Repository navigation
Companion note, version 1.37
tex/diaz_prove2me.tex, 44 pages. The PDF is attached as diaz_prove2me-v1.37.pdf.
In this version: the Hilbert-function half of (ME), machine-checked under an explicit unmixedness hypothesis.
Since v1.36 the formal proof of Roy's strong six exponentials theorem rests on one hypothesis, (ME), the elementary
multiplicity estimate. Thirty results of 29 October 2026 prove its Hilbert side, at finite level: inequalities
between dimensions of boxes of polynomials, with no Hilbert polynomials (DiazModulus.torus_hilbert_side).
- What is proved: the chain bound of DALAG Theorem 8.8 at finite level (
DiazModulus.boxHilb_le_of_height); the
model of 𝔾ₐ^{A₀} × 𝔾ₘ^{A₁} inside K[X, Y, Z] (reduced monomials are a basis modulo the torus ideal, its generators
form a weakly regular sequence, translations act by automorphisms, the exact count on a box); and the Hilbert side
of DALAG Proposition 5.14 on the boxes of (ME), λ·2^{r(r−1)/2} ≤ 2^{rd} D₀^{d₀} (2D₁)^{d₁}. - What is assumed: (U), Macaulay's unmixedness, in the form that every associated prime of the quotient by a weakly
regular sequence has height at most its length. It is an explicit hypothesis of the chain bound and of the Hilbert
side, and it is not proved formally. - What is not proved: (U), the multiplicity side of (ME), and the assembly of (ME) from its two sides. So (ME) is
still a hypothesis, and Roy's theorem is still machine-checked only under it. Section 1 says so. - These are routine finite-level forms of DALAG Chapters 5 and 8 (chapter 8 is by D. Roy); none was found in this form in the
sources read.
Later the same day, (U) for polynomial rings over an algebraically closed field was proved on Prove2Me (round C1.2,
DiazModulus.unmixedHeight_mvPolynomial). It is not in this version; it enters the mirror and the note in the next.
Appendix A. One more row with two identifiers; 223 identifiers in all. The mirror holds 559 results, all
building in CI on Lean's three standard axioms.