Should the Egyptian final ى match ي in a vocabulary word? (الدكتور على against الدكتور علي) #1337

55 of 401 King Saud University directory names (measured 2026-11-07) write the final ى (alef maqsura) where others write ي: القحطانى, على for علي. A blanket fold in _normalize is not safe, because ى is a letter of its own at the end of موسى and مصطفى, so folding it would change real names. The tatweel, which carries no letter, was folded in #1331 on that basis.

Not measured yet: whether any shipped vocabulary word is missed because of it. The Persian letters were handled in #1331 by listing spellings as same_as, and the same could be done here for the words where the final ي matters.

Questions

• Which vocabulary words end in ي and would be missed when typed with ى?
• Do they ship as same_as spellings in ar.toml, as Read Persian titles and Persian spellings of the Arabic ones (#1151) #1331 did for the Persian letters?
• Does the match's _fold need the same treatment as the vocabulary, or only the vocabulary?

What to measure: the 55 names, split into those where the final ى ends a word the vocabulary lists and those where it is an ordinary name word.

Follows #1151.

🤖 Generated with Claude Code