Smart / dynamic language file generation: 2nd language, international Symbols, translate.google quality & lexical similarity

[ImprovedTube]

(Big topic / worth converting to a separate repo. Also might "convert to discussion")

...previously:

English in all the locale files for consistency

• We made a locale.py for this before https://github.com/code-for-charity/ImprovedTube-for-YouTube/blob/master/py/locale.py )
  • 💡 browser extensions default to displaying the label name when in no locale file #1490

Ahh nice! Might be neat to not update the locales at all, and instead have locale priorities that gets generated when the plugin is loaded at installation/updates (and held in storage; leaving the way predefined by the browsers). So if someone adds something to the en locale, it should not exist in any others (and that makes it easier for people to know what needs to be translated). Instead module loads, goes through all locales to produce object of each key, and fills it out based on language priority, so if someone's priorities are 'it->es->en->all others' then it first looks for key in it and if not there, then looks in es, etc. Would only happen on startup so runtime is fast.

• Next to "en" (international english) there could be "as easy as possible" (with symbols / unicode. For people knowing very few words, kids & illiterates)

• Common L2: en | fr | ar | zh | ru | id | es | hi | de | iranian Persian | Pt | Malay | Urdu | Punjabi | Bengali | .... | ... | ... (depending L1 & place)
  pt_BR ➜ pt ➜ es
  javanese ➜ id
  Malagasy ➜ fr

  • No L2? Lexical similartity

while italian has few L2-speakers, it could help a French-only-speaker.
indonesian ⇆ malay
0.89 fr ⇆ it
0.85 es ⇆ pt

• • regional language ➜ official language | macro-language
    en_US | Nigerian Pidgin | en_UK | en-** ➜ en
    chinese varieties | zh-** ➜ chinese
    arabian dialects | ar-** ➜ arabian
    pt_BR ➜ pt

• Google-translate understandability (best & worst)