoordinatesThree, canonicalEigenCoordinates25Three,
Coordinates4.mk.injEq]
constructor
· ring
constructor
· ring
constructor <;> ring
/-- The inverse eigenbasis sends a nonzero vector to a nonzero vector. -/
theorem canonicalCoordinatesFromEigen_ne_zero
{K : Type*} [Field K] [CharP K 3] {e : K}
(he : IsCyclotomicFive25Three e) {U : Coordinates4 K}
(hU : U ≠ zeroCoordinatesThree) :
canonicalCoordinatesFromEigen25Three e U.x U.y U.z U.w ≠
zeroCoordinatesThree := by
intro h
have h' := congrArg (canonicalEigenCoordinates25Three e) h
have hleft := canonicalEigenCoordinates_fromEigen
(e := e) (u₁ := U.x) (u₂ := U.y) (u₃ := U.z) (u₄ := U.w) he
apply hU
rw [hleft] at h'
rcases U with ⟨u1, u2, u3, u4⟩
simpa [zeroCoordinatesThree, canonicalEigenCoordinates25Three] using h'
/-- The forward eigenbasis sends a nonzero vector to a nonzero vector. -/
theorem canonicalEigenCoordinates_ne_zero
{K : Type*} [Field K] [CharP K 3] {e : K}
(he : IsCyclotomicFive25Three e) {P : Coordinates4 K}
(hP : P ≠ zeroCoordinatesThree) :
canonicalEigenCoordinates25Three e P ≠ zeroCoordinatesThree := by
intro h
have h' := congrArg (fun U : Coordinates4 K =>
canonicalCoordinatesFromEigen25Three e U.x U.y U.z U.w) h
have hleft := canonicalCoordinatesFromEigen_eigen (e := e) (P := P) he
apply hP
rw [hleft] at h'
simpa [zeroCoordinatesThree, canonicalCoordinatesFromEigen25Three] using h'
/-- Apply the inverse eigenbasis and normalize the resulting canonical
homogeneous vector. -/
def eigenToCanonicalProjective
{K : Type*} [Field K] [DecidableEq K] (e : K)
(P : NormalizedProjective4 K) : NormalizedProjective4 K :=
normalizeCoordinatesThree
(canonicalCoordinatesFromEigen25Three e
(normalizedCoordinatesThree P).x (normalizedCoordinatesThree P).y
(normalizedCoordinatesThree P).z (normalizedCoordinatesThree P).w)
/-- Apply the forward eigenbasis and normalize the resulting eigenvector. -/
def canonicalToEigenProjective
{K : Type*} [Field K] [DecidableEq K] (e : K)
(P : NormalizedProjective4 K) : NormalizedProjective4 K :=
normalizeCoordinatesThree
(canonicalEigenCoordinates25Three e (normalizedCoordinatesThree P))
/-- The mutually inverse eigenbasis matrices induce a genuine equivalence on
first-nonzero normalized projective coordinates. -/
def eigenCanonicalProjectiveEquiv
{K : Type*} [Field K] [CharP K 3] [DecidableEq K]
(e : K) (he : IsCyclotomicFive25Three e) :
NormalizedProjective4 K ≃ NormalizedProjective4 K where
toFun := eigenToCanonicalProjective e
invFun := canonicalToEigenProjective e
left_inv P := by
unfold eigenToCanonicalProjective canonicalToEigenProjective
let U := normalizedCoordinatesThree P
have hU : U ≠ zeroCoordinatesThree :=
normalizedCoordinatesThree_ne_zero P
have hV : canonicalCoordinatesFromEigen25Three e U.x U.y U.z U.w ≠
zeroCoordinatesThree := canonicalCoordinatesFromEigen_ne_zero he hU
rcases normalizeCoordinatesThree_spec hV with ⟨a, ha, hspec⟩
