Repository navigation
Should the differential compare initials(), which moved in #408 with every role field identical? #484
Description
Activity
Closed by #494. Answers to the three questions in the issue:
-
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_ambiguitiesagree 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_initialstherefore lists nothing else;validate_rulesrefuses 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_initialsfor no added discrimination. Full record indocs/design/decisions.md, "differential-ledger, the initials view". -
Do the
initials()ignores the folded-word ordering that thefamilyfield applies (der, y vangivesy. d. v.vs the facade'sy. 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 nameinitials()ignores the folded-word ordering that thefamilyfield applies (der, y vangivesy. d. v.vs the facade'sy. v. d.) #408 moves at the default order,der, y van, diffs on roles underfix(#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 DonaldandSte Marieread the particle as the given name #360'slos(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), andHumanName('Scott E. Werner').initials()drops the middle initial — the facade readsEas the conjunction #462 (14 names, fixed in the same PR). -
Is
initials()a template for other derived views,family_basesay? 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.
-
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_ambiguitiesprecedent) would catch render-layer drift on real names at the baseline boundary.Open questions:
_RULE_FIELDSso ledger rules can classify initials-only diffs?initials()ignores the folded-word ordering that thefamilyfield applies (der, y vangivesy. d. v.vs the facade'sy. v. d.) #408-era movements need classification at the 1.4.0/2.0.0 baselines when the field lands?initials()the only derived view worth this, or a template for others (family_base?)