The threefold wall transport, on ℙ³ #
Stability/ThreefoldWallTransport.lean carries a polarised threefold class into
the four-coordinate (α, β) charge plane, and Models/Threefold/ProjectiveSpace.lean
builds the ℙ³ model. Nothing connected them: the model was referenced only
inside its own file, and no polarization was defined for it at all. So the
threefold charge arithmetic had no worked example, exactly as the surface
arithmetic had none before Examples/Surface/RankOneRealization.lean.
This file supplies the polarization and evaluates the transport.
The coordinates #
With ∫H³ = 1 and linear-section coordinates (a, b, c, e), the four degrees
come out as
∫H³ch₀ = a, ∫H²ch₁ = b, ∫H·ch₂ = -b/2 + c, ∫ch₃ = b/6 - c + e,
which is threefoldChCoeff read against ∫H³ = 1. The four
p3_toNumClass_deg* lemmas are that computation, and everything else in the file is a consequence of it together
with the general theorems of the transport.
The Bayer--Macrì--Toda datum, and what a witness can honestly be #
BMTData has no witness anywhere in the tree, so every consequence of it is
vacuous. It cannot be given an honest one by proving the inequality: the
conjecture is false for some threefolds — the blow-up of ℙ³ at a point,
Schmidt — and true for ℙ³ itself only by a real theorem this repository does
not contain.
p3BMTSanity is therefore a sanity witness in the exact style of
k3BogomolovSanity: its TiltSemistable predicate is defined to be the
conclusion, so nonneg holds tautologously. It asserts nothing whatsoever
about sheaves or about tilt stability, and must not be read as the
Bayer--Macrì--Toda inequality for ℙ³. What it buys is that the transported
statements below are about an inhabited structure.
The polarization #
∫_{ℙ³} H³ = 1.
The hyperplane class as a polarization of ℙ³. The first polarization
of a threefold model in the repository.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The transported class #
The four degrees of a class on ℙ³ #
The transport of (a, b, c, e) is (a, b, -b/2 + c, b/6 - c + e). These four
lemmas are the computation the rest of the file rests on.
Not a simp lemma: the generic Threefold.toNumClass_deg* projections are
already simp and rewrite the left-hand side first, so the normal-form linter
rejects the attribute. p3_reZ and p3_imZ name these explicitly.
Not a simp lemma: the generic Threefold.toNumClass_deg* projections are
already simp and rewrite the left-hand side first, so the normal-form linter
rejects the attribute. p3_reZ and p3_imZ name these explicitly.
Not a simp lemma: the generic Threefold.toNumClass_deg* projections are
already simp and rewrite the left-hand side first, so the normal-form linter
rejects the attribute. p3_reZ and p3_imZ name these explicitly.
Not a simp lemma: the generic Threefold.toNumClass_deg* projections are
already simp and rewrite the left-hand side first, so the normal-form linter
rejects the attribute. p3_reZ and p3_imZ name these explicitly.
The charge family on the model #
The threefold charge family of ℙ³, indexed by (α, β).
Equations
- One or more equations did not get rendered due to their size.
Instances For
The real part of the ℙ³ charge, in linear-section coordinates.
The imaginary part of the ℙ³ charge, in linear-section coordinates.
A sanity witness for the Bayer--Macrì--Toda datum #
A sanity witness, not geometry.
TiltSemistable is taken to be the conclusion itself, so nonneg holds
tautologously. This asserts nothing whatsoever about sheaves or about tilt
stability: it exists only to show BMTData is inhabitable and that the
transported theorems fire on a concrete model. Do not read it as the
Bayer--Macrì--Toda inequality for ℙ³, which is a real theorem this repository
does not contain, nor as evidence for the conjecture, which is false for some
threefolds.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The transported inequality, on the model. Its content is exactly the
content of p3BMTSanity, which is none.