Repository navigation
Should a two-word given name like María José or Min Ho fill given, instead of given + middle? #704
Description
Today the parser calls the first given-name word the given name and puts the rest in middle. That's the English-language convention (John Quincy Adams). Several naming traditions treat a multi-word given name as one unit instead: Spanish María José and Juan Antonio, Portuguese Maria do Carmo, Arabic Muhammad Salim, Korean Min Ho, CLDR's own given field (Cornelia Margarita).
Measured by the report-only view added in #703 (cldr_agreement.py --split): 265 CLDR renderings and 3 rows of the conventions corpus split given and middle differently from what the source states. Most of the 265 come from this convention, not a defect: Cornelia Margarita (given) + Sofia Aletta (given2), where the parser reads given Cornelia, middle Margarita Sofia Aletta. GND and Wikidata record every given name in one field, so they don't state the split and aren't scored on it.
Options:
| Option | middle means |
Effect |
|---|---|---|
| (a) keep today's convention | every given-name word after the first | the 265 are recorded as accepted differences and the split can be ratcheted now |
| (b) one given unit | only an explicitly secondary name | breaks what callers expect from middle in English names |
| (c) per language | (a) by default; a language pack declares compound given names (es, pt, ar, ko) |
the default stays as it is; packs get it right for their culture |
Open questions:
- If (c): what does a pack declare: a list of compound given names, a rule (
María+ one more word), or both? - Where do a Sikh religious name (
Kaur) and an English given name beside a Chinese one (Lim Mei Ling Jessica) go? Neither is a middle name. - The conventions corpus states a few rows with a stopgap
given2(Tran Thi Mai,Harpreet Kaur Dhillon,Lim Mei Ling Jessica). Restate them once this is decided.
Replaces the open questions in #702. The split view is in #703.
Activity