rw [hspec, canonicalEigenCoordinates_scale,
canonicalEigenCoordinates_fromEigen he]
exact normalize_scale_normalized P ha
right_inv P := by
unfold eigenToCanonicalProjective canonicalToEigenProjective
let V := normalizedCoordinatesThree P
have hV : V ≠ zeroCoordinatesThree :=
normalizedCoordinatesThree_ne_zero P
have hU : canonicalEigenCoordinates25Three e V ≠ zeroCoordinatesThree :=
canonicalEigenCoordinates_ne_zero he hV
rcases normalizeCoordinatesThree_spec hU with ⟨a, ha, hspec⟩
rw [hspec, canonicalCoordinatesFromEigen_scale,
canonicalCoordinatesFromEigen_eigen he]
exact normalize_scale_normalized P ha
/-! ## Dense and Kummer solution types -/
section Generic
variable {K : Type*} [Field K] [CharP K 3] [DecidableEq K]
/-- Affine solutions `(r,s)` on the dense Segre chart. -/
abbrev DenseSolutions (e : K) :=
{p : K × K // canonicalDenseSegreCubic25Three e p.1 p.2 = 0}
/-- Solutions `(r,t)` of the denominator-cleared Kummer equation. -/
abbrev KummerSolutions (e : K) :=
{p : K × K // canonicalKummerPolynomial25Three e p.1 p.2 = 0}
/-- Dense-chart solutions on the special fibre `r=0`. -/
abbrev DenseZeroSolutions (e : K) :=
{p : K × K //
canonicalDenseSegreCubic25Three e p.1 p.2 = 0 ∧ p.1 = 0}
/-- Dense-chart solutions away from `r=0`. -/
abbrev DenseNonzeroSolutions (e : K) :=
{p : K × K //
canonicalDenseSegreCubic25Three e p.1 p.2 = 0 ∧ p.1 ≠ 0}
/-- Kummer solutions on the special fibre `r=0`. -/
abbrev KummerZeroSolutions (e : K) :=
{p : K × K //
canonicalKummerPolynomial25Three e p.1 p.2 = 0 ∧ p.1 = 0}
/-- Kummer solutions away from `r=0`. -/
abbrev KummerNonzeroSolutions (e : K) :=
{p : K × K //
canonicalKummerPolynomial25Three e p.1 p.2 = 0 ∧ p.1 ≠ 0}
/-- Split the dense solution set according to whether `r` vanishes. -/
def densePartitionEquiv (e : K) :
DenseSolutions e ≃ DenseZeroSolutions e ⊕ DenseNonzeroSolutions e where
toFun p := if h : p.1.1 = 0 then
Sum.inl ⟨p.1, p.2, h⟩
else Sum.inr ⟨p.1, p.2, h⟩
invFun p := p.elim (fun q => ⟨q.1, q.2.1⟩) (fun q => ⟨q.1, q.2.1⟩)
left_inv p := by
dsimp
split <;> rfl
right_inv p := by
rcases p with p | p
· simp [p.2.2]
· simp [p.2.2]
/-- Split the Kummer solution set according to whether `r` vanishes. -/
def kummerPartitionEquiv (e : K) :
KummerSolutions e ≃ KummerZeroSolutions e ⊕ KummerNonzeroSolutions e where
toFun p := if h : p.1.1 = 0 then
Sum.inl ⟨p.1, p.2, h⟩
else Sum.inr ⟨p.1, p.2, h⟩
invFun p := p.elim (fun q => ⟨q.1, q.2.1⟩) (fun q => ⟨q.1, q.2.1⟩)
left_inv p := by
dsimp
split <;> rfl
right_inv p := by
rcases p with p | p
· simp [p.2.2]
· simp [p.2.2]
/-- The dense zero fibre consists only of `(r,s)=(0,0)`. -/
def denseZeroEquivUnit {e : K} (he : IsCyclotomicFive25Three e) :
DenseZeroSolutions e ≃ Unit where
toFun _ := ()
invFun _ := ⟨(0, 0), (denseSegre_at_zero he).2 rfl, rfl⟩
left_inv p := by
rcases p with ⟨⟨r, s⟩, hcurve, hr⟩
change r = 0 at hr
subst r
change canonicalDenseSegreCubic25Three e 0 s = 0 at hcurve
