Zulip Chat Archive
Stream: Is there code for X?
Topic: cohomology of the shift is shift of the cohomology
Kevin Buzzard (Dec 18 2024 at 17:48):
Richard Hill asked me this in our last London Learning Lean meeting:
import Mathlib
variable (C : CochainComplex Ab ℤ)
example : C⟦(2 : ℤ)⟧.homology n ≅ C.homology (n + 2) := by
sorry
Is this easy with our current definitions?
Joël Riou (Dec 18 2024 at 17:54):
This is in the file Algebra.Homology.HomotopyCategory.ShiftSequence
.
Adam Topaz (Dec 18 2024 at 17:58):
import Mathlib
variable (C : CochainComplex Ab ℤ)
open CategoryTheory
noncomputable
example : C⟦(2 : ℤ)⟧.homology n ≅ C.homology (n + 2) :=
CochainComplex.ShiftSequence.shiftIso _ _ _ _ (Int.add_comm _ _) |>.app _
Adam Topaz (Dec 18 2024 at 17:59):
Probably just beyond what exact?
can currently do :-/
Kevin Buzzard (Dec 18 2024 at 18:00):
But not beyond #Is there code for X? ! Thanks a lot!
Joël Riou (Dec 18 2024 at 18:04):
The preferred phrasing should be:
open HomologicalComplex
noncomputable example {A : Type*} [Category A] [Abelian A] (C : CochainComplex A ℤ) (n : ℤ) :
C⟦(2 : ℤ)⟧.homology n ≅ C.homology (n + 2) :=
((homologyFunctor _ _ 0).shiftIso _ _ _ (add_comm 2 n)).app C
Joël Riou (Dec 18 2024 at 18:16):
The notion Functor.ShiftSequence
which is used here is not in the mathematical literature. It means that we attach to the functor homologyFunctor _ _ 0 : CochainComplex Ab ℤ ⥤ Ab
the whole series of functors homologyFunctor _ _ n : CochainComplex Ab ℤ ⥤ Ab
for all n : ℤ
and we record isomorphisms about their behaviour with respect to the shift functors (and compatibilities of these isomorphisms). This is particularly useful in the triangulated context (e.g. the homotopy category, or the derived category): when we have a homological functor like F : DerivedCategory A ⥤ C
, we should get long exact sequences like FX ⟶ FY ⟶ FZ ⟶ F(X⟦1⟧)
for any distinguished triangle, but in the case F := DerivedCategory.homologyFunctor _ 0
, that very exact sequence is H^0 X ⟶ H^0 Y ⟶ H^0 Z ⟶ H^0(X⟦1⟧)
. The use of this Functor.ShiftSequence
structure allows to get a better looking long exact sequence ... ⟶ H^0 X ⟶ H^0 Y ⟶ H^0 Z ⟶ H^1 X ⟶ ...
.
Richard Hill (Dec 18 2024 at 18:23):
Thanks, that's really useful.
Last updated: May 02 2025 at 03:31 UTC