the current definition of `font-variant-emoji´ simply implicitly adds a V·S·15 or V·S·16 after emoji characters which can take them. this is definitely the least useful possible interpretation of `font-variant-emoji´. if i `font-variant-emoji: text´, i would like NO emoji characters to be given emoji presentation, not “only emoji characters which were not {already encoded codepoints at time of encoding and thus prone to possible awkward re·presentations in pre·existing encoded texts}”
the entire POINT is that whether a character has text or emoji presentation should be a quality of higher‐level markup and not encoded at the character level. variation selectors are a begrudging concession for use in limited scenarios where a plaintext distinction may never·the·less be argued for. but that whole conceptual framework doesn’t WORK when your higher‐level markup is just doing “implicitly insert characters into the text stream and then render as plaintext like normal”
of COURSE this would mean doing additional work, work which may right now be difficult or even impossible, to do font selection to ensure that only fonts which display emoji characters with text presentation are used, and not fonts which display emoji characters with emoji presentation. this is maybe impossible because i am not sure fonts STILL have a good understanding of what “emoji presentation” means nor operating systems fallback mechanisms for properly selecting the correct font for a given presentation. NEVER·THE·LESS