zenmemes

New post

PHP Latin Diacritic Transliteration Policy V1

This document defines the approved Latin-diacritic transliteration table for v3 compose normalization.

It is the policy source of truth for the transliteration change described in docs/plans/php_latin_diacritic_transliteration_slices_v1.md (source unavailable).

Goal

Reduce friction for common Latin-script text while keeping the canonical ASCII-only write contract unchanged.

This policy applies only to browser-side compose normalization. It does not change backend validation, canonical storage format, or the rule that unsupported non-ASCII characters still require explicit cleanup.

Policy Boundary

This transliteration policy is intentionally narrow:

  • it applies only to characters explicitly listed in this document

  • it is meant for Latin letters with diacritics plus a small set of approved ligatures/special letters

  • it is not a general Unicode transliteration system

  • it is not an "extended ASCII" compatibility mode

  • characters not listed here remain unsupported in V1

Runtime Rule

During compose normalization:

  1. existing smart-punctuation replacements still run

  2. the transliteration table in this document runs

  3. any remaining non-ASCII characters are treated as unsupported

Field Scope

This policy applies to authored compose fields that already pass through browser normalization:

  • board_tags

  • subject

  • body

For board_tags, transliteration runs before the existing lowercase/strip/collapse normalization. board_tags remain a canonicalized field, and collisions caused by transliteration plus existing normalization are acceptable in V1.

Approved Mapping Table

A

  • À Á Â Ã Ä Å Ā Ă Ą Ǎ -> A

  • à á â ã ä å ā ă ą ǎ -> a

C

  • Ç Ć Ĉ Ċ Č -> C

  • ç ć ĉ ċ č -> c

D

  • Ď Đ -> D

  • ď đ -> d

E

  • È É Ê Ë Ē Ĕ Ė Ę Ě -> E

  • è é ê ë ē ĕ ė ę ě -> e

G

  • Ĝ Ğ Ġ Ģ -> G

  • ĝ ğ ġ ģ -> g

H

  • Ĥ Ħ -> H

  • ĥ ħ -> h

I

  • Ì Í Î Ï Ĩ Ī Ĭ Į İ Ǐ -> I

  • ì í î ï ĩ ī ĭ į ǐ -> i

J

  • Ĵ -> J

  • ĵ -> j

K

  • Ķ -> K

  • ķ -> k

L

  • Ĺ Ļ Ľ Ŀ Ł -> L

  • ĺ ļ ľ ŀ ł -> l

N

  • Ñ Ń Ņ Ň -> N

  • ñ ń ņ ň -> n

O

  • Ò Ó Ô Õ Ö Ø Ō Ŏ Ő Ǒ -> O

  • ò ó ô õ ö ø ō ŏ ő ǒ -> o

R

  • Ŕ Ŗ Ř -> R

  • ŕ ŗ ř -> r

S

  • Ś Ŝ Ş Š -> S

  • ś ŝ ş š -> s

T

  • Ţ Ť Ŧ -> T

  • ţ ť ŧ -> t

U

  • Ù Ú Û Ü Ũ Ū Ŭ Ů Ű Ų Ǔ -> U

  • ù ú û ü ũ ū ŭ ů ű ų ǔ -> u

W

  • Ŵ -> W

  • ŵ -> w

Y

  • Ý Ŷ Ÿ -> Y

  • ý ŷ ÿ -> y

Z

  • Ź Ż Ž -> Z

  • ź ż ž -> z

Ligatures And Special Letters

  • Æ -> AE

  • æ -> ae

  • Œ -> OE

  • œ -> oe

  • ẞ -> SS

  • ß -> ss

Out Of Scope

The following are explicitly out of scope for V1:

  • Cyrillic transliteration

  • Greek transliteration

  • transliteration for non-Latin scripts generally

  • symbol-name rewriting such as ™ -> tm

  • currency-name rewriting

  • emoji aliases

  • punctuation not already covered by the existing smart-punctuation map

  • any character not listed in the approved table above

Notes On Coverage

This table is intended to cover characters commonly encountered by English-speaking users in borrowed words and nearby Western/Northern/Central European Latin orthographies.

It is intentionally not exhaustive for all Latin-script languages and not intended to express language-specific pronunciation or orthographic rules. It is a deterministic ASCII fallback table, not a linguistic transliteration engine.

Examples

  • Café -> Cafe

  • François -> Francois

  • Smörgåsbord -> Smorgasbord

  • Dvořák -> Dvorak

  • Łódź -> Lodz

  • Œuvre -> OEuvre

  • straße -> strasse

Examples that remain unsupported after this table:

  • Привет

  • γειά

  • 東京

  • 🙂

Change Control

If a new transliteration pair is desired later:

  • add it explicitly to this document

  • justify why it belongs in the narrow approved set

  • add regression coverage for it

The implementation must not transliterate characters that are not explicitly approved here.