ICANN Approves Internationalized Chinese Domain Names

philalethiac writes "Millions of Chinese language users will soon be able to access the Internet using Chinese script following a decision today by ICANN's Board of Directors to approve a set of Chinese language internationalized domain names."

Comments

! Ha!

[KWTm]

! ("shou3" = number one; "biao1" = to announce/post)

Re:! Ha!

[donscarletti]

Ah, pinyin, the writing system that 1.3 billion people can write, and only primary school kids can read. To be able to write in a valid language and never be fear that it will be read by anyone important is liberating. For example:

xie2xian4dian3 shi4 zhai2nan2 de xin1wen2 dan1shi4 wang2zhan4 de ji4shu4 fang1mian4 chu1chou4, er2qie4 nei4rong2 shi4 gou3pi4. wo3 gan4le4 zhong1xiao4 mo4xi1ge1zhuan3 de ma1 de bi1. (Slashdot: News for Nerds, Stuff that matters. Commander Taco's mother is a classy lady)

Though on a more serious note, this is a little bit worrying. OK, ICANN is allowing Chinese domain names, this is no huge problem to me, since I can read and write Chinese anyway. But the Chinese will be pissed off when Japanese start using Kana and they are no longer able to enter the correct domain names to look up porn. I think this just screws the world all over in the long run, at least EVERYONE knows ascii.

Re:! Ha!

[sydneyfong]

Took me a while to decode. Looks like I'm barely primary school level. :) My first language is Cantonese so my Putonghua isn't quite there...

For those who are interested, this is the real translation:

Slash-Dot is Otaku's news. But the website's technological aspect stinks and the content is dog fart. I fucked Commander Mexico-Roll's mother's pussy.

Nice one.

Really?

[Anonymous Coward]

ICANN haz internationalized Chinese domain name?

ICANN speak Chinese but Slashdot can't

[KWTm]

I guess, until Slashdot enables the UTF character set like everyone else has for the past decade or so,

1. There will be some domain names that we can't link to on Slashdot
2. No one will get my First Post joke.

Re:ICANN speak Chinese but Slashdot can't

[xC0000005]

You can link but it will need to be to the punycode versions for the domain portion, I suspect. Just wait for international email addresses - it's viagra ads in all the languages of the world.

Re:ICANN speak Chinese but Slashdot can't

[tlhIngan]

I guess, until Slashdot enables the UTF character set like everyone else has for the past decade or so,

1. There will be some domain names that we can't link to on Slashdot

Slashdot did allow Unicode. Then things like like this happened. Blame the comment trolls for forcing Slashdot to use a whitelist of characters allowed.

As for domain names, from what I see, they start with a standard prefix (I think it's "xn--") followed by the Unicode codepoints. Just so they're compatible across all systems. Browsers can choose to display the codepoints, or, I'm seeing an option to not do that, so you can tell Paypal.com from xn--blahblabblah.xn--blah.

Do they resolve to cn or are they seperate?

[pavon]

I have heard conflicting information about this. I know the new ccTLDs for China (they approved two - traditional and simplified) are aliases for each other (resolve to the same sites), but are they also both aliases for the existing cn ccTLD or do they resolve to an entirely new domain? If they are separate, why did they choose to do it this way? It seems like it would only cause confusion.

Oh, and damn slashdot and it's lack of unicode support. It would be nice to be able to type the damn things when talking about them.

Re:Do they resolve to cn or are they seperate?

[diakka]

I would assume that deciding to do it separately would be the only logical decision. Traditional to Simplified mapping is not 1:1. There are a decent number of cases where two or more Traditional characters map to 1 simplified character. There are also other cases that are 2:2. Managing the transformations centrally would likely be a nightmare.

Time to revisit oldschool phishing attacks

[Mattpw]

With all the non latin address character sets being approved I imagine there is a world of new opportunities which completely void all the "inspect the address bar" education which was pushed on the general public for so many years. ICANN has managed to turn the net into a pretty much anything goes place, almost every major company is practically extorted into buying the new extension flavour of the month to prevent spammers and fraudsters sending seemingly legitimate email and the general public is left completely confused with no guiding address principals.

