Email or username:

Password:

Forgot your password?
Hector's posts Post Back to profile
Hector Martin

Speaking of Unicode identifiers being a stupid idea: I have not seen a single Unicode/punycode URL in my almost 10 years in Japan, in real life.

Not. Once. Not in the hostname portion, not in the path portion. Never.

Nobody wants that nonsense here. Seriously. It's a silly novelty and only creates practical problems (and security issues).

You know how Japanese ads and billboards direct people to complex pages/URLs? They give you a search term to plug into Google.

(To clarify, you do get Unicode terms in path fields for things like wikis, but never as part of URLs people are expected to type out, and I've seriously never seen punycode domains.)

11 Jun 2023 at 6:13 | Open on social.treehouse.systems
17 comments
Nicolás Alvarez

@marcan In Argentina, .com.ar domains started allowing áéíóúüñ in domain names, and gave registration priority to people who already had the un-accented version registered.

I don't think I have *ever* seen any legitimate domain name using them. People will just assume "links don't have accents" and type it without so you'd need to register both anyway.

11 Jun 2023 at 6:18 | Open on social.treehouse.systems
Nicolás Alvarez

@marcan I think switching from .gov.ar (from blindly adopting the global TLDs) to .gob.ar was a good move though, despite causing similar issues during the transition (users "knowing" domains spell it gov so websites needed to be available in both, etc).

11 Jun 2023 at 6:21 | Open on social.treehouse.systems
rin
@marcan maybe true in case of cjk, but having lived in russia I have definitely seen cyrillic domains used in ads, albeit much rarer than latin ones
11 Jun 2023 at 6:28 | Open on post.ebin.club
Gaelan Steele

@marcan the search terms on billboards thing is common in the UK too. always seems terrifying - so many ways it can go wrong!

11 Jun 2023 at 6:31 | Open on queer.party
Meriel :leafeon:

@marcan I've seen them once in Germany, and it was because I had to fix an ancient tool for managing DNS zones to correctly use UTF-8 and punycode behind the scenes. lol

11 Jun 2023 at 7:01 | Open on elekk.xyz
bex :neocat_flag_nb_256:

@marcan I’ve actually seen like.. two here in Sweden that uses åäö, everyone else just uses a and o as before..

11 Jun 2023 at 7:05 | Open on smellslike.sparklef.art
ダイハツ！ムーヴコンテ

@marcan I've seen a few punycode domains, but most often on prc sites.

11 Jun 2023 at 7:15 | Open on social.treehouse.systems
lj·rk

@marcan I think it's a lot bigger in the Arabic world? I've seen a few domains that used their characters, and punycode integration into browsers make it quite painless. It's pretty ugly of a hack, but I also don't see the issue, tbqh.

11 Jun 2023 at 7:23 | Open on todon.eu
HEROBRINE7GAMER

@ljrk @marcan Usually big websites use ASCII characters for domain names only because arabic is RTL, but for some reason use arabic for sub-categories or article names which renders it in percent-encoding and makes writing the url impossible. For example, Firefox on android does not support punycode integration and displays only in percent-encoding. That is painful. (afaik firefox for desktop requires arabic support packages for punycode integration)

11 Jun 2023 at 10:01 | Open on fosstodon.org
Nicolás Alvarez

@herobrine7gamer @marcan @ljrk punycode is only a thing in domain names, for the rest of the URL it's just Unicode encoded in UTF-8

11 Jun 2023 at 14:49 | Open on social.treehouse.systems
jaseg

@marcan I very occassionally see ones involving umlaut characters (äöü) in DE. They are an integral part of German and are in lots of common German names, and while there exists a standard romanization that everyone knows about, it looks kind of ugly.

11 Jun 2023 at 8:46 | Open on chaos.social
scarcraft

@marcan
Here in spain we have free .es domains for the city councils. I assume that is nearly a mandatory rule to obtain certain money funds. As an example, my city has the logroño.es

11 Jun 2023 at 13:54 | Open on mas.to
Hector Martin

@scarcraft ñ is basically the only one I'd bother with in Spanish, it feels so wrong to write that as "n". But it feels like the exception that proves the rule.

Case in point: it's aragon.es, not aragón.es :)

11 Jun 2023 at 13:58 | Open on social.treehouse.systems
Landon Epps

@marcan I do see Unicode domains in search results. They tend to be single purpose websites geared toward SEO. I bet it‘s because Google prioritizes a result if the URL matches the search term exactly (e.g. wimax比較.com).

Interestingly [kanji].com is often used in logos, but the actual URL is ASCII. I see 価格.com the most, but there are plenty of examples. The Unicode domain in their logo doesn’t exist, not even to redirect. The actual URL is kakaku.com.
11 Jun 2023 at 16:08 | Open on mastodon.social
Hector Martin

@landonepps Oh, it never really hit me but kakaku.com is indeed a very interesting case.

As for the other one... Yeah, I mean, of course such domains exist. I just mean I've never seen one used by any notable company, advertiser, service, etc., and in particular I don't think I've ever seen one used on paper/offline with the intention of being typed in.

11 Jun 2023 at 16:14 | Open on social.treehouse.systems
Landon Epps

@marcan Yeah, I've also never seen a notable company use a Unicode domain. I'm pretty sure it's just an SEO hack where they make their domain a common search term.

11 Jun 2023 at 16:17 | Open on mastodon.social
Григорий Клюшников

Cyrillic IDN domains in the .рф TLD are a thing in Russia. They aren't very popular, but they do exist, including on ads.

13 Jun 2023 at 10:32
Go Up