Let V be the finitely supported sequences over a field F — the free module with basis e₀, e₁, e₂, … The forward shift Sₖ sends eₙ to eₙ₊ₖ; the backward shift Bₖ sends eₙ₊ₖ to eₙ and kills e₀ … eₖ₋₁.
Actual kernels and cokernels: LinearMap.ker, the quotient by LinearMap.range, and Module.finrank — not a count kept by hand. bshift_comp_shift records that Bₖ ∘ Sₖ = id, which is why one is injective and not surjective and the other surjective and not injective.
It was committed first to book11/, on the strength of a placement map that gave Volume XI the role “the algebraic floor”; then to book33/, on the reading that the floor ladder's top rung was 33. Neither was right. Two documents were numbering the same word differently, and a file cannot be addressed correctly until that is settled. The rule now is that the rung number is the volume number, so index theory is XXVIII and book33 is reserved for noncommutative geometry — a different subject, with its own source on the shelf.
| ch | title | state |
|---|---|---|
| 1 | An Operator With a One-Sided Inverse — the shift, its kernel, its cokernel, and the sign | live |
| 2 | Infinity Minus Infinity — why finite dimensions give index zero, and how the shift escapes it (Zois Lecture 1; Mathlib LinearMap.index; 7 theorems) | live |
| 3 | The Winding Number — Ind Tf = −wn(f) read through Zois, observed on eleven symbols, and what Mathlib lacks to formalise the bridge (Noether's own 1921 paper, Math. Ann. 82, now held) | live |
| 4 | Riemann–Roch and the Letter K — why K-theory exists; the even-dimensional index, on the sphere and on curves of genus 0–4 (Blackadar, Rosenberg, Dugger; RiemannRoch.lean) | live |
| 5 | The Space of Fredholm Operators — kernels jump, the index does not; Atiyah’s appendix and why the word “index” (FredholmSpace.lean) | live |
| 6 | The Circle Comes Back — Bott periodicity, and why its generator is a winding number (CliffordSmall.lean) | live |
| 7 | One Number, Two Ways — the Atiyah–Singer index theorem, stated from Blackadar; its odd and even cases are ch 3 and ch 4; tested on line bundles over projective space (IndexTwoWays.lean) | live |
sorryAx. A clean axiom report is not a reading of the statement: per R20, a theorem can assume its conclusion and still report clean. Follow the link before citing one as evidence.