Repository navigation
Publish an Erdős 647 campaign report with verified frontier artifacts and claim discipline #3
Description
Summary
Create a public campaign report for Erdős problem 647 that explains what has been kernel-verified, what remains prior art, what is experimental, and how the unresolved wall is being mapped.
This should be a durable research page or long-form blog post, not a transient progress update.
Verified results to present
- Shift-8 classification:
(n - 8) / 8is prime or twice a prime. - Strengthened 13-coefficient modular sieve: 48 survivors modulo 46189.
- Prime-23 refinement: 528 survivors modulo
46189 * 23. - Prime-29 refinement: 768 survivors modulo
46189 * 29. - Explain per-prime refinement and CRT composition as the scalable verification pattern.
Prior-art and trust boundaries
- Hughes's public development already establishes the Stage-1 96-survivor / 41-open-residue framework and must be credited.
- Idén's finite computation and Hughes's Python frontier certificates are external computational evidence, not Lean kernel evidence in this project.
- The 41-residue frontier is not equivalent to solving the open problem.
- Companion lemmas, formal refinements, and finite certificates must not be described as a resolution.
- Novelty claims require citation and prior-art review before publication.
Required sections
- Statement and historical context
- Known computational exclusions
- Prior formal work
- Our proof-search workflow
- Kernel-verified results table
- The 41-residue wall
- Why admissible prime constellations survive finite congruence sieves
- Current in-flight shifts and next obligations
- Trust legend: kernel verified, externally cited, experimentally reproduced, open
- Reproducibility links and hashes
Visuals
- Progression diagram from 2520 divisibility to modular sieve to sub-AP refinements
- Residue-grid or frontier visualization
- Artifact table with theorem/problem/episode IDs, environment hash, proof status, and disclosure status
- Wall map showing verified territory and remaining bridge obligations
Acceptance criteria
- Every mathematical claim has a visible trust label.
- Prior work is linked and credited.
- Kernel-verified artifacts link to public dossiers or safe summaries.
- The page contains a last-reviewed date and can be updated incrementally.
- No statement implies Erdős 647 is solved.
npm run buildpasses.
Activity