Repo naming: no rename — assemble `compliance-lecture-style`, then archive this repo

[quantecon-services]

Important

Decision recorded 2026-08-26 — the rename is off. Execution is tracked in #7. Full reasoning in the resolution comment and addendum below, and in the amendment proposed on QEP-3 (QuantEcon/qeps#7).

Decision

• This repo keeps its name (audit.2026-05.style-guide, grandfathered under QEP-3 §1.4) and will be archived once absorbed. An audit is an event; archiving keeps its Pages site and every posted issue link live — no links break, no re-syncs needed.
• The living document becomes a new repo, QuantEcon/compliance-lecture-style — a standing conformance ledger under the compliance-{domain} type proposed to QEP-3. Boundary: status-* reports what machines observe; compliance-* records what a rubric adjudicates.
• Seeding: main plus the in-flight 2026-08 pass on claude/project-review-lecture-updates-6o9a2f, once that update completes.

The original rename proposal (audit-lectures-style-guide) and its checklist are preserved in this issue's edit history; the premise corrections are in the comments below. Remaining work: see #7.

[mmcky]

The naming authority for this is QEP-3 (qeps#7), which isn't cited above — and its draft registry already covers audit-. Raised there as a field test: QuantEcon/qeps#7 (comment)

To answer the family question directly: there is no audit- family today — this repo is the only one in the org with "audit" in its name, archived repos included. And it is not part of reports- either: that prefix has one member, reports-activity, a private Action that generates weekly org activity reports, sitting in project-reporting's activity plane. QEP-3's field test proposes renaming it to reporter-activity, so reports- is a prefix on its way out, not a family to join.

What the draft registry does say: §2 registers audit-{yyyy-mm}-{topic} — "dated point-in-time audit … frozen once published" — and §1.4 grandfathers the dotted names while requiring new members to use the dash form. So the compliant rename of this repo is audit-2026-05-style-guide: the dots become dashes, the date stays. audit-lectures-style-guide takes the registry's prefix while dropping the date its grammar requires and asserting a lifecycle the row rules out.

And the durability argument made above — the cadence-regenerated cross-period series is the asset — is QEP-3's status-* graduation rule almost verbatim ("once its numbers are collected automatically on a cadence it graduates to a status-* repo named for the domain it measures"). If that argument holds, the destination is a status-* sibling of status-lectures and status-translations, named for the domain measured — not a de-dated audit-*. The QEP comment sets out both readings and proposes the amendment that would settle it.

Two smaller notes:

• ROADMAP.md §1 records an interim decision (2026-05) to keep the dated convention, publish a new dated repo per period, and specifically avoid an in-place rename because it breaks the published Pages URL and the posted issue links. This issue reverses that decision without saying so — worth stating what changed. §2's own tentative leaning was style-audits, a third name again.
• The migration checklist is sound and independent of which name wins — in particular the github.io-URLs-don't-redirect point, and that editing contributions/issues/*.md is only half the job without re-syncing the live action-style-guide issues.

Suggestion: hold the rename until QEP-3 resolves the lifecycle question. The QEP is under revision right now, so the timing is good.

[quantecon-services]

Field test concluded — direction agreed with @mmcky: no rename. The living document becomes a new compliance-lecture-style repo; this repo keeps its name and is archived once absorbed. An amendment registering the new type is being proposed on QEP-3 (QuantEcon/qeps#7). This supersedes the rename proposed above. Three parts: a factual correction, the decision, and a revised checklist.

1. The premise doesn't hold on main

Verified against main @ 18716da:

• lectures/data/ does not exist — no history.csv, no rule_reach_history.csv, anywhere in the tree or its history.
• There is no trend chart. lectures/charts.md carries four static charts rendered from an inline data block, and itself states that moving the data to data/scores.csv "so trends can be charted across audit dates" is future work for "when this becomes a recurring monthly pass" (same in UPDATE.md at L74 and L198).
• There is no 2026-08 snapshot. Every audit artifact dates 2026-05-27/29; the only later commit is a docs pointer (2026-06-11). The only "2026-08" strings in the repo are hypothetical examples of a future dated repo name (ROADMAP.md L27, UPDATE.md L206). No tags, no other content branches.
• Pages deploys from main, so the charts page cited above shows no cross-period series either.