Re:Time to revisit oldschool phishing attacks

[Trepidity]

There are some attempts to mitigate the problem, though you're right that it can be one. Some registrars are limiting the characters that can appear in their domain, and there's a push to make that more widespread. One approach is to limit to "local" scripts only, so e.g. Cyrillic or Latin in .ru, but no Telegu or CJK in .ru. That greatly limits the number of clashing pairs compared to allowing all of Unicode. Some registers also have policies on not permitting certain known clashes, such as allowing two domains to be registered that are identical, except for one having a Latin 'a' where the other has a Cyrillic 'a' (which look identical in most fonts).

Firefox and Opera will only display the internationalized Unicode name for TLDs that are whitelisted as having a "safe" policy on the subject, and will display the punycode for other domains. Here is Mozilla's current policy.

Re:Time to revisit oldschool phishing attacks

[LambdaWolf]

With all the non latin address character sets being approved I imagine there is a world of new opportunities which completely void all the "inspect the address bar" education which was pushed on the general public for so many years.

Seems like a good browser feature would be to highlight any non-ASCII characters in the address bar in a contrasting color, such as red or bright green. Then it would take only a minimal amount of additional education to understand that it means something is amiss, unless you're clearly expecting an address composed of foreign characters.

Regarding the expressiveness of Chinese characters

[tenchikaibyaku]

This looks like a perfect opportunity to highlight this recent post at the Pinyin News blog, closely related to the issue at hand! (Disclaimer: I'm not affiliated with the blog in any way, but as a former student of Japanese I can relate to the general message.)

Re:left-to-right-top-to-bottom-you-silly-foreigner

[Mashiki]

You might have missed it for the last little while, but English is pretty much the defacto trade language anywhere you go. But no, people don't get worked up over the intrusion of foreign languages into English. English in itself is highly mailable, which is why it's considered a trade language. French on the other hand, gets bent out of shape because they see it as pollution of the language. They're all about purity.

compromise idea to prevent regional isolation

[ad454]

Looks like the domain names will be encoded using punycode instead of the cleaner UTF8 encoding:

http://en.wikipedia.org/wiki/Internationalized_domain_name
http://en.wikipedia.org/wiki/Punycode

However, my biggest concern is that the use of non-ascii characters in domain names breaks the whole International nature of the web, and imposes regional barriers. Your mail client and mail server software might not be too happy with you trying to send an e-mail to "joe@.jp" or "joe@.jp-r14k153opxc" in punycode. (Crap, it looks like slashdot does not accept international characters in comment submission, so you can't read this: "日本人".)

Remember that very few people have rendering and fonts for every written language on the planet, so most people will be cut off from many websites.) With the current IPv4 shortage, one can no longer reliable just use an IP address to access a specific website, e-mail address, etc., since a single IP address can host many domain names.

Personally I think that the best compromise solution would be to only allow non-ascii characters for domain names in different languages if there are submitted with a paired up romainization version that can be equally accepted for the same domain. So using my previous example, one could equally specify ".jp" in Japanese Kanji, ".jp-yn9d427hcvb" in punycode, or "nihonjin .jp" in Romanji. That way you can still cater to a local/regional audience, and still allow everyone else in the planet to reach you.

For those that argue that it does not matter if a domain name is only specified in a foreign language, if all of the hosted content is in the same foreign language forget about all of current International collaboration in Mathematical, Scientific, Engineering, Programming, and other fields. (You can write an entire math proof or software program using only symbols without a single human word.)

Even for individual one-on-one e-mail communications between people in different countries that are able to communicate in a common language this would still be a problem, since a large percentage of e-mail accounts are hosted with a user's local ISP, that in future may leave them stuck with a non-ASCII e-mail address that would cut them off from the rest of the world.

Re:compromise idea to prevent regional isolation

[_merlin]

However, my biggest concern is that the use of non-ascii characters in domain names breaks the whole International nature of the web

Requiring everything to be ASCII breaks with the whole international nature of the web by forcing everyone to use English alphabet characters.

Re:compromise idea to prevent regional isolation

[ad454]

Requiring everything to be ASCII breaks with the whole international nature of the web by forcing everyone to use English alphabet characters.

Everyone has to use English ASCII characters for top level domains (*.com, *.jp, *.cn, ...) and protocols (http, https, ftp, ...), so everyone online in every country has to continue to use ASCII whether they want to or not, even after these International domain names are in common use.

BTW, I never said that everything had to be in English ASCII, just something like a domain name or e-mail address that is used to identify a website or person should be.

