The threefold wall transport, on the quintic #
Examples/Threefold/ProjectiveSpaceWalls.lean connected the transport of
Stability/ThreefoldWallTransport.lean to the ℙ³ model. This file does the
same for the quintic Calabi--Yau threefold, which was in the same position: a
model referenced only inside its own file, with no polarization defined for it.
Why the second threefold is worth doing #
ℙ³ and the quintic differ in exactly the two places the transport touches.
- The degree.
∫H³ = 5rather than1, so every one of the four degrees carries a factor theℙ³computation could not see. A transport that silently assumed unit degree would pass onℙ³and fail here. - The Todd class. The quintic has
td₁ = 0and∫td₃ = χ(O) = 0, whereℙ³has every coefficient nonzero. The transport does not read the Todd class, and these statements are the check that it does not: the four degrees below depend on the Chern coefficients and the degree alone.
The coordinates #
With ∫H³ = 5 and linear-section coordinates (a, b, c, e),
∫H³ch₀ = 5a, ∫H²ch₁ = 5b, ∫H·ch₂ = 5(-b/2 + c), ∫ch₃ = 5b/6 - 5c + e,
the last because threefoldChCoeff 5 divides the point coordinate by the
degree, which the integral then undoes. That cancellation is the one place the
quintic computation is not the ℙ³ one with a factor of five.
No second sanity witness #
ProjectiveSpaceWalls.p3BMTSanity already shows BMTData is inhabitable, and a
second tautological witness would add nothing. quintic_Q_toNumClass_nonneg
therefore takes a supplied BMTData as a hypothesis, which is the honest shape:
the Bayer--Macrì--Toda inequality is not proved here for the quintic or for
anything else, and it is false for some threefolds.
The polarization #
∫_X H³ = 5 on the quintic.
The hyperplane class as a polarization of the quintic.
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 the quintic #
The transport of (a, b, c, e) is (5a, 5b, 5(-b/2 + c), 5b/6 - 5c + e).
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. quintic_reZ and quintic_imZ name these explicitly.
Not a simp lemma, for the reason given at quintic_toNumClass_deg0.
Not a simp lemma, for the reason given at quintic_toNumClass_deg0.
Not a simp lemma, for the reason given at quintic_toNumClass_deg0.
The degree cancels the /5 that threefoldChCoeff 5 puts on the point
coordinate, which is why this one is not the ℙ³ answer times five.
The charge family on the model #
The threefold charge family of the quintic, indexed by (α, β).
Equations
- One or more equations did not get rendered due to their size.
Instances For
The real part of the quintic charge, in linear-section coordinates.
The imaginary part of the quintic charge, in linear-section coordinates.
The transported Bayer--Macrì--Toda inequality #
The transported inequality on the quintic, taking the datum as a hypothesis.
No witness is constructed here. ProjectiveSpaceWalls.p3BMTSanity already
shows the structure is inhabitable, and a second tautological witness would add
nothing. The inequality itself is not proved for the quintic or for anything
else, and it is false for some threefolds.