Transitions between the abstract blow-up charts #
The blowing-up π : M' → M of a closed submanifold is characterised abstractly (IsBlowUp): over
an adapted chart φ of the centre, blow-up charts Φ_i of every index exist (exists_chart) and
cover π⁻¹(φ.source) (cover), each reading ψ ∘ φ ∘ π = π_i ∘ ψ ∘ Φ_i (comm) with the model
chart domain as target (mem_target_iff). Nothing says directly which points of M' lie in which
chart. This module derives the transition rule from the axioms — Kollár's proof of the uniqueness of
blow-up sequences needs the two lifts f'(q), g'(q) to be read in one chart of Bl_Y M
[Kol07, Theorem 97, proof]:
IsBlowUpChart.mem_source_of_mem_transitionDomain: ifp ∈ Φ_i.sourceand the chart coordinatesψ (Φ_i p)lie in the transition domainT_ik(u_{σ k} ≠ 0, ori = k), thenp ∈ Φ_k.sourceandψ (Φ_k p) = T_ik (ψ (Φ_i p))— the abstract charts are glued by the transition maps of the model blow-up (blowUpTransition), the chart changes of Bierstone–Milman's description of the blowing-up [BM88, Definition 4.1].
The proof is the uniqueness of lifts (IsBlowUp.eqOn_of_comp_eq, from the density of the
complement of the exceptional divisor): the candidate q ↦ Φ_k⁻¹(ψ⁻¹(T_ik(ψ(Φ_i q)))) is a
continuous map over π on the open set of transverse points, so it is the identity there.
A point of the blow-up chart of index i whose coordinates lie in the transition domain
T_ik (u_{σ k} ≠ 0, or i = k) lies in the blow-up chart of index k over the same adapted
chart, with coordinates T_ik of the old ones (the chart changes of the blowing-up,
[BM88, Definition 4.1]). By the uniqueness of lifts (IsBlowUp.eqOn_of_comp_eq).