The coordinate swap between two charts at a point #
For two charts e, e' of the maximal atlas centred at p, the coordinate swap φ_p = e'⁻¹ ∘ e
is the chart change of Hironaka/Manifold/Chart/Atlas.lean (PartialDiffeomorph.chartChange),
an analytic isomorphism from U(p) = e.source ∩ e⁻¹(e'.target) onto
V(p) = e'.source ∩ e'⁻¹(e.target) — two neighbourhoods of p, in general distinct — with
e' ∘ φ_p = e (IsCoordinateSwap). It fixes p, pulls the coordinate functions of e' back to
those of e (Kollár's φ^*(x'_1, x_2, …, x_n) = (x_1, x_2, …, x_n) [Kol07, 95]), induces the
pullback of germs φ_p^* : 𝒪_{M,p} →+* 𝒪_{M,p} — the stalk map germMapOn of an analytic map
fixing p — and satisfies the Taylor identity T_e ∘ φ_p^* = T_{e'}: the e-expansion of
g ∘ φ_p is the e'-expansion of g, because (g ∘ φ_p) ∘ e⁻¹ = g ∘ e'⁻¹ near 0 = e p = e' p,
so T_e ∘ φ_p^* satisfies the specification IsTaylorHom of the Taylor homomorphism of the chart
e' and equals it by uniqueness (IsTaylorHom.eq). Kollár's formal automorphism φ^* of
𝒪̂_{M,p} ≅ 𝕜[[x]] [Kol07, 95] is T_{e'} ∘ T_e⁻¹, the composite of the two completion
isomorphisms taylorCompletionEquiv; with the Taylor identity this reads T_e ∘ φ_p^* = φ^* ∘ T_e.
Włodarczyk's coordinate change φ_{uv} between two tangent directions is the same construction
[Wlo09, Lemma 5.5.3, proof, step (0)]. It is used to compare hypersurfaces of maximal contact
given by different charts (Hironaka/Resolution/Analytic/MaximalContact/SwapRealizes.lean).
The chart change of two charts centred at p is their coordinate swap in the sense of the
specification IsCoordinateSwap.
For two charts of the maximal atlas centred at p, the coordinate swap φ_p = e'⁻¹ ∘ e exists
as an analytic isomorphism between two neighbourhoods of p.
The coordinate swap maps its source into the source of e'.
The coordinate swap of two charts centred at p fixes p.
V(p) = Φ.target is a neighbourhood of p.
The coordinate functions of e' pull back along the swap to those of e (Kollár's
φ^*(x'_1, x_2, …, x_n) = (x_1, x_2, …, x_n), [Kol07, 95]).
The pullback of germs along the coordinate swap: the stalk map of an analytic map
(germMapOn) for φ_p, which fixes p.
The pullback of germs is determined by its specification (stalkToGerm is injective).
T_e ∘ φ_p^* = T_{e'}, the primitive form of Kollár's T_p ∘ φ_p^* = φ^* ∘ T_p [Kol07, 95]:
the e-expansion of g ∘ φ_p is the e'-expansion of g, both at 0 = e p = e' p.
Kollár's formal automorphism φ of 𝒪̂_p ≅ 𝕜[[x]] [Kol07, 95]: the composite T_{e'} ∘ T_e⁻¹
of the two completion isomorphisms carries the e-expansion of every germ to its
e'-expansion.