Changelog

This standard uses semantic versioning. Specification changes are tagged alongside reference-implementation releases.

Unreleased (working draft toward v1)

Additions made while hardening the parser against real-world submission data:

• Added the unspecified special preference (asked but no usable set given — distinct from none and from a null/absent preference).
• Added the optional exclude/exc flag to represent negative preferences (“just not they/them”).
• Generalized concatenation handling to recover sets from non-subjective and scrambled form order via a form→set reverse lookup (him/her → he/him, she/her).
• Defined separator normalization (, ; & and or whitespace as set separators; \ like /; empty elements dropped).
• Defined filler/placeholder stripping (he/him is fine → he/him; N/A → null).
• Defined partial/unrecognized/empty handling (recognized set + custom remainder; garbage preserved as custom; empty/placeholder → null).
• Expanded the parsing test suite to 47 cases across 22 categories drawn from real submission data.
• Split the canonical JSON Schema into its own @openpronoun/schema package; the @openpronoun/conformance fixtures depend on it.
• Added @openpronoun/zod: a Zod mirror of the schema (TypeScript schemas + inferred types), kept in sync with @openpronoun/schema via a parity test.

0.1 — 2025-05-30

• Initial draft.
• Core English five-form data model (PronounSet, PronounPreference).
• Special preferences: any, none, ask, custom.
• Optional fields: context, privacy, language, and ranking.
• Parsing and normalization rules.
• Display and canonical stringification conventions.
• UX/accessibility, internationalization, and security/privacy guidance.
• Conformance requirements (RFC 2119), JSON Schema, and shared test suite.