{
 "tm": "1RB1RA_1LC0RE_1LE0LD_0LB1LB_1RA0LF_1LD---",
 "bucket_reason": "The funded seams stage hit its stop rule (seam alphabet still growing at 1e10), so W4-8 is parked. New evidence suggests a different route might work, and only you can fund it.",
 "question": "Keep W4-8 parked, or fund one 2 h native search for a closed tape language that only has to forbid 00 <B inside the tape?",
 "options": [
  {
   "id": "park",
   "label": "Keep parked (STUCK)",
   "consequence": "Keeps the token extractor, the 1e10 token file, the growth tables, the fuzz and census results and the validated rules on file. Siblings 61 and 704 stay parked with it."
  },
  {
   "id": "invariant_search",
   "label": "One stage (at most 2 h): native closed-language search",
   "consequence": "Search natively (no Coq) for a regular tape language over phase-tracked tokens (u-phase, 1-runs, (10)-runs, short pieces), closed under the machine and free of 00 <B inside the tape. Start from the safety property, not from the orbit's seam list. Stop rule: no closed language in 2 h means park for good."
  }
 ],
 "recommended": "park",
 "rationale": [
  "Seam words keep growing at every abstraction (1,028 verbatim, 270 pieces at 1e10); no template is near.",
  "Yet re-drawn run lengths never halt (0/1,200): one local invariant (no 00 <B inside) would suffice.",
  "Standard CTL/NG deciders already missed this holdout, so a 2 h search has only a fair chance."
 ],
 "evidence": [
  {
   "kind": "link",
   "text": "Segment map to 1e10, right part: grey deposits turn into dark fossil u-runs (card Fig 5)",
   "href": "img/segmap_1e10_right.png"
  },
  {
   "kind": "fact",
   "text": "Distinct seam words 513 / 783 / 1,028 and elementary pieces 182 / 226 / 270 at 1e8 / 1e9 / 1e10.",
   "href": "work/seams_growth.txt"
  },
  {
   "kind": "fact",
   "text": "Run-length fuzz from real cycle starts: 0/1,200 halts within 5e7 steps (residues kept, free, phases free).",
   "href": "work/fuzz_lengths.out"
  },
  {
   "kind": "fact",
   "text": "C reads 0 33,051 times to 1e10, always at the blank left edge, never inside the tape.",
   "href": "work/census_1e10.txt"
  }
 ],
 "effort_if_yes": "One stage (about 2 h) of native search. If a closed language is found: a CTL-style Coq check of that certificate (about 1-2 stages), which covers 61 and 704 if the language contains their orbits.",
 "updated": "2026-10-11",
 "answer": {
  "option": "park",
  "note": "Dennis: apply recommendations (2026-10-12)",
  "date": "2026-10-12"
 }
}