The audit has run once; there is no cross-period series yet to protect. The "Why" section above describes ROADMAP.md §1's intended end-state as if already built. That intent is real — it now has a concrete plan (below) — but the decision record shouldn't rest on artifacts that don't exist. (The checklist's link counts are also slightly off: fresh grep finds 17 Pages refs and 14 github.com refs; re-derive at execution.)

2. The decision

Running this case against QEP-3's draft registry produced two conclusions, now proposed as an amendment on QuantEcon/qeps#7:

This repo never renames. Not to audit-lectures-style-guide (takes the registry's audit- prefix while dropping the date its grammar requires); not to the compliant dash form audit-2026-05-style-guide (a pointless rename of a frozen, grandfathered artifact); and not to status-* (this repo's numbers are rubric-adjudicated scores — agent judgment with calibration drift as a managed risk — not machine-observed facts; publishing them under status- would dress judgment as fact). It keeps audit.2026-05.style-guide for life. An audit is an event; this repo is that event's record.

The living document is a new repo: compliance-lecture-style. Proposed to QEP-3 as a new type, compliance-{domain}: the standing conformance ledger — assembled from one or more audits, re-measured in place per pass (the UPDATE.md runbook becomes a skill), with versioned history seeded from each absorbed audit. Once the ledger is assembled, absorbed audit repos are archived, never renamed — content and Pages stay public and citable. Consequences:

• This repo's Pages site stays reachable at its current URL — the "Still open" question above resolves to a flat yes, the 2026-05 report stays published.
• The audit links in the live QuantEcon/action-style-guide issues #18–#21 keep working untouched.
• The link-editing and issue re-sync steps in the migration checklist above become unnecessary. Zero breakage.

This is ROADMAP.md §2.3's Option β — the tentative leaning recorded in May — now with a registered type to assemble into. It explicitly supersedes both §1's interim decision (a new dated repo per period) and the rename proposed in this issue.

3. Revised checklist

Nothing moves until pass 2 is real. Then, in order:

• QEP-3 amendment accepted (compliance-{domain} row; assemble-and-archive lifecycle) — see QuantEcon/qeps#7.
• Build the pass-2 skill (successor to UPDATE.md) and rubric (successor to lectures/spec.md).
• Assemble QuantEcon/compliance-lecture-style: skill + rubric + charts machinery; seed data/history.csv with the 2026-05 row; run pass 2 to generate the current-state reports; link back here for the 2026-05 per-lecture prose.
• Move the contributions/ maintenance loop (mirrors of Proposal: 7 new style rules surfaced by lecture audit action-style-guide#18–#21) to the ledger or project-style-guide — an archived repo can't re-sync them.
• Add a README pointer here → the ledger; close this issue.
• Archive this repo. (Verify at execution that the Pages site persists post-archive — expected GitHub behavior, but it is load-bearing here.)

---

Generated by Claude Code

[quantecon-services]

Addendum — the in-flight pass exists on a branch, and it becomes the ledger's seed.

After the comment above was posted, @mmcky pointed to the working branch claude/project-review-lecture-updates-6o9a2f, where another session is applying the 2026-08 update (249 of 348 lectures reviewed at time of writing). It carries exactly the artifacts this issue's "Why" section cited: lectures/data/history.csv (seeded with the 2026-05 rows) and rule_reach_history.csv (2026-08 rows over the 348-lecture corpus), plus lectures/data/snapshot.json pinning the corpus, a tools/qestyle_*.py deterministic checker suite, and per-lecture reviews/*.json.

Two refinements to the comment above:

1. §1 scoped correctly: the factual correction stands for main and the published site — but the issue's premise was describing this unmerged in-flight state, not aspiration. The artifacts are real; they just aren't the published record yet. (The naming conclusion is unchanged — if anything the branch strengthens it: its own layout already separates machine evidence (tools/, lectures/data/) from adjudication (reviews/, per-lecture prose), which is precisely the compliance- anatomy.)

2. Checklist item 3 amended (per @mmcky): the in-flight update will be completed, and QuantEcon/compliance-lecture-style will be seeded from main plus that branch — the history CSVs, snapshot, qestyle tooling, and current-state reports — rather than waiting to generate a fresh pass 2. The remaining sequence (move the contributions/ loop, README pointer, close this issue, archive this repo with Pages persistence verified) is unchanged.

---

Generated by Claude Code