Should Georgian script count as caseless? (ჯეკ MA gives family '' and suffix 'MA', where محمد MA gives family 'MA') #1222

Georgian is written in Mkhedruli, which has no capital letters in use. Unicode 11 added Mtavruli (U+1C90–U+1CBF) as its uppercase, and since then Python classes every Mkhedruli letter as lowercase ('დ'.islower() is True). So the parser's one-case test reads a Georgian name as written wholly in lower case, and a capital beside it, Latin or Mtavruli, makes the name "mixed case": the case rules then read evidence the writer never chose.

Measured on master 93de2cb:

rules/S6 (the case signal at the suffix slot) says a name written wholly in one case "says nothing about any word in it", and "so does a script with no case to write in". Arabic and hangul get that; Georgian does not, because Unicode gives it a case its writers do not use. The same flip reaches every rule that asks whether a name is in one case: P8 (a single-letter connective against the initial it spells), P11 (a lowercase particle opening a family-first name), T4 (the bare letter as an initial), and C4/C6 after a comma.

Real data rarely mixes them: 4 of 5,242 Wikidata labels of Georgian citizens write a Mtavruli first letter in the Latin habit (Უჩა გოგოლაძე), and a Georgian name with a Latin credential (ნინო ბურჯანაძე PhD) reads the same either way. So this is a question of principle more than of a measured population.

Questions

• Should the one-case test treat Mkhedruli (and Mtavruli) as caseless, as it treats Arabic and hangul — for example by leaving Georgian letters out when asking whether the name is in one case?
• Is any Georgian writing a case signal? Mtavruli is used for whole words in capitals, so an all-Mtavruli name is one case either way (see the case-repair bug capitalized() on all-Mtavruli Georgian writes a capital first letter: ᲜᲘᲜᲝ ᲑᲣᲠᲯᲐᲜᲐᲫᲔ gives Ნინო Ბურჯანაძე #1223).