The postal system in most countries allow one to mail a letter using romainized characters in addition to local language characters. For example, in Japan I can send a letter from one city to another (locally) using an address like "Akihabara 1-2-3-567, Chiyoda-ku, Tokyo, 111-1111", in addition to Japanese.

Re:compromise idea to prevent regional isolation

[Anonymous Coward]

Circular reasoning. One could just as easily say that TLDs should be allowed to be non-ASCII as well (and who says they won't be?), which resolves the current dependencies on English. But that wouldn't make you happy because you use English and you want everyone else to use it too. Only for the domain or email address, you say... but those just happen to be the most important parts, right?

Your postal system example is just further evidence of systemic bias. Yes, you can send mail in Japan using addresses written in English. Can you send mail in the US using katakana? The point is that DNS is equally important to nearly every country that uses it. Forcing everyone to use a de facto trade language is not good enough. Why isn't it good enough? Because people and businesses end up with transliterations like this: http://www.imdb.com/title/tt0852721/ and they may not even realize that their product, business, or name is being filtered because they're not native English speakers, trade language or no.

The correct solution if you want to get in contact with a foreign domain? Use a translator.

Re:compromise idea to prevent regional isolation

[Trepidity]

In practice, everyone does anyway. I don't know of any country where the people who use the internet haven't already developed at least some informal way of writing in the Latin alphabet, at least for short snippets like addresses. Many seem to prefer it even when alternatives are available--- for example, Facebook supports UTF-8 status updates, but my Greek cousins use Greek transliterated into ASCII more often than they use the Greek alphabet.

Re:compromise idea to prevent regional isolation

[ashenden]

You don't get it: gTLDs and ccTLDs are being translated (aliased) as well. When this is done, for, say, the Japanese user, there will be no need for any ASCII, whatsoever. As for mapping to ASCII, all IDNs are mapped to punycode, which is ASCII, but it will be invisible. And mixed scripts aren't allowed, so phishing fears are overblown; it won't be any worse than it is today. IDNs should have been a part of the original DN structure, but better late than never. It's simply idiotic to have an entire website in Japanese, except for the DN.

Re:left-to-right-top-to-bottom-you-silly-foreigner

[Trepidity]

While I don't like to raise too much sturm und drang about it, as a native English speaker I must still take some affront at the chutzpah with which these dirty foreigners waltz into our tongue, thinking they have carte blanche to sully our language.

This is a terrible idea

[identity0]

As a person who can read/write Japanese(similar to, but a bit different from Chinese) characters, I don't know why ICANN thought this was a good idea. It's not like the actual contents of pages had to be in Latin characters, so "Allowing use of other languages" is not really an issue. Only the address had to be in Latin characters.

Having all internet users use the 26 (x2 for capitals) letters of the Latin charset and 10 numbers is a much, much simpler than having everyone try to learn all the letters of all the character sets out there.

This is going to make administration harder.

If you started getting hacking attacks from .com, would you even know how to type that into your firewall? If you got an email from @.com, do you think you could describe the address over the phone to a colleague? From the preview, it appears Slashdot is filtering out Japanese characters I used for the addresses. The above examples would be tokyo.com and shujin@osaka.com if they were forced to be in latin. And that's something that's usable by both Japanese and foreigners, whereas the Japanese-character addresses are for 'Japanese only'.

I hope ICANN reconsiders and returns to latin+numbers only addresses.

Re:Why?

[BiggerIsBetter]

Everything that's even remotely interesting on the Internet is either blocked in China or English language. Native Chinese sites is nothing but censored propaganda anyway.

There are plenty of Chinese folks all around the world who would like to use this. It's not just the mainland.

Re:left-to-right-top-to-bottom-you-silly-foreigner

[sznupi]

Hm, there was some of it ("what for / useless / why those people won't just learn our script") on the occasion of last such ICANN news (regarding TLDs in, among others, Arabic script IIRC)

Yeah, the language is about communication. And in todays world, there are lots of people for whom even Latin alphabet itself looks like, say, Georgian alphabet to you. Accidentally, they are often amongst those with most to gain, if they had less roadblocks in communication.

Zhuyin fuhao

[tepples]

Kana developed out of man'yougana, the old "rebus" method of using Chinese characters for their sounds to spell Japanese words. Katakana were partial characters, and hiragana were cursive. Chinese has its own analogous system, called zhuyin fuhao, whose alphabet begins bo-po-mo-fo.