The blow-up of the origin restricted to the cusp #
"The restriction of a smooth blow-up sequence need not be a smooth blow-up sequence" [Kol07, 30.2].
On the model of HironakaExamples/Sequence/CuspModel.lean (X = 𝔸²_k, Z = V(x, y),
S = V(y² − x³)):
isSmooth_seqOrigin: blowing up the origin is a smooth blow-up sequence; its centerV(x, y) ≅ Spec kis smooth overk(smooth_origin_subschemeι: the subscheme isSpec (k[x, y]/(x, y))andk → k[x, y]/(x, y)is an isomorphism, every polynomial being congruent to its constant term), the ambient𝔸²_kbeing smooth overk(smooth_affineSpaceToSpec_two);not_smooth_cusp: the restricted sequence starts atS, which is not smooth overk. If it were, the local ring𝒪_{S,0} = 𝒪_{𝔸²,0}/(y² − x³)would be regular; but𝒪_{𝔸²,0}is regular,y² − x³is a nonzero element of𝔪²at the origin (x, y ∈ 𝔪), and the quotient of a regular local ring by a nonzero element of𝔪²is never regular (HironakaExamples/Local/SingularQuotient.lean).
The predicate IsSmooth is Kollár's clause on the centers; the ambient clause of
[Kol07, Notation 19], the smoothness of the stages, is the one that fails here.
The affine plane is smooth over k (of relative dimension 2).
1 ∉ (x, y).
Every polynomial is congruent to its constant term modulo (x, y).
The composite k → k[x, y] → k[x, y]/(x, y) is bijective.
The origin V(x, y) ⊂ 𝔸²_k is smooth over k: its subscheme is Spec (k[x, y]/(x, y)), and
k → k[x, y]/(x, y) is an isomorphism.
Blowing up the origin of 𝔸²_k is a smooth blow-up sequence.
The cuspidal cubic is not smooth over k: the scheme at which the restricted sequence of
[Kol07, 30.2] starts is singular.