Showcase¶
The programme in one view: three mathematical execution modes, one continuity layer, eight governed domains, explicit handoffs, and a refusal to confuse momentum with completion.
Make mathematical work cumulative, inspectable, and difficult to fool.
MATH-PROGRAMME preserves governance, decisions, terminology, publication, archival state, and the authoritative integrated artifact. It is not a fourth proof stage.
The central transformation¶
| Stage | Governing question | Required artifact | Change achieved |
|---|---|---|---|
| Curiosity | What might be worth studying? | Lead note | A direction becomes visible |
| Reconstruction | What was actually asked? | Source map | Ambiguity is removed |
| Campaign | What exact obligation can be attacked? | Work Package | Search becomes structured |
| Evidence | What proof, computation, or obstruction exists? | Claim ledger | Support becomes inspectable |
| Handoff | What can now be checked? | Certification packet | Dependencies become explicit |
| Certification | What crossed the declared boundary? | Checked artifact | Local reliance becomes warranted |
| Integration | What is the authoritative current account? | Ledger, reviews, ADRs, public page | Meaning survives revision |
What the programme refuses¶
Discovery may be exciting without being authoritative.
Understanding helps a proof; it does not replace one.
The bridge must itself be proved.
Completed, published, selected, and archived describe artifacts unless a mathematical support route says otherwise.
Status vocabulary¶
These compact labels summarize mathematical support. They are not the same as artifact lifecycle or campaign disposition. See the Programme Status Taxonomy.
Domain portfolio¶
| Domain | Mathematical state | Current programme boundary |
|---|---|---|
| 01 · Union-Closed Sets | Open conjecture | Foundational demonstration domain; local proofs and bounded certificates only |
| 02 · Navier–Stokes Critical Integrability | Open problem | Equation-specific critical-integral routes; no regularity theorem |
| 03 · Hodge Conjecture | Open conjecture | Source and equivalence normalization; no new algebraicity result |
| 04 · Birch–Swinnerton-Dyer | Open conjecture | WP00–WP04 promoted; selected restricted target remains unproved |
| 05 · Poincaré Reconstruction | Solved classical theorem | Qualified reconstruction archive; no new proof or complete formalization |
| 06 · Yang–Mills Existence and Mass Gap | Open problem | Source-normalized axiomatic dossier; no continuum construction or physical gap theorem |
| 07 · P versus NP | Open problem | Machine and encoding lock; no equality, separation, algorithm, or unrestricted lower bound |
| 08 · Riemann Hypothesis | Open conjecture | Function and zero normalization; no proof, disproof, or newly certified zero range |
The catalogue is not a scoreboard. Different domains may legitimately produce a source audit, a false-proof atlas, a negative result, a selected target, a bounded certificate, or an archival dossier.
Executable fixture 001 · Exact algebraic identity¶
One claim, three support boundaries
Replace the inequation by an inverse variable.
Replay sparse polynomial arithmetic over exact rationals.
Depends on both semantic translation and checked identity.
x − 1 = t(x² − 1) + (1 − x)(t(x + 1) − 1)The fixture proves that a serialized witness can be checked and adversarially mutated without promoting the surrounding semantic implication beyond its audited bridge.
Executable fixture 002 · Radical membership¶
The exponent and the universe both matter
No nonzero nilpotent elements.
The dual numbers contain a nonzero nilpotent.
The checker distinguishes ideal membership from radical membership and preserves the false broader statement with its countermodel.
Fixture 003: The logarithmic GCD kernel¶
Classical mathematics · certified formal artifact
General GCD-matrix theory already supplies the incidence-factorization criterion.
Lean checks positive semidefiniteness and the exact finite-support Gram realization.
The publication gate preserves every exclusion and makes no novelty claim.
- not a new theorem
- not a novel kernel
- not a first proof
- not a first feature representation
- not a first Lean formalization
Publication changed visibility, not mathematical history. The theorem remains classical; the formal and editorial contribution is stated without a priority claim.
Review posture¶
I know which domain I am reading, what is proved, what is computed, what is conjectural, what failed, what was ruled out, which artifact is authoritative, and what must happen next.