@KitRedgrave @sydneyfalk @troubleMoney @zatnosk @squirrel
So, again, he issue is program names, not some document copy. The should be a widely-accessible,distinguishable, unambiguous identifier. It's likely going to have to be accommodated in various package and repository formats.

As such, you'll do yorself favours by sticking to small, unambiguous charactersets.

I've just checked the Debian project's Debian Policy manual.

1/

@squirrel @zatnosk @troubleMoney @sydneyfalk @KitRedgrave Specifically:

"Package names (both source and binary, see Package) must consist only of lower case letters (a-z), digits (0-9), plus (+) and minus (-) signs, and periods (.). They must be at least two characters long and must start with an alphanumeric character."

https://www.debian.org/doc/debian-policy/#the-package-name

There are places for internationalisation. And there are places it doesn't apply well.

International Air Traffic Control is another example.

3/end/

@KitRedgrave @sydneyfalk @troubleMoney @zatnosk @squirrel @dredmorbius There's similar restrictions on DNS labels, hence the punycode hack, which has to be implemented in all software that wants to work with internationalized domain names (but is transparent to everything that doesn't).
I don't know a good solution. I can't enter cyrillic characters, you can't type äöüß, never mind chinese... (ctd)

@KitRedgrave @sydneyfalk @troubleMoney @zatnosk @squirrel @dredmorbius ...but the desire for everyone to be able to use their own languages and character sets is legitimate. It's just not in any of the designs the Internet as a technology stack is working with, and many things are not easy to replace.