have hs := (denseSegre_at_zero he).1 hcurve
change s = 0 at hs
subst s
rfl
right_inv p := by cases p; rfl
/-- The cleared Kummer zero fibre consists of `t=0` and the one extraneous
parameter introduced by clearing `r⁴`. -/
def kummerZeroEquivBool {e : K} (he : IsCyclotomicFive25Three e) :
KummerZeroSolutions e ≃ Bool where
toFun p := if p.1.2 = 0 then false else true
invFun b := if b then
⟨(0, extraneousKummerParameter25Three e),
(kummer_zero_fibre_iff he).2 (Or.inr rfl), rfl⟩
else ⟨(0, 0), (kummer_zero_fibre_iff he).2 (Or.inl rfl), rfl⟩
left_inv p := by
rcases p with ⟨⟨r, t⟩, hcurve, hr⟩
change r = 0 at hr
subst r
change canonicalKummerPolynomial25Three e 0 t = 0 at hcurve
rcases (kummer_zero_fibre_iff he).1 hcurve with ht | ht
· change t = 0 at ht
subst t
simp
· change t = extraneousKummerParameter25Three e at ht
subst t
simp [extraneousKummerParameter25Three_ne_zero he]
right_inv b := by
cases b <;> simp [extraneousKummerParameter25Three_ne_zero he]
/-- Away from `r=0`, the substitution `t=r²s` is an equivalence between
dense Segre solutions and Kummer solutions. -/
def denseNonzeroEquivKummerNonzero (e : K) :
DenseNonzeroSolutions e ≃ KummerNonzeroSolutions e where
toFun p := ⟨(p.1.1, p.1.1 ^ 2 * p.1.2), by
constructor
· have h := denseSegre_to_kummer_identity e p.1.1 p.1.2
rw [p.2.1, mul_zero] at h
exact h.symm
· exact p.2.2⟩
invFun p := ⟨(p.1.1, p.1.2 / p.1.1 ^ 2), by
constructor
· exact (denseSegre_eq_zero_iff_kummer e p.2.2).2 p.2.1
· exact p.2.2⟩
left_inv p := by
apply Subtype.ext
apply Prod.ext
· rfl
· exact mul_div_cancel_left₀ p.1.2 (pow_ne_zero 2 p.2.2)
right_inv p := by
apply Subtype.ext
apply Prod.ext
· rfl
· exact mul_div_cancel₀ p.1.2 (pow_ne_zero 2 p.2.2)
/-- Re-index Kummer pairs by the quotient parameter `t`, exposing the root
fibre used by the one-dimensional certificate. -/
def kummerSolutionsEquivSigma (e : K) :
KummerSolutions e ≃
(Σ t : K, {r : K // canonicalKummerPolynomial25Three e r t = 0}) where
toFun p := ⟨p.1.2, ⟨p.1.1, p.2⟩⟩
invFun p := ⟨(p.2.1, p.1), p.2.2⟩
left_inv p := rfl
right_inv p := by rcases p with ⟨t, r, h⟩; rfl
/-! ## Normalized charts on the Segre quadric -/
/-- First-nonzero charts on the two factors of the Segre quadric. The four
constructors represent the dense chart, the two boundary rulings, and their
intersection. -/
inductive NormalizedSegrePoint (K : Type*) where
| dense (r s : K)
| leftBoundary (s : K)
| rightBoundary (r : K)
| corner
deriving DecidableEq, Fintype
/-- Map the four normalized Segre charts through the inverse eigenbasis to
canonical homogeneous coordinates. -/
def normalizedSegreCoordinates (e : K) :
NormalizedSegrePoint K → Coordinates4 K
| .dense r s => canonicalSegreCoordinates25Three e 1 r 1 s
| .leftBoundary s => canonicalSegreCoordinates25Three e 0 1 1 s
| .rightBoundary r => canonicalSegreCoordinates25Three e 1 r 0 1
| .corner => canonicalSegreCoordinates25Three e 0 1 0 1
