investigation(qa): pre-register #1068's section leave-one-out for qa on fixture #34, with the cost estimated before any run — written in the research repo, hash and estimate posted on #1068
I maintainer di solito rispondono entro 1 giorno
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 4/5
- Tempo stimato
- 3-5 giorni
- Idoneità per principianti
- 18/100
- Tipo di issue
- Funzionalità
- Chiarezza
- Abbastanza chiara
- Stato di attività
- Attiva
- Stack tecnologico
- github, typescript
- Ambito
- developer-experience
Direzione di ricerca
Start with the blocked issue #1645 and the research repository’s rules for experiments and decisions; this issue requires work in the separate private admarble/smpl-sequant repository. The requested files are experiments/E8-qa-leave-one-out.md, decisions/log.md, and journal.md, plus a comment on #1068. Done means a committed, owner-approved preregistration with its cost estimate and a comment carrying its hash, arms, run count, and estimate without a fixture identifier; the issue says a separate owner decision is required before any run.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
Epic: graph-2026-11-simplify Tier: judgment Execution: runner Blocked by: #1645
- Blocked by: #1645
Problem
D-4: pre-register #1068 now for qa's sections on fixture #34, with the cost estimated before any run; exec and fullsolve later. Every later cut of qa's 3,845 lines waits on this measurement, so that the cut is chosen by data. Fixture #34 is the one fixture with headroom: /qa named 1 of 5 planted defects in 8 of 8 runs. Its ground truth lives only in admarble/smpl-sequant (private), and the held-out fixture #1272 names stays unread while sections are chosen. Founding doc P-10, OQ-4.
Decisions
| ID | Decision | Rejected |
|---|---|---|
| P-1 | The pre-registration is experiments/E8-qa-leave-one-out.md in admarble/smpl-sequant, under that repo's rules (committed before any scored run; amendments dated) |
sequant's docs/investigations/ (ground truth would sit beside it on a public repo) |
| P-2 | Sequant gets one comment on #1068 with the commit hash, the arms, the run count and the cost estimate, carrying the marker SEQUANT_PREREG, and no fixture identifier |
no comment (the tracker would not know the measurement exists) |
| P-3 | Firing it is a separate owner decision after the estimate (§10) | firing on approval of this node |
Acceptance Criteria
- AC-1: The pre-registration is committed in the research repo. Evidence:
git -C ~/Projects/smpl-sequant log --format=%h -- experiments/E8-qa-leave-one-out.md | wc -l | tr -d ' '→≥ 1 - AC-2: It has the sections the research repo's method requires. Evidence:
grep -cE '^## (Question|Arms|Fixture|Decision rule|Cost estimate|Stop rules)$' ~/Projects/smpl-sequant/experiments/E8-qa-leave-one-out.md→6 - AC-3: The owner approved it in the research repo's decision log. Evidence:
grep -c 'E8' ~/Projects/smpl-sequant/decisions/log.md→≥ 1 - AC-4: #1068 carries the comment. Evidence:
gh api repos/sequant-io/sequant/issues/1068/comments --jq '[.[].body|select(test("SEQUANT_PREREG"))]|length'→≥ 1 - AC-5: The comment carries no fixture identifier. Evidence:
gh api repos/sequant-io/sequant/issues/1068/comments --jq '[.[].body|select(test("SEQUANT_PREREG"))]|last' | python3 -c 'import json,os,sys;t=json.load(open(os.path.expanduser("~/Projects/smpl-sequant/experiments/E6/fixture/ground-truth.json")));b=sys.stdin.read().lower();print(sum(i.lower() in b for d in t["defects"] for i in d["identifiers"]) if b.strip() not in ("","null") else "no-comment")'→0
isGateTestEvidence classifies AC-2 and AC-5 as gates (parse-check, 2026-11-04). This node changes nothing in sequant's tree and runs no /qa phase (Execution: runner), so §6i never reads them and no SEQUANT_MUTATION record applies; the runner runs each command and pastes its output on the issue.
May touch
admarble/smpl-sequant: experiments/E8-qa-leave-one-out.md (new), decisions/log.md, journal.md. sequant-io/sequant: one comment on #1068.
Non-Goals
- Running it (§10), choosing sections to cut, exec and fullsolve (D-4: later).
- Any read of #1272's held-out fixture.
Done when
A committed, owner-approved pre-registration with a cost estimate exists, and #1068 points to it without naming the fixture's content.
- Lingua principale
- TypeScript
- Stelle
- 1
- Fork
- 1
- Merge medio
- 11h 3m
- PR unite (30g)
- 84
Preparare l'ambiente
- Nessun Dockerfile né file Docker Compose
- Nessun modello di pull request
- Leggi la guida per i contributori
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di sequant-io/sequant
-
graph-2026-11-simplify
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
sequant-io/sequant#1650 ·
I maintainer di solito rispondono entro 1 giorno
-
run: --phases does not suppress the UI-label test phase (determinePhasesForIssue adds it to an explicit list)Forse già presa @admarble l’ha presa oggi. Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
sequant-io/sequant#1631 ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 75/100
sequant-io/sequant#1601 ·
I maintainer di solito rispondono entro 1 giorno
-
bug
Difficoltà 2/5 1-3 ore Idoneità per principianti 86/100
sequant-io/sequant#1582 ·
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 1/5 Meno di un'ora Idoneità per principianti 76/100
sequant-io/sequant#1381 ·
I maintainer di solito rispondono entro 1 giorno
Tutte le issue di sequant-io/sequant
Issue simili
-
content triage
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 82/100
tensorflow/tfjs#8754 ·
-
tool-submission
Difficoltà 1/5 Meno di un'ora Idoneità per principianti 83/100
BraveOPotato/NoSignups#2348 ·
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
code-yeongyu/oh-my-openagent#12784 ·
I maintainer di solito rispondono entro 1 giorno
-
external-issue to-triage
Difficoltà 2/5 1-3 ore Idoneità per principianti 71/100
I maintainer di solito rispondono entro 1 giorno