Logarithmic derivations in coordinates #
[Kol07, 87]: in local coordinates with S = (x₁ = 0), the sheaf Der_X(−log S) of logarithmic
derivations is generated by x₁ ∂₁, ∂₂, …, ∂ₙ, and D(−log S)(I) is generated by I and their
values on I. Here the coordinate system is RegularCoords on a regular local ring R, with the
hypotheses of Hironaka/Algebra/Local/CoordsOver.lean that the coordinate derivations ∂ᵢ are
k-linear (IsLinearOver) and span the k-derivations (SpansDerivations: every k-derivation
is δ = ∑ᵢ δ(xᵢ) ∂ᵢ, eq_sum_smul_pderivOver), and S = V(x_h).
- The
k-derivations preserving(x_h)are theR-span ofx_h ∂_hand the∂_j,j ≠ h(preserving_span_x_eq). The generators preserve(x_h)(x_smul_pderivOver_preservesIdeal,pderivOver_preservesIdeal_of_ne, fromHironaka/Algebra/Local/LogDeriv.lean); converselyδ = ∑ᵢ δ(xᵢ) ∂ᵢwithδ(x_h) = b · x_h, soδ = b · (x_h ∂_h) + ∑_{j ≠ h} δ(x_j) ∂_j. - Kollár's ideal
D(−log (x_h))(I)over allk-derivations preserving(x_h)(Ideal.logDerivative,Hironaka/Scheme/IdealSheaf/Derivative/Logarithmic.lean) is the coordinate logarithmic derivative idealc.Dlog h IofHironaka/Algebra/Local/LogDeriv.lean(logDerivative_eq_Dlog), and the iterates agree (logDerivativeIter_eq_Dlogpow), just asderivative_eq_DofHironaka/Algebra/Derivative/Basic.leanidentifiesD(I)withc.D I.
Used for the stalks of the logarithmic derivative sheaves on manifolds
(Hironaka/Manifold/IdealSheaf/LogDerivLemmas.lean).
x_i ∂_i preserves (x_h) for every i (x_smul_pderiv_mem_logDerE,
Hironaka/Algebra/Local/LogDeriv.lean).
∂_j preserves (x_h) for j ≠ h (pderiv_mem_span_x_of_ne').
A k-derivation is the R-combination ∑ᵢ δ(xᵢ) ∂ᵢ of the k-linear coordinate
derivations, when these span the k-derivations.
[Kol07, 87], Der_X(−log S) = ⟨x₁ ∂/∂x₁, ∂/∂x₂, …, ∂/∂xₙ⟩ in coordinates with S = (x₁ = 0):
the k-derivations preserving (x_h) are the R-span of x_h ∂_h and the ∂_j, j ≠ h.
The ideal D(−log (x_h))(I) of [Kol07, 87] over all k-derivations preserving (x_h) is the
coordinate logarithmic derivative ideal c.Dlog h I of Hironaka/Algebra/Local/LogDeriv.lean.
The iterates: D^s(−log (x_h))(I) = c.Dlogpow h s I.