[derek73]

Part of #762, final round. This batch covers the R section plus the two rendering files.

What changed

Six files in docs/design/decisions/ gain a ## Standing block and a ## Log heading, and their keys leave tests/v2/decisions_without_standing.json:

• R1 (fields are views of the parsed words)
• R2 (the all-particles family_base divergence)
• R3 (initials honor the fold)
• R5 (the case-repair gate)
• render-default
• comparison-surface

Each Standing line states what the tree does today. Every example a line cites was parsed on this branch: for view lines, initials(), capitalized() and capitalized(force=True) were checked; for lines about them, replace() and revise() too. The facade examples were run through HumanName.

Edit-kind-1 head notes

Added where a later decision reversed an entry in whole or in part:

• R1:
  • parse("Smith, MD PhD").suffix returns "MD, PhD" — a comma the input never had #429: its between-piece rule was superseded by parse("John Smith MD PhD").suffix returns "MD, PhD" — a run without a comma renders with one #436/Should the suffix separator be derived from the comma the writer typed, rather than from segment shape? #437, and its single caller of segment_suffix_reading by C1's Rethink the suffix-comma rule (C1): decide the comma's structure once, after the words are tagged #613.
  • parse("John Smith MD PhD").suffix returns "MD, PhD" — a run without a comma renders with one #436/Should the suffix separator be derived from the comma the writer typed, rather than from segment shape? #437, five things superseded:
    • its one call site, by C1's Rethink the suffix-comma rule (C1): decide the comma's structure once, after the words are tagged #613;
    • Smith, Ph. D. Smith, by C13's Should a comma part opening with a credential be a suffix comma? John Smith, PhD Jones reads family John Smith #603;
    • Smith, MD nee Jones PhD, by M2's Rethink the maiden-marker rule (M2): a marker stands after the current name and takes words up to the trailing suffix run #601;
    • the 'PhD, FACS' readings of Smith, MD - PhD - FACS and John Doe, MD - PhD - FACS under the default policy, by S8's Should a name word after a credential join the suffix? John Smith PhD Jones reads middle Smith PhD #602;
    • the Smith row under extra_suffix_delimiters={' - '}, by Should a name word after a credential join the suffix? John Smith PhD Jones reads middle Smith PhD #602 as well.
• R2:
  • 2026-08-16 Should a family name that is all particles have an empty family_base? (parse("Anh Do").initials() == 'A.' — the surname vanishes) #385: fully SUPERSEDED by Should a particle that joins nothing still count as a particle? (Jong van der and Jong, van der already disagree) #404/Should a family name that is all particles have an empty family_base? (parse("Anh Do").initials() == 'A.' — the surname vanishes) #385.
  • Its "Recorded 2026-08-16" bullet and its pre-merge coherence pass: the claims Should a particle that joins nothing still count as a particle? (Jong van der and Jong, van der already disagree) #404/Should a family name that is all particles have an empty family_base? (parse("Anh Do").initials() == 'A.' — the surname vanishes) #385 reversed are marked.
  • 2026-08-29 rules.md#R3's "a conjunction never initials" is unqualified: Juan de y has base y and initials J. #461: the given group it left unsettled, and its readings that "may not move", superseded by 2026-09-20 rules.md#R3's "a conjunction never initials" is unqualified: Juan de y has base y and initials J. #461.
• R3: HumanName('Scott E. Werner').initials() drops the middle initial — the facade reads E as the conjunction #462's shape test was superseded by HumanName.initials() re-derives the conjunction decision instead of honoring the parse's tags: john e smith gives j. s. from the facade and j. e. s. from parse() #528. HumanName.initials() re-derives the conjunction decision instead of honoring the parse's tags: john e smith gives j. s. from the facade and j. e. s. from parse() #528's juan y garcia and JUAN Y GARCIA readings were superseded by rules.md#R3's "a conjunction never initials" is unqualified: Juan de y has base y and initials J. #461.
• R5:
  • The capitalized() lowercases a word that family_base and initials() treat as an ordinary name (ANH DO -> Anh do) #407-arc extraction: the suffix's place in the gate is now A capitalized roman-numeral suffix shouldn't stop no-force capitalization — "juan garcia III" stays lowercase #492's.
  • Two 2026-08-29 entries: their mechanism was reversed by Should Jose E Maria Santos join like Jose e Maria Santos does? (the single-letter connective's Latin-capital veto was never adjudicated) #383/JOSE E MARIA SANTOS reads the conjunction E as an initial; an all-caps name carries no case signal to read it from #479.
• R4: the 2026-08-29 CONJUNCTION CARVE-OUT entry. Its R3 initials premise was superseded by R3's 2026-09-20 rules.md#R3's "a conjunction never initials" is unqualified: Juan de y has base y and initials J. #461.

Edit-kind-2 correction

R3's #585 entry cites the Devanagari example ज़ाकिर. The text had stored it decomposed (U+091C U+093C), which initials ज., so it read as the very bug #585's review fixed. The example is now written with the U+095B escape, and a note says what was wrong.

