So, yes, language tags would be useful here, but the
"correct" transliteration would depend on the receiver's
language!
And, of course, since the receivers UA is providing the transliterations,
perfect knowledge of their desirability and nature is possible, given apropriate
dictionary aparatus.
BTW, your character specifications were interestingly displayed by my UA,
apparently these were Kuten bracketed by control characters? Guess I will have
to break out my archived 2022-JP draft and add it to my reading list.
--
dana s emery <de19(_at_)umail(_dot_)umd(_dot_)edu>