Repository navigation
Description
Under two_family_names the family reads the same with or without a family comma, and the surname views do not (measured on master c878443, Parser(policy=Policy(two_family_names=True))):
Juan García Márquez → family 'García Márquez', surname1 'García', surname2 'Márquez'
García Márquez, Juan → family 'García Márquez', surname1 'García Márquez', surname2 ''
rules/R10 (the second-surname view) accepts this: "A family comma marks nothing, even where the part before it holds two family names, since nothing in it tells them from one compound name." rules/O6 (two family names) stands down at a family comma, which "decides by itself".
The case for filling them. The compound-name doubt is the same without the comma, and there the switch already decides it: Andrew Lloyd Webber reads surname1 'Lloyd', surname2 'Webber' under the switch, where Lloyd Webber, Andrew reads surname1 'Lloyd Webber'. The caller has declared that a family holds two names. The comma adds certainty about where the family ends, which is the half the switch otherwise takes on trust.
How often (2026-11-18, orcid-researchers, the free-text credit name, holders in 20 Spanish-speaking countries, read under locales.ES): 2,005 credit names are written with one comma; 811 have two words before it; in 607 those two words are the holder's own family box. All 607 read surname2 ''. ORCID's credit names lean towards bibliographic bylines, where the comma form is common.
Proposed. Under two_family_names, where the part before a family comma holds exactly two name units, the first makes surname1 and the second surname2, as the comma-less name reads them. No field moves: family is already both.
Questions.
- Which heads count as two units?
García Márquezdoes.de la Vega Ruizreads as one unit without the comma too (Juan de la Vega Ruiz→ surname2 ''), so it would stay one. A head of three words (García Márquez López, Juan) has no comma-less reading to follow. - Does the given-name veto apply? Without a comma it keeps
José Luis Pérezfrom readingLuisas a surname. Before a comma both words are family by the writer's own mark, so a listed given name there (Martín López, Ana) should arguably still split. - Is this the view rule reading the comma's family by itself, or O6 marking the units? The view reads O6's mark today, and O6 does not run at a family comma. Where the change lands was not checked in the code.
- The same Accepted clause leaves a connective and a hyphen empty (
Ortega y Gasset,García-López), which decisions/R10 holds as an open question with no issue. Those are marked by the writer even with the switch off. Decide them here, or apart? - With the switch off, nothing changes:
Lloyd Webber, Andrewstays one unit.
Because: a caller who asks for the first surname should get the same answer for one person written two ways.
Activity