2021-04-29_2 Yves Andre On the canonical, fpqc and finite topologies

Yves André (CNRS), “On the canonical, fpqc and finite topologies: classical questions, new answers (and conversely)”

Tags: #seminar_notes #blog

Reference: Yves André (CNRS), “On the canonical, fpqc and finite topologies: classical questions, new answers (and conversely)”. Princeton/IAS NT seminar.

attachments/image_2021-04-29-16-36-39.png
  • What does it mean for an algebra to be faithfully flat over another algebra?

  • p-adic Hilbert functor : see Bhatt-Lurie.

    • Use this to get “almost” results, then use prismatic cohomology]] techniques (where one has Frobenius) to remove the “almost”.
  • Extracting \(p\)th roots? Passing from \(k[g_1, \cdots, g_n]\) to \(k[g_1^{1/p}, \cdots, g_n^{1/p}]\), I think…

  • F-pure and strongly F-regular singularities are characteristic \(p\) analogs of log-canonical and log-terminal singularities in the minimal model program.

  • Tilting]] : pass from mixed characteristic to characteristic \(p\). Try to use simpler proofs/theorems from characteristic \(p\) situation.

    • Going forward: some limiting process after inverting \(p\)..? Going backward: take Witt Vectors]].
  • What properties of schemes descend along faithfully flat morphism? See EGA. However, what properties descend for the fpqc topology?

    • What is a faithfully flat morphism?
  • See Faltings almost purity theorem.

  • Commutative algebra: see excellent regular domains, integral vs algebraic closures. - Cohen-Macaulay rings and modules

  • Can have \(\mathrm{\operatorname{fpqc}}\) coverings that are not fppf coverings.

  • What is a regular scheme?

    • Theorem: any finite covering of a regular scheme is an \(\mathrm{\operatorname{fpqc}}\) covering.

    • Very nontrivial in characteristic zero.

    • Noether normalization can show some finite coverings of \({\mathbb{A}}^3_{/k}\) are not \(\mathrm{\operatorname{fppf}}\) coverings.

  • Sometimes local or coherent cohomology classes

  • Grothendieck’s descent?

    faithfully flat implies something is an equivalence.

#seminar_notes #blog