Rendered from docs/obligations/0163-ten-decision-records-claim-a-q-number-the-decision-register-does-not-carry.md in the Headwater
corpus. Every document on this half of the site is typed by the taxonomy
the descriptor names: corpus.json.
Ten decision records claim a Q number the decision register does not carry
Context
docs/spec/09-decisions.md is the decision register of this repository, and its sections are keyed on a question number. On 2026-09-06 it carried 36 of them, and they ran Q1 to Q32 and then Q40 to Q43.
Ten decision records name a question number in their file name that the register has no section for. Enumerated on 2026-09-06 by taking the two sets and comparing them: Q33 Q34 Q35 Q36 Q37 Q38 Q39 Q44 Q50 Q51. The reading is a set difference rather than a sample:
comm -13 <(grep -oE '^## Q[0-9]+' docs/spec/09-decisions.md | sed 's/## //' | LC_ALL=C sort -u) \
<(ls docs/decisions/ | grep -oE '^[0-9]{4}-q[0-9]+' | sed 's/^[0-9]*-q/Q/' | LC_ALL=C sort -u)
The gap grew by three between 2026-09-06 and 2026-10-01, and it is now 13 of 28. The same set difference, run again on 2026-10-01, gives Q33 Q34 Q35 Q36 Q37 Q38 Q39 Q44 Q50 Q51 Q61 Q64 Q66. The file names under docs/decisions/ name 28 distinct Q numbers, across 100 records. The register now carries 38 sections: Q1 to Q32, Q40 to Q43, Q65 and Q67. So the register gained two sections and the records gained three Q numbers that it does not carry. The title keeps "Ten", because the file name is the slug of the title at birth, and a rename is one of the answers below.
The two sides answer to nothing. A Q number in a file name is not a facet, so no rule reads it. A section heading of the register is prose, and section.required.missing reads whether a required heading is there rather than which headings a document owes. So neither half of the pair is a declaration, and a strict run has never had anything to say about the gap.
The convention is already retired for new records, and that is the reason a ruling is owed rather than a repair. HW-DR-0045, 0046, 0047, 0048, 0052, 0053 and 0054 carry no Q number at all. Each of those was ruled in conversation rather than filed as an open question on the board. The register gains nothing from them, because its sections are Q-keyed. The ten above are the residue of the older convention, and whether they are a gap or a fossil is the question.
Obligation
A reader who follows a Q number out of a decision record file name reaches nothing. That is the cost, and it is small: the file name is a convenience and the identifier is what a relation names.
The larger cost is that the register asserts a shape it does not hold. It reads as the record of every question this repository has ruled, and it is the record of 38 of them.
What is still owed is a ruling by the owner on the Q key, and waiting_on: ruling names it. The three answers the owner can give are under Discharge. An agent gives none of them, because each one is a judgment about who reads the register.
Discharge
A ruling on the Q key, and one of three answers discharges this.
The register owes a section for each of the ten. The discharge is prose, plus a rule that reads a Q number out of a file name and holds the register to it.
The Q key is retired. The discharge is a rename of the ten file names and of every citation of them. The register then says that its sections cover the questions filed on the board and no others.
The gap is accepted. The discharge is one sentence in the register stating the set it covers, so no reader infers a completeness the file does not claim.
Nothing here proposes one. The value of a decision register is a judgment about who reads it, and this record is the measurement rather than the answer.