The shifted hom-complexes of C^dg, checked against Mathlib #
DGCategory.shiftD and DGCategory.shiftComp are stated for an arbitrary dg
category, so on Cdg A they must come out as the operations Mathlib already
has. This file checks that, rather than leaving the agreement to a comment.
Three things are recorded.
shiftDonCdg Aisδwith the sign(-1) ^ m, andshiftCompisCochain.compwith the sign the derivation in..Shiftforced. Both arerfl, so the general definitions genuinely restrict to the model.The sign on the shifted differential is the same
(-1) ^ mthat Mathlib's ownδ_shiftcarries. That lemma is about shifting the objects of a cochain —Cochain.shift a : Cochain (K⟦a⟧) (L⟦a⟧) n— which is a different operation from shifting the hom-complex, so this is a real cross-check between two independent constructions and not a restatement of one.shiftD_eq_delta_shiftrecords where the two operations meet: the sign(-1) ^ mproduced byCochainComplex.shiftFunctoron the hom-complex and the sign produced byCochain.shifton the objects agree, which is what makes the two conventions composable rather than merely coexistent.
shiftD on C^dg is δ with the shift's sign.
shiftComp on C^dg is Cochain.comp with the Koszul sign that
DGCategory.shiftComp's derivation forced.
The cross-check. Cochain.shift shifts the objects of a cochain and
δ_shift says that costs (-1) ^ m. shiftD shifts the hom-complex and costs
(-1) ^ m by CochainComplex.shiftFunctor. The two constructions are
independent; the signs agree.