/-- The four-term cubic predicate on a normalized Segre chart. -/
def IsNormalizedSegreCurve (e : K) : NormalizedSegrePoint K → Prop
| .dense r s => canonicalDenseSegreCubic25Three e r s = 0
| .leftBoundary s => canonicalSegreCubic25Three e 0 1 1 s = 0
| .rightBoundary r => canonicalSegreCubic25Three e 1 r 0 1 = 0
| .corner => canonicalSegreCubic25Three e 0 1 0 1 = 0
/-- Normalized projective Segre parameters satisfying the curve equation. -/
abbrev SegreSolutions (e : K) :=
{p : NormalizedSegrePoint K // IsNormalizedSegreCurve e p}
/-- Curve points on the left boundary ruling. -/
abbrev LeftBoundarySolutions (e : K) :=
{s : K // canonicalSegreCubic25Three e 0 1 1 s = 0}
/-- Curve points on the right boundary ruling. -/
abbrev RightBoundarySolutions (e : K) :=
{r : K // canonicalSegreCubic25Three e 1 r 0 1 = 0}
/-- The normalized Segre solution set is the disjoint union of its dense
chart, two boundary rulings, and their corner. -/
def segreSolutionsEquivCharts (e : K) : SegreSolutions e ≃
DenseSolutions e ⊕
LeftBoundarySolutions e ⊕ RightBoundarySolutions e ⊕ Unit where
toFun p := by
rcases p with ⟨p, hp⟩
cases p with
| dense r s => exact Sum.inl ⟨(r, s), hp⟩
| leftBoundary s => exact Sum.inr (Sum.inl ⟨s, hp⟩)
| rightBoundary r => exact Sum.inr (Sum.inr (Sum.inl ⟨r, hp⟩))
| corner => exact Sum.inr (Sum.inr (Sum.inr ()))
invFun p := by
rcases p with p | p
· exact ⟨.dense p.1.1 p.1.2, p.2⟩
· rcases p with p | p
· exact ⟨.leftBoundary p.1, p.2⟩
· rcases p with p | p
· exact ⟨.rightBoundary p.1, p.2⟩
· exact ⟨.corner, segre_boundary_corner e⟩
left_inv p := by
rcases p with ⟨p, hp⟩
cases p <;> rfl
right_inv p := by
rcases p with p | p
· rfl
· rcases p with p | p
· rfl
· rcases p with p | p
· rfl
· cases p
rfl
/-- The left boundary ruling contributes one solution. -/
def leftBoundaryEquivUnit {e : K} (he : IsCyclotomicFive25Three e) :
LeftBoundarySolutions e ≃ Unit where
toFun _ := ()
invFun _ := ⟨0, (segre_boundary_a_zero he).2 rfl⟩
left_inv p := by
rcases p with ⟨s, hs⟩
have h := (segre_boundary_a_zero he).1 hs
subst s
rfl
right_inv p := by cases p; rfl
/-- The right boundary ruling contributes one solution. -/
def rightBoundaryEquivUnit {e : K} (he : IsCyclotomicFive25Three e) :
RightBoundarySolutions e ≃ Unit where
toFun _ := ()
invFun _ := ⟨0, (segre_boundary_c_zero he).2 rfl⟩
left_inv p := by
rcases p with ⟨r, hr⟩
have h := (segre_boundary_c_zero he).1 hr
subst r
rfl
right_inv p := by cases p; rfl
end Generic
/-! ## The `F81` projective count -/
/-! ## Restricting the projective eigenbasis to the curve
The preceding count lives naturally on the four normalized charts of
`P¹ × P¹`. This final structural layer identifies those charts with the
normalized determinant quadric in eigen-coordinates, restricts the
projective eigenbasis equivalence to the cubic equation, and thereby
transports the count to the stored canonical quadric-cubic model.
