Skip to content

Locale packs for Alexandre Dumas fils and Jiří Dienstbier ml.: generational words that collide outside their language #684

Open

Description

@derek73

#661 and #679 added junior/senior words for 45 languages to the default suffix vocabulary. Four real, standard forms were left out, because each collides with a name or word in other languages written in the same Latin script:

Words Language Collision
fils / père fr Fils is a surname (767 US Census bearers): Marie Fils would lose her family name
ml. cs, sk, sl, hr a bare trailing ML (initials) would read as a suffix
st. cs, sk, sl St. is Saint, already a title and an ambiguous particle

A locale pack is where a word goes when its claim is only safe once the caller has declared the language (the #604 nic precedent; mechanisms.md#LOCALE-PACKS-PURE-DATA). In a parse the caller has declared French, Dumas fils is unambiguous.

Proposal: a locales.FR pack adding fils/père, and one Central European pack (or CS/SK/SL/HR packs) adding ml., and st. if the St. overlap can be settled. Each needs its docstring's declared deviations, a DEVIATES predicate, and test rows.

Open questions:

  • Does st. belong even in a pack? Jiří Dienstbier st. is standard Czech, but a pack can only add vocabulary, and st is already a title and a particle.
  • One pack per language, or one regional pack? The words are shared across cs/sk/sl/hr.
  • Are two words enough to justify a pack's own surface, or should this wait for the v3 pack work?

Recorded in decisions.md#generational-words.

Metadata

Metadata

Assignees

Labels

enhancementvocabularyPer-word vocabulary model and language packs

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions