You mean that there are other languages with other alphabets? And some of those look like ours? The horror!
So, should UNICODE be squeezed into ASCII so that it can be used in hostnames? Which UNICODE? My position on internationalized domain names may be unwelcome: You don’t use an ASCII subset? Tough shit.
One of the reasons for the ASCII subset was to be able to transfer the object identifier from the computer to paper and back via my fingers and a pen, or from this computer to that computer without having recourse to copy and paste. It’s certainly much easier to do that with DNS names than, say, X.208’s Object Identifiers. I will admit that the non-Latin world has had to adjust to our alphabet, much as I have had to adjust to Arabic numerals. That’s what happens when you trade: you pick a common language and use it.
Though, now that they’ve stopped teaching writing in the schools, we’ll soon be illiterate, so it won’t much matter.
(On the other hand, I think it’s funny how all these problems come from trying to approximate analog forms within the digital namespace.)