-/
section CanonicalRestriction
variable {K : Type*} [Field K] [CharP K 3] [DecidableEq K]
/-- The four Segre charts as normalized eigen-coordinates. They are the
rank-one matrices `(ac,ad;bc,bd)` after normalizing the first nonzero entry
of the two projective factors. -/
def normalizedSegreEigenPoint :
NormalizedSegrePoint K → NormalizedProjective4 K
| .dense r s => .xChart s r (r * s)
| .leftBoundary s => .zChart s
| .rightBoundary r => .yChart 0 r
| .corner => .wChart
/-- The determinant equation `u₂u₃=u₁u₄` on a first-nonzero normalized
eigenvector. This is the diagonalized canonical quadric. -/
def IsNormalizedEigenQuadric (P : NormalizedProjective4 K) : Prop :=
let U := normalizedCoordinatesThree P
U.y * U.z - U.x * U.w = 0
/-- The normalized Segre charts parametrize the normalized determinant
quadric bijectively. The inverse reads off the dense chart, then the two
boundary rulings, with the determinant equation removing the redundant
coordinate in each case. -/
def normalizedSegreEquivEigenQuadric :
NormalizedSegrePoint K ≃
{P : NormalizedProjective4 K // IsNormalizedEigenQuadric P} where
toFun p := by
refine ⟨normalizedSegreEigenPoint p, ?_⟩
cases p <;>
simp [normalizedSegreEigenPoint, IsNormalizedEigenQuadric,
normalizedCoordinatesThree]
all_goals ring
invFun P := by
rcases P with ⟨P, hP⟩
cases P with
| xChart y z w => exact .dense z y
| yChart z w => exact .rightBoundary w
| zChart w => exact .leftBoundary w
| wChart => exact .corner
left_inv p := by
cases p <;> rfl
right_inv P := by
rcases P with ⟨P, hP⟩
apply Subtype.ext
cases P with
| xChart y z w =>
change y * z - 1 * w = 0 at hP
simp only [normalizedSegreEigenPoint]
congr
linear_combination hP
| yChart z w =>
change 1 * z - 0 * w = 0 at hP
have hz : z = 0 := by simpa using hP
subst z
rfl
| zChart w => rfl
| wChart => rfl
omit [CharP K 3] [DecidableEq K] in
/-- Scaling a homogeneous vector multiplies the canonical quadric by the
square of the scalar. This is the projective invariance needed after
first-nonzero normalization. -/
theorem canonicalQuadric25Three_scale (a : K) (P : Coordinates4 K) :
canonicalQuadric25Three (scaleCoordinatesThree a P) =
a ^ 2 * canonicalQuadric25Three P := by
rcases P with ⟨x, y, z, w⟩
simp only [scaleCoordinatesThree, canonicalQuadric25Three]
ring
omit [CharP K 3] [DecidableEq K] in
/-- Scaling a homogeneous vector multiplies the canonical cubic by the
cube of the scalar. Hence its zero locus is unchanged by projective
normalization through a nonzero scalar. -/
theorem canonicalCubic25Three_scale (a : K) (P : Coordinates4 K) :
canonicalCubic25Three (scaleCoordinatesThree a P) =
a ^ 3 * canonicalCubic25Three P := by
rcases P with ⟨x, y, z, w⟩
simp only [scaleCoordinatesThree, canonicalCubic25Three]
ring
/-- The curve in normalized eigen-coordinates: the determinant quadric
together with the canonical cubic pulled back through the inverse
eigenbasis. -/
def IsNormalizedEigenCurve (e : K)
(P : NormalizedProjective4 K) : Prop :=
let U := normalizedCoordinatesThree P
U.y * U.z - U.x * U.w = 0 ∧
canonicalCubic25Three
(canonicalCoordinatesFromEigen25Three e U.x U.y U.z U.w) = 0
/-- Projectivizing the inverse eigenbasis transports the determinant-cubic
curve exactly to the normalized canonical quadric-cubic intersection.
The only scalars introduced are the nonzero normalization scalar, the
nonzero quadric multiplier, and their homogeneous powers. -/
theorem isNormalizedEigenCurve_iff_canonical
(e : K) (he : IsCyclotomicFive25Three e)
(P : NormalizedProjective4 K) :
IsNormalizedEigenCurve e P ↔
IsCanonicalNormalizedThree (eigenToCanonicalProjective e P) := by
let U := normalizedCoordinatesThree P
