> xn--ime-zza.eu
> šime.eu
Looks like HN needs Punycode support [1] when it displays the domain name after the article title.
> šime.eu
Looks like HN needs Punycode support [1] when it displays the domain name after the article title.
That is the domain name.
I'm glad that Hacker News, the site for hackers, displays the domain name, and not the rendered display version of the domain name. Other places can and should do it differently, it's the right choice here.
I can't understand why you think the underlying ASCII encoding should be used, instead of the Unicode domain that actually gets rendered in the browser's address bar.
That makes as much sense as saying non-ASCII characters in the comments section should be displayed as raw ASCII bytes instead of correctly interpreted as UTF-8 characters. Or that domains shouldn't even be displayed as letters but just as sequences of numbers that are the code points.
The whole point of a domain name is to be semantically meaningful. Otherwise we'd just use IP addresses.
In a domain name, punycode is the encoding. The actual domain name, the one you'd need to run DNS and such, is what you see. Also, rendering it into unicode is a vector for attacks. So it's the wrong choice here, where no one wants to deal with the aggravation of knowing that confusables can be slipped into domain names at any time.
But it's not something we should see. It's terrible UX for the reasons I presented. I don't ever want to see the underlying gibberish in a context where I'm reading and commenting on the web.
This is a forum for discussing news articles from sites. Gibberish is not what you ever want to see.
And the idea of being a vector for attacks doesn't hold water. Yes it's true in a theoretical sense, but anything can be a vector for attacks. If the Unicode version of the domain name is safe enough for major browsers to display in the address bar, then it's safe enough to display to the even more tech-savvy HN readership.
(Also your comment seems to misunderstand the analogy between UTF-8 and Punycode. You say "UTF-8 is the encoding, so it makes sense to display it" -- but the point is HN doesn't display it. HN displays Unicode characters, not the underlying UTF-8 bytes. Which is the same way you want to display Unicode characters for a domain name, not the underlying Punycode bytes.)
You're mistaking misunderstanding for disagreement.
We understand each other perfectly well, we simply disagree.