Review

design-docs-reviewer (opus) ran four rounds. Rounds 1–3 found 8, 8 and 2 problems; round 4 found nothing.

• Round 1:
  1. R2's unjoined-mark line was stated for every view, but case repair keeps the connective lowercase (R4's carve-out). It is now scoped to initials().
  2. R5's gate line said "written in more than one case", which A capitalized roman-numeral suffix shouldn't stop no-force capitalization — "juan garcia III" stays lowercase #492 narrowed to the words outside the suffix.
  3. R3's initials() drops the accent from a decomposed (NFD) name: émile zola gives e. z. #585 examples did not hold: the decomposed ज़ and a composed ジ.
  4. R3's HumanName.initials() re-derives the conjunction decision instead of honoring the parse's tags: john e smith gives j. s. from the facade and j. e. s. from parse() #528 head was unmarked against rules.md#R3's "a conjunction never initials" is unqualified: Juan de y has base y and initials J. #461.
  5. R1's parse("John Smith MD PhD").suffix returns "MD, PhD" — a run without a comma renders with one #436/Should the suffix separator be derived from the comma the writer typed, rather than from segment shape? #437 head was unmarked.
  6. R2's "Recorded 2026-08-16" bullet stated the opposite of the Standing block in the present tense.
  7. R5's boundary line ignored rules/R5's newer boundary rows.
  8. Two Elsewhere lines cited examples their target entries don't contain.
• Round 2:
  1. A positional "the next line's" pointer.
  2. R1's Should a name word after a credential join the suffix? John Smith PhD Jones reads middle Smith PhD #602 note covered only one policy and one name.
  3. "Under the same lexicon" referred back to the wrong antecedent.
  4. R4's CONJUNCTION CARVE-OUT head was unmarked against rules.md#R3's "a conjunction never initials" is unqualified: Juan de y has base y and initials J. #461.
  5. The edit-kind-2 note called the decomposed-nukta initial "correct".
  6. A Rethink the suffix-comma rule (C1): decide the comma's structure once, after the words are tagged #613 note sat in an entry's body rather than its head, and pointers dropped "(Derek),".
  7. A note said no entry recorded a move.
  8. An Elsewhere line pointed into its own file.
• Round 3:
  1. The Smith, MD nee Jones PhD move is recorded by M2's Rethink the maiden-marker rule (M2): a marker stands after the current name and takes words up to the trailing suffix run #601.
  2. The differential-ledger Replace the differential ledgers with a committed snapshot of corpus readings #664 Elsewhere pointer named three twin entries; it now carries the twin key.

The MD - PhD - FACS rows are pinned under the " - " delimiter (_SD in tests/v2/cases.py), and the R1 note now says which reading moved under which policy.

Stale text found and left

• rules/R3 against the tree (needs an issue, not a docs edit). rules/R3 says "a letter written as a base letter followed by its combining accent initials whole". But NFC writes U+095B as ज plus nukta, and that form initials ज., dropping the nukta. The precomposed U+095B keeps it. So NFC-normalized Devanagari loses its nukta in the initial, contradicting that sentence.
• decisions/P8, line 32: the continuation of its 2026-09-13 Should Jose E Maria Santos join like Jose e Maria Santos does? (the single-letter connective's Latin-capital veto was never adjudicated) #383 + JOSE E MARIA SANTOS reads the conjunction E as an initial; an all-caps name carries no case signal to read it from #479 entry says both views give "J. G." for JUAN Y GARCIA. Since rules.md#R3's "a conjunction never initials" is unqualified: Juan de y has base y and initials J. #461 both give "J. Y. G.". P8 is outside this batch, so it is left unmarked.
• decisions/R1's Keep citation comments short: quote, clause, pointer #760 entry attributes the nee move to "the comma now read as a family comma". For Smith, … that comma was already a family comma; the move is Rethink the maiden-marker rule (M2): a marker stands after the current name and takes words up to the trailing suffix run #601's. The diagnosis is plausible, not confirmed, so it is left as written.
• decisions/differential-ledger's own tombstones read "SUPERSEDED by 2026-10-13 Replace the differential ledgers with a committed snapshot of corpus readings #664", which names three twin entries. That file belongs to a parallel batch.
• No Standing line covers Has case repair now earned a column in the snapshot, as initials() did? #826, which is still open.

Tests

• test_design_docs, test_doc_citations and test_rules_doc pass.
• The full suite on 3.11: 22,772 passed, 125 skipped, 3 xfailed.
• The full suite on 3.15: 22,770 passed, and two clock-timed benchmark rows failed (unmatched_opens, maiden_pairs). Rerun alone, both pass twice on this branch and twice on an origin/master worktree. The branch changes no code.

---

Generated by Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.05%. Comparing base (9be7261) to head (cc2b0ef).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #830   +/-   ##
=======================================
  Coverage   99.05%   99.05%           
=======================================
  Files          47       47           
  Lines        4659     4659           
=======================================
  Hits         4615     4615           
  Misses         44       44           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.