The marked transform of a smaller ideal is contained in the birational transform #
In the proof of Theorem 107, "if ord N(I) ≥ m, we can apply order reduction (68) to N(I),
until its order drops below m" [Kol07, 111, Step 1]: a round of order reduction for the
nonmonomial part N(I) at d = max-ord N(I) ≥ m is a sequence of order d ≥ m for N(I),
hence of order ≥ m for (I, m), since ord_Z I ≥ ord_Z N(I). At the later stages of the round
the center Z_i has ord_{Z_i} J_i = d for the birational transform J_i of J = N(I), and the
marked transform I_i of (I, m) must have ord_{Z_i} I_i ≥ m; this module supplies the
per-stage inclusion I_i ⊆ J_i. One step is
π^{-1}_*(I, m) = (π^* I : F^m) ⊆ (π^* J : F^m) ⊆ (π^* J : F^d) = π^{-1}_* J for I ⊆ J,
m ≤ d = ord_Z J (the colon is monotone in the ideal and antitone in the divisor; the birational
transform of J is its marked transform at the exact order d,
weakTransform_eq_markedTransform_of_smooth), and ord_{Z_i} I_i ≥ ord_{Z_i} J_i = d ≥ m
follows because the order is antitone in the ideal.
markedTransform_le_weakTransform_of_le: one blow-up.markedTransformSeq_le_weakTransformSeq_of_le: along a smooth blow-up sequence of orderdforJ, at every stage.isOrderGeSeq_of_isOrderSeq_of_le: a smooth blow-up sequence of orderd ≥ mfor(X, J, E)is a smooth blow-up sequence of order≥ mfor(X, I, m, E)wheneverI ⊆ J(same centers, same boundaries).
For I ⊆ J, m ≤ d and a smooth center D with ord_D J = d, the marked transform
π^{-1}_*(I, m) = (π^* I : F^m) of [Kol07, Definition 60] is contained in the birational
transform π^{-1}_* J = (π^* J : F^d) of [Kol07, 58].
Along a smooth blow-up sequence of order d ≥ m for (X, J, E), the marked transforms of
(I, m) are contained in the birational transforms of J at every stage whenever I ⊆ J; the
per-stage form of "ord_Z I ≥ ord_Z N(I)" in [Kol07, 111, Step 1].
A smooth blow-up sequence of order d ≥ m for (X, J, E) is a smooth blow-up sequence of
order ≥ m for (X, I, m, E) whenever I ⊆ J: the same centers, smooth and with simple normal
crossings with the same boundaries, and ord_{Z_i} I_i ≥ ord_{Z_i} J_i = d ≥ m at every stage by
the per-stage inclusion. This is how the order reduction of N(I) in [Kol07, 111, Step 1] is a
sequence of order ≥ m for (I, m).