Should the differential compare initials(), which moved in #408 with every role field identical?

[derek73]

The gate compares the seven role fields plus _ambiguities. Initials are a derived render view, so #408's reordering (M. H. A. A. → M. A. A. H. on folded-middle names) was invisible to it by construction — tests/v2/cases.py's header records that the table has no initials column and the pin lives in rules.md#R3 + test_render.py.

Both surfaces expose initials(); comparing it as a pseudo-field (the _ambiguities precedent) would catch render-layer drift on real names at the baseline boundary.

Open questions:

• Does the pseudo-field join _RULE_FIELDS so ledger rules can classify initials-only diffs?
• Do the initials() ignores the folded-word ordering that the family field applies (der, y van gives y. d. v. vs the facade's y. v. d.) #408-era movements need classification at the 1.4.0/2.0.0 baselines when the field lands?
• Is initials() the only derived view worth this, or a template for others (family_base?)

[derek73]

Closed by #494. Answers to the three questions in the issue:

1. Does the pseudo-field join _RULE_FIELDS? Yes, as _initials, and it stands alone: it enters a name's diff only when the seven roles and _ambiguities agree on every compared surface (a role move drags its initials with it, so that movement is the role diff's consequence, not render drift). A rule listing _initials therefore lists nothing else; validate_rules refuses the mix. Strict subset semantics were measured and rejected: 53 / 43 / 36 existing rules at 1.4.0 / 2.0.0 / 2.1.0 would have needed _initials for no added discrimination. Full record in docs/design/decisions.md, "differential-ledger, the initials view".

2. Do the initials() ignores the folded-word ordering that the family field applies (der, y van gives y. d. v. vs the facade's y. v. d.) #408-era movements need classification at 1.4.0 / 2.0.0? No. At 1.4.0 they ARE parity, and the one corpus name initials() ignores the folded-word ordering that the family field applies (der, y van gives y. d. v. vs the facade's y. v. d.) #408 moves at the default order, der, y van, diffs on roles under fix(#379) at all three baselines, so the pseudo-field never enters for it. What the column did surface and classify: the R2 all-particle readmission (27 names, three ledgers), Mc Donald and Ste Marie read the particle as the given name #360's los (1), the facade initialing each word of a joined run since 2.0.0 (98 names at 1.4.0, recorded late in the 2.3.0 log), and HumanName('Scott E. Werner').initials() drops the middle initial — the facade reads E as the conjunction #462 (14 names, fixed in the same PR).

3. Is initials() a template for other derived views, family_base say? Not added. A second view is two rows per surface (worker template and tree side) plus its own roles-identical guard, and nothing needs it today; the entry above records the decision rather than the issue.