let V := canonicalCoordinatesFromEigen25Three e U.x U.y U.z U.w
have hU : U ≠ zeroCoordinatesThree :=
normalizedCoordinatesThree_ne_zero P
have hV : V ≠ zeroCoordinatesThree :=
canonicalCoordinatesFromEigen_ne_zero he hU
rcases normalizeCoordinatesThree_spec hV with ⟨a, ha, hspec⟩
have hcoords :
normalizedCoordinatesThree (eigenToCanonicalProjective e P) =
scaleCoordinatesThree a V := by
exact hspec
have hquad :
canonicalQuadric25Three
(normalizedCoordinatesThree (eigenToCanonicalProjective e P)) = 0 ↔
U.y * U.z - U.x * U.w = 0 := by
rw [hcoords, canonicalQuadric25Three_scale,
canonicalQuadric25_fromEigen he]
simp [ha, canonicalSegreQuadricScalar25Three_ne_zero he]
have hcubic :
canonicalCubic25Three
(normalizedCoordinatesThree (eigenToCanonicalProjective e P)) = 0 ↔
canonicalCubic25Three V = 0 := by
rw [hcoords, canonicalCubic25Three_scale]
simp [ha]
unfold IsNormalizedEigenCurve IsCanonicalNormalizedThree
change (U.y * U.z - U.x * U.w = 0 ∧ canonicalCubic25Three V = 0) ↔ _
rw [hquad, hcubic]
omit [DecidableEq K] in
/-- On every rank-one eigenvector, vanishing of the pulled-back canonical
cubic is equivalent to vanishing of the four-term Segre cubic. The
comparison scalar `e³+1` is nonzero on the fifth-cyclotomic locus. -/
theorem canonicalCubic25_fromEigen_eq_zero_iff_segre
{e a b c d : K} (he : IsCyclotomicFive25Three e) :
canonicalCubic25Three
(canonicalCoordinatesFromEigen25Three e
(a * c) (a * d) (b * c) (b * d)) = 0 ↔
canonicalSegreCubic25Three e a b c d = 0 := by
have h := canonicalCubic25_segre
(e := e) (a := a) (b := b) (c := c) (d := d) he
change canonicalCubic25Three
(canonicalCoordinatesFromEigen25Three e
(a * c) (a * d) (b * c) (b * d)) =
(e ^ 3 + 1) * canonicalSegreCubic25Three e a b c d at h
rw [h]
simp [cyclotomicFive25Three_cubicScalar_ne_zero he]
omit [DecidableEq K] in
/-- The normalized four-chart Segre curve is exactly the determinant-cubic
curve in normalized eigen-coordinates. Each chart is a rank-one matrix;
the determinant vanishes identically and the preceding scalar comparison
handles the cubic. -/
theorem isNormalizedSegreCurve_iff_eigenCurve
{e : K} (he : IsCyclotomicFive25Three e)
(p : NormalizedSegrePoint K) :
IsNormalizedSegreCurve e p ↔
IsNormalizedEigenCurve e (normalizedSegreEigenPoint p) := by
cases p with
| dense r s =>
have hc := canonicalCubic25_fromEigen_eq_zero_iff_segre
(e := e) (a := 1) (b := r) (c := 1) (d := s) he
change canonicalSegreCubic25Three e 1 r 1 s = 0 ↔
s * r - 1 * (r * s) = 0 ∧
canonicalCubic25Three
(canonicalCoordinatesFromEigen25Three e 1 s r (r * s)) = 0
constructor
· intro h
exact ⟨by ring, by simpa using hc.mpr h⟩
· intro h
exact hc.mp (by simpa using h.2)
| leftBoundary s =>
have hc := canonicalCubic25_fromEigen_eq_zero_iff_segre
(e := e) (a := 0) (b := 1) (c := 1) (d := s) he
change canonicalSegreCubic25Three e 0 1 1 s = 0 ↔
0 * 1 - 0 * s = 0 ∧
canonicalCubic25Three
(canonicalCoordinatesFromEigen25Three e 0 0 1 s) = 0
constructor
· intro h
exact ⟨by ring, by simpa using hc.mpr h⟩
· intro h
exact hc.mp (by simpa using h.2)
| rightBoundary r =>
have hc := canonicalCubic25_fromEigen_eq_zero_iff_segre
(e := e) (a := 1) (b := r) (c := 0) (d := 1) he
change canonicalSegreCubic25Three e 1 r 0 1 = 0 ↔
1 * 0 - 0 * r = 0 ∧
canonicalCubic25Three
(canonicalCoordinatesFromEigen25Three e 0 1 0 r) = 0
constructor
· intro h
exact ⟨by ring, by simpa using hc.mpr h⟩
· intro h
exact hc.mp (by simpa using h.2)
| corner =>
have hc := canonicalCubic25_fromEigen_eq_zero_iff_segre
(e := e) (a := 0) (b := 1) (c := 0) (d := 1) he
change canonicalSegreCubic25Three e 0 1 0 1 = 0 ↔
0 * 0 - 0 * 1 = 0 ∧
canonicalCubic25Three
