Comparing the repository's perfect-complex notions #
The repository uses three deliberately different notions:
schemePerfect Xis the absolute thick envelope of finite locally free coherent sheaves inD(Coh X);schemeRelativePerfect pis pseudo-coherence plus local finite Tor amplitude over a chosen morphismp, insideDqc(X);Coh.TwoTermPerfectDeterminantData Fis explicit presentation data for a degree-zero coherent sheaf, restricted to a two-term finite locally free resolution.
Only the valid forgetful direction involving the presentation data is proved
here: a two-term resolution gives an absolute perfect object, which then maps
to the defining essential image of perfect objects in Dqc(X). Relative
perfection depends on the base morphism and is not identified with absolute
perfection without an additional geometric theorem.
Relative perfection always includes pseudo-coherence. It does not, by
itself, identify the object with the absolute perfect locus in Dqc(X).
The degree-zero coherent-derived object represented by the target sheaf of a two-term finite locally free resolution.
Equations
- _D.derivedObject = (DerivedCategory.singleFunctor (AlgebraicGeometry.Coh X) 0).obj F
Instances For
A two-term finite locally free resolution places its degree-zero target in the absolute perfect thick envelope.
Forget two-term presentation data to the canonical absolute perfect derived category.
Equations
- D.toSchemePerfectDerivedCategory = { obj := D.derivedObject, property := ⋯ }
Instances For
The absolute perfect object supplied by a two-term resolution, mapped to
the honest Dqc(X) locus.
Equations
Instances For
The Dqc(X) object obtained from a two-term finite locally free
resolution belongs to the canonical absolute-perfect essential image.