Number formatting in foreign languages

[Omikhleia]

Why does SILE have the 3 "hacks" SU.formatNumber.en, SU.formatNumber.tr, SU.formatNumber.eo (with methods such as nth(), string() etc.) - not covering many languages, thus - while it also has ICU (via the "justenoughicu" wrapper to Lua) and could cover, with little efforts, a whole lot of languages with less ad hoc coding?

Just playing 5mn with ICU...

> print(icu.format_number_ext(1984, "en-US", 0)) -- 1984
> print(icu.format_number_ext(1984, "en-US", 1)) -- 1,984
> print(icu.format_number_ext(1984, "en-US", 5)) -- one thousand nine hundred eighty-four
> print(icu.format_number_ext(1984, "en-US", 6)) -- 1,984th

> print(icu.format_number_ext(1984, "fr-FR", 0)) -- 1984
> print(icu.format_number_ext(1984, "fr-FR", 1)) -- 1 984
> print(icu.format_number_ext(1984, "fr-FR", 5)) -- mille neuf cent quatre-vingt-quatre
> print(icu.format_number_ext(1984, "fr-FR", 6)) -- 1 984e

Or for useful purposes, not doubts 😁 :

> print(icu.format_number_ext(1984, "eo", 0)) = 1984
> print(icu.format_number_ext(1984, "eo", 1)) = 1 984
> print(icu.format_number_ext(1984, "eo", 5)) = mil naŭcent okdek kvar
> print(icu.format_number_ext(1984, "eo", 6)) = 1 984.

No, seriously:

>  print(icu.format_number_ext(1984, "sr-Cyrl", 5)) -- једна хиљаду деветсто осамдесет и четири
>  print(icu.format_number_ext(1984, "sr-Latn", 5)) -- jedna hiljadu devetsto osamdeset i četiri

Was there a good rationale not using ICU, and having those above-mentioned limited language hooks?
If not, should I make a PR?

EDIT: Of course eventually to unleash the power of the beast, we'd need using locales instead of the current use-my-country-code-as-a-language...

[alerque]

I think there was a good reason for not using ICU, but it no longer applies. Those hacks predate having the full ICU module as a dependency and have just kind of grown along with the source even since it's been available for a long time. There is no sane reason not to use it now.