Snapshot rescue and modernization
CompletedFour historical solver snapshots load on modern SBCL with ASDF, package isolation, and cleaned archival structure.
This dashboard tracks the project against its main mission: validate the integrated constraint solver against the PhD results, expand that validated umbrella to other historical result families, and converge on one trustworthy code line.
Green steps are complete, amber marks the active front, and muted cards show the work still ahead.
Four historical solver snapshots load on modern SBCL with ASDF, package isolation, and cleaned archival structure.
The experiment plots, ACL-vs-SBCL overlays, and tracked graph assets regenerate from a fresh checkout.
The core suites, artifact validation, AO assertions, ff provenance, and dashboard sync now form the supported merge gate.
yj-sun/Csp, PrevResults, and NewData4b-Batch are tied together well enough to support integrity checks and roadmap decisions.
Shared core and AO warning families were reduced until the supported validation gate now runs cleanly; remaining legacy notes sit outside that mainline path.
qcsp3 now holds the PhD baseline, AO baseline, and documented ff relationship together under one supported integrated story. M1 is accepted.
One trusted code path now covers the validated historical surface with explicit boundaries, documented deltas, and synchronized public-facing status artifacts.
With 1.0.0 published, the next mainline is disciplined intake of new materials plus cleanup of warning noise that still lives outside the supported release gate.
Post-1.0 work can then absorb harder historical questions like Hanoi-4 behavior and repeatability once the relevant artifacts and references are ingested.
Archive work now becomes the main practical front: classify new finds through incoming/, deduplicate overlaps, strengthen provenance, and use leads like the identified Lacie1 terrain materials to recover blocked historical domains.
Every new step still runs through the same validation spine so the published 1.0 baseline stays trustworthy while newly ingested materials are reviewed deliberately and post-release cleanup stays bounded.