(canonicalCoordinatesFromEigen25Three e 0 0 0 1) = 0
constructor
· intro h
exact ⟨by ring, by simpa using hc.mpr h⟩
· intro h
exact hc.mp (by simpa using h.2)
/-- The pulled-back canonical cubic on normalized eigen-coordinates. It is
kept separate here because the determinant equation is already stored in
the subtype produced by the Segre-quadric equivalence. -/
def IsNormalizedEigenCubic (e : K)
(P : NormalizedProjective4 K) : Prop :=
let U := normalizedCoordinatesThree P
canonicalCubic25Three
(canonicalCoordinatesFromEigen25Three e U.x U.y U.z U.w) = 0
/-- Inside the determinant quadric, the four-term Segre cubic cuts out the
same locus as the pulled-back canonical cubic. -/
def segreSolutionsEquivEigenQuadricCubic
{e : K} (he : IsCyclotomicFive25Three e) :
SegreSolutions e ≃
{Q : {P : NormalizedProjective4 K // IsNormalizedEigenQuadric P} //
IsNormalizedEigenCubic e Q.1} :=
normalizedSegreEquivEigenQuadric.subtypeEquiv (fun p => by
change IsNormalizedSegreCurve e p ↔
IsNormalizedEigenCubic e (normalizedSegreEigenPoint p)
have h := isNormalizedSegreCurve_iff_eigenCurve he p
constructor
· intro hp
exact (h.mp hp).2
· intro hc
exact h.mpr ⟨(normalizedSegreEquivEigenQuadric p).2, hc⟩)
/-- Flatten the nested “point on the quadric, then on the cubic” subtype to
the conjunction defining the normalized eigen-coordinate curve. -/
def eigenQuadricCubicEquivEigenCurve (e : K) :
{Q : {P : NormalizedProjective4 K // IsNormalizedEigenQuadric P} //
IsNormalizedEigenCubic e Q.1} ≃
{P : NormalizedProjective4 K // IsNormalizedEigenCurve e P} where
toFun Q := ⟨Q.1.1, Q.1.2, Q.2⟩
invFun P := ⟨⟨P.1, P.2.1⟩, P.2.2⟩
left_inv _ := rfl
right_inv _ := rfl
/-- The normalized Segre curve and the normalized eigen-coordinate curve are
equivalent as finite types, not merely related by a surjection. -/
def segreSolutionsEquivEigenCurve
{e : K} (he : IsCyclotomicFive25Three e) :
SegreSolutions e ≃
{P : NormalizedProjective4 K // IsNormalizedEigenCurve e P} :=
(segreSolutionsEquivEigenQuadricCubic he).trans
(eigenQuadricCubicEquivEigenCurve e)
/-- Restrict the projective eigenbasis equivalence to the curve equations.
Homogeneity ensures that first-nonzero renormalization preserves both loci. -/
def eigenCurveEquivCanonicalCurve
(e : K) (he : IsCyclotomicFive25Three e) :
{P : NormalizedProjective4 K // IsNormalizedEigenCurve e P} ≃
{P : NormalizedProjective4 K // IsCanonicalNormalizedThree P} :=
(eigenCanonicalProjectiveEquiv e he).subtypeEquiv
(isNormalizedEigenCurve_iff_canonical e he)
/-- Composite structural equivalence from the normalized `P¹ × P¹` Segre
model to the canonical normalized quadric-cubic model. -/
def segreSolutionsEquivCanonicalCurve
(e : K) (he : IsCyclotomicFive25Three e) :
SegreSolutions e ≃
{P : NormalizedProjective4 K // IsCanonicalNormalizedThree P} :=
(segreSolutionsEquivEigenCurve he).trans
(eigenCurveEquivCanonicalCurve e he)
end CanonicalRestriction
/-- The executable ternary-table predicate and the actual field-valued
canonical predicate define equivalent point types over `F81`. -/
def f81ExecutableCanonicalEquivActual :
{P : NormalizedProjective4 F81 //
IsCanonicalNormalized25Ternary f81Operations P} ≃
{P : NormalizedProjective4 F81 // IsCanonicalNormalizedThree P} :=
Equiv.subtypeEquivRight isCanonicalNormalized25Ternary_f81_iff
end
end MazurProof.RationalPointsN25QuotientKummerThreeProjective
Source
B. Mazur, Modular curves and the Eisenstein ideal, Publ. IHES 47 (1977); Lean formalization of the order-25 case for this mission. Modules in the MazurProof/Keystone/Tate families are adapted from anthropics/fermats-last-theorem (Apache-2.0) via the xiangyazi24/FLT fork.