Skip to content

docs(backlog): file the scoring gap that re-opens at the filing rate (BACKLOG #1444) - #838

Open
wshallwshall wants to merge 1 commit into
mainfrom
claude/manager-9291ef
Open

docs(backlog): file the scoring gap that re-opens at the filing rate (BACKLOG #1444)#838
wshallwshall wants to merge 1 commit into
mainfrom
claude/manager-9291ef

Conversation

@wshallwshall

Copy link
Copy Markdown
Collaborator

Ledger only. One commit, one file, 25 added lines. BACKLOG #1444 is filed as a DECISION, not a defect — the items involved are behaving correctly and the ledger simply has no gate.

The mechanism

A scoring pass is a snapshot. Nothing in CI requires a value or difficulty at filing time, so the gap re-opens as fast as items are filed.

Measured 2026-09-04 against main at 685d4f548: 278 open items, 5 with no score — #1424, #1427, #1429, #1430, #1439 — every one filed 2026-09-03. PR 754 closed this to zero on 2026-09-03, so the gap went 0 to 5 in about a day.

Re-measured here with parse_items rather than a hand-rolled scan, against the reporting seat's figures. Both the count and the exact five item numbers agreed independently.

Why it is a decision

#1427's own banner says value and difficulty belong to a scoring pass. That is a deliberate convention, so:

  • Moving the gate to filing time needs a ruling before it needs code, and would red on the five existing items on day one — unshippable without a grandfather baseline or changed-lines-only scope. Same shape as #1435 (open in PR 831, not yet on main), where an unscoped ledger gate measured 22 day-one failures and was deliberately not built for that reason.
  • Re-running the pass periodically keeps the convention intact but is unbounded work with no set cadence, and the gap re-opens between every run — which is exactly what just happened.

Scoring the current five would not settle it. That is another snapshot and it decays the same way. The thing to fix is the absence of a gate or a cadence.

Two small things

The item carries its own score, so it is not an instance of itself.

The first draft cited #1435 as an in-page anchor link, which dangles because #1435 is not on main yet. Caught by dangling_citation_check, and confirmed mine rather than pre-existing by a controlled run: the checker exits 0 on unmodified main and exited 1 with the change applied. Now written as a plain reference that says it is not yet on main, per the convention for citing a number that exists but has not landed.

Checks

backlog_status_check OK (665 items, each declaring exactly one status). dangling_citation_check exit 0. link_check OK (5035 relative links across 321 files). All pre-commit hooks passed. No code touched, so ruff and mypy have nothing to check.

🤖 Generated with Claude Code

…(BACKLOG #1444)

A scoring pass is a SNAPSHOT. Nothing in CI requires a value or difficulty
at filing time, so the gap re-opens as fast as items are filed.

Measured 2026-09-04 against main at 685d4f5: 278 open items, 5 with no
score -- #1424, #1427, #1429, #1430, #1439 -- every one filed 2026-09-03.
PR 754 closed this to zero on 2026-09-03, so the gap went 0 to 5 in about
a day. Re-measured here with parse_items rather than a hand-rolled scan,
against the reporting seat's figures; both the count and the exact five
item numbers agreed.

Filed as a DECISION rather than a defect. The items are correct: #1427's
own banner states that value and difficulty belong to a scoring pass. So
moving the gate to filing time needs a ruling before it needs code, and
would red on the five existing items on day one -- unshippable without a
grandfather baseline, the same shape as #1435. Re-running the pass keeps
the convention but is unbounded work with no set cadence.

Scoring the current five would not settle it. That is another snapshot,
and it decays the same way.

The item carries its own score, so it is not an instance of itself.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@wshallwshall wshallwshall added the reviewed A reviewer has read this. Removed automatically when new commits arrive. label Sep 4, 2026
@wshallwshall
wshallwshall added this pull request to the merge queue Sep 4, 2026
@github-merge-queue
github-merge-queue Bot removed this pull request from the merge queue due to no response for status checks Sep 5, 2026
@wshallwshall
wshallwshall added this pull request to the merge queue Sep 5, 2026
@wshallwshall
wshallwshall removed this pull request from the merge queue due to a manual request Sep 5, 2026
@github-actions github-actions Bot added the ci-red A required check went red. Attribute it before retrying. label Sep 5, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ci-red A required check went red. Attribute it before retrying. reviewed A reviewer has read this. Removed automatically when new commits arrive.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant