Join 3,512 readers in helping fund MetaFilter (Hide)


□□□ □□□□ □□□□ □□□□□□□□□
October 27, 2009 4:34 PM   Subscribe

Net set for 'language shake-up' - ICANN plans to allow internationalized domain names with non-Latin characters.
posted by Burhanistan (61 comments total) 9 users marked this as a favorite

 
I call metafilter.☺
posted by Jimbob at 4:39 PM on October 27, 2009 [1 favorite]


METAFILTER.Z҉A҉L҉G҉O̚̕̚
posted by Horace Rumpole at 4:40 PM on October 27, 2009 [23 favorites]


The Artist Formerly Known As Prince can finally have a webpage!
posted by qvantamon at 4:41 PM on October 27, 2009 [1 favorite]


I did not slog through the swamps in Vietnam at all so Obama could hand over the Good Ol' American Internet to a One World GovernmentCharacter Encoding Scheme!!!eleven
posted by DU at 4:42 PM on October 27, 2009


And a thousand malware vendors rejoiced!
posted by PenDevil at 4:42 PM on October 27, 2009 [18 favorites]


What, again? Is it really going to happen this year?

The proposal would allow domain names written in Asian, Arabic or other scripts.

Yeah, and such exotic scripts as German, French, Spanish, and pretty much every damn language but American English.
posted by Nelson at 4:43 PM on October 27, 2009


Yeah, and such exotic scripts as German, French, Spanish, and pretty much every damn language but American English.

Naïve and façade called, said they thought they were your friends.
posted by qvantamon at 4:46 PM on October 27, 2009 [13 favorites]


My dream of making Mëtälfïltër.cöm can now come to fruition!
posted by idiopath at 4:49 PM on October 27, 2009 [4 favorites]


What about German's double-S letter, or the updside down exclamation in Spanish? Where do they stand, or were they already allowed?
posted by mannequito at 4:50 PM on October 27, 2009


I want a blackletter domain name.
posted by dunkadunc at 4:50 PM on October 27, 2009 [1 favorite]


the updside down exclamation in Spanish? Where do they stand

In their heads.
posted by qvantamon at 4:51 PM on October 27, 2009


Damn, I'm actually finally going to have to learn to read Japanese now. I've relied heavily on romaji and even English URLs to navigate Japanese web sites for years.
posted by darksasami at 4:53 PM on October 27, 2009 [1 favorite]


This is gonna suck for phishing, because there's a lot of characters that look really similar to Latin ones. This will allow hostile websites to imitate legitimate ones in all kinds of new ways.
posted by Mitrovarr at 4:53 PM on October 27, 2009 [15 favorites]


Registering øl.com. Noooooo! Those Danish bastards beat me to it!
posted by Dumsnill at 4:54 PM on October 27, 2009


One might want to read up on some of the anti-spoofing measures that have been developed to deal with the side-effects of IDNs.
posted by jedicus at 4:54 PM on October 27, 2009 [3 favorites]


I want a blank domain name. no characters, no letters, no numbers, nothing. Whenever you don't enter a URL in your address bar, you'd arrive at my site.

Yeah, that's what I want.
posted by flapjax at midnite at 4:56 PM on October 27, 2009 [10 favorites]


Wait, what? punycode has been around for ages. Example: http://点心和烤鸭.w3.mag.keio.ac.jp
posted by Rhomboid at 4:56 PM on October 27, 2009


Yeah, regular IDNs have worked for a long time. It looks like the only new thing in this announcement is non-ASCII country code TLDs, like .조선글 for Korea and .РФ for Russia.
posted by teraflop at 4:59 PM on October 27, 2009 [2 favorites]


Rhomboid, this is for international tlds... you can do that up to the .jp, but currently the .jp has to be the latin letter j and the letter p. After this, you can have a site that ends in .[kanji for Japan]
posted by qvantamon at 5:00 PM on October 27, 2009


Okay, well that's a rather different kind of announcement than the vague and all-encompassing "ICANN to allow internationalized domains."
posted by Rhomboid at 5:02 PM on October 27, 2009


𝕴𝖒 𝖘𝖔 𝖙𝖔𝖙𝖆𝖑𝖑𝖞 𝖗𝖊𝖌𝖎𝖘𝖙𝖊𝖗𝖎𝖓𝖌 𝖒𝖊𝖙𝖆𝖋𝖎𝖑𝖙𝖊𝖗.𝖈𝖔𝖒.
posted by dunkadunc at 5:16 PM on October 27, 2009 [8 favorites]


"Whenever you don't enter a URL in your address bar, you'd arrive at my site."

I've been visiting for years. You need to update.
posted by mr_crash_davis mark II: Jazz Odyssey at 5:16 PM on October 27, 2009 [3 favorites]


I want a blank domain name. no characters, no letters, no numbers, nothing.

I want the blink tag to be my domain name. Even in print.
posted by DU at 5:17 PM on October 27, 2009 [1 favorite]


What about German's double-S letter

If you use ß in a domain it gets converted into 'ss' prior to doing the domain lookup. So ß.com converts to ss.com, which is a blog about Social Security and Medicare.

or the updside down exclamation in Spanish?

¡.com (that's upside-down-exclamation-point.com) redirects to the punycode address xn--7a.com, which doesn't go anywhere but is registered to some person in Switzerland.

And I'd like to take this opportunity to give a shout out to ☃.net, as featured on Metafilter.
posted by jedicus at 5:32 PM on October 27, 2009


Interesting. Punycode seems to encode ß as ss, which doesn't convert back.
posted by oaf at 5:32 PM on October 27, 2009


𝖀𝖓𝖎𝖈𝖔𝖉𝖊 𝖓𝖊𝖛𝖊𝖗 𝖈𝖊𝖆𝖘𝖊𝖘 𝖙𝖔 𝖎𝖒𝖕𝖗𝖊𝖘𝖘 𝖒𝖊.
posted by Jimbob at 5:35 PM on October 27, 2009 [1 favorite]


Interesting. Punycode seems to encode ß as ss, which doesn't convert back.

Actually ß is converted to ss per nameprep (RFC3491), which is a separate process from punycode (RC3492).
posted by jedicus at 5:43 PM on October 27, 2009 [2 favorites]


If I were slightly less lazy, and slightly less scrupulous, and slightly less confused by the complexities of character encoding, I would make a killing as part of the wave of Web 2.0 domain prospecting that's about to begin.

I'm not, though.
posted by ixohoxi at 6:24 PM on October 27, 2009


thepiratebay.☠
posted by qvantamon at 6:47 PM on October 27, 2009 [7 favorites]


chile's NIC has allowed the "ñ" for a while, now
posted by signal at 6:47 PM on October 27, 2009


http://www.metafilter.

(iPhone users only)
posted by patr1ck at 6:52 PM on October 27, 2009 [1 favorite]


> HE COMES.
posted by clarknova at 7:04 PM on October 27, 2009


Dibs on Deathtöngue.com!
posted by Chocolate Pickle at 7:25 PM on October 27, 2009


Oh god. The last thing Korea needs is Korean web addresses.
posted by GilloD at 7:39 PM on October 27, 2009


Oh Patri1ck. 


posted by june made him a gemini at 8:10 PM on October 27, 2009


I hope spam filters can cope with unicode too
posted by Dub at 8:11 PM on October 27, 2009


So, this will create entire reams of domain names that most users in the world can't type? Great! I'm sure this will lead to no problems at all.
posted by Electrius at 9:31 PM on October 27, 2009


this will create entire reams of domain names ....

Reams. How quaint!
posted by flapjax at midnite at 9:50 PM on October 27, 2009 [1 favorite]


So, this will create entire reams of domain names that most users in the world can't type? Great! I'm sure this will lead to no problems at all.

I'd imagine that if you're going to http://порно.РФ, you'll either be following a link, or already know how to type the characters.
posted by cmonkey at 10:05 PM on October 27, 2009


No, that's not the issue. The issue is that there are all kinds of unicode characters that look very similar to latin characters but aren't. For example: www.ігѕ.com or www.irs.com. Quick, which is real? Here's a whole bunch of others. Browsers, email clients, etc. all have to be coded with extensive heuristics to cope with phishing. But IDN has been around for years, it's not like this is really something new.
posted by Rhomboid at 10:19 PM on October 27, 2009


Oh, I'd totally jump on ☺.net (I'm not a company, so not .com).

I also want that reversing character .org.

∞.∞?
posted by Eideteker at 10:28 PM on October 27, 2009


Wow, this feels kind of like a Tower of Babel moment happening. As a (prematurely) old man, I dislike change. That, and unicode is a mystery to me.
posted by Ghidorah at 10:46 PM on October 27, 2009


This sounds like it will lead even more Balkanization of the Internet unless there is some kind of workaround for those who can't read/input non-Roman characters.

At the very least, phishing attacks will skyrocket. jedicus' link doesn't make me feel at all confident that registrars will be scrupulous enough to prevent spoofed addresses from being registered -- we all remember Network Solutions, right?
posted by armage at 11:01 PM on October 27, 2009


Can't wait to get my . ☃ domain.
posted by cj_ at 11:35 PM on October 27, 2009


For example: www.ігѕ.com or www.irs.com. Quick, which is real?

Neither. It's irs.gov.
posted by oaf at 11:42 PM on October 27, 2009 [1 favorite]


Ghidorah, you're late to the party. I'm on the hardline resistance against non-latin characters. My vsername is a protest against this so-called "ioo" letter that everyone is talking abovt, telling me I shovld vse that after "q".

Fvck yov commies! My alphabet has 23 letters and I like it THAT VVAY.
posted by qvantamon at 11:54 PM on October 27, 2009 [2 favorites]


How do you tell one www.(row of squares) from another?
posted by JHarris at 3:44 AM on October 28, 2009


If you want to get into a fantastic amount of detail about the security implications of IDNs, check out Unicode Technical Report #36, Unicode Security Considerations.
posted by chrismear at 4:24 AM on October 28, 2009


So, this will create entire reams of domain names that most users in the world can't type?

If by "most users in the world" you mean "monolingual, culturally-unaware Americans" then yes. However, a counterpoint:

So, this will create entire reams of domain names that most users in the world can read.
posted by DU at 5:03 AM on October 28, 2009 [5 favorites]


A little balkanisation isn't necessarily a bad thing for old people who aren't all that familiar with the English alphabet and want easier access to government services. As to security models... if you're relying on the average user to distinguish between websites based on the URL or the link text, you are doing it very much not correctly. This will affect a few people who are savvy enough to check URLs before clicking links and not savvy enough to know the source of the link and whether to trust it, or to check for security certificates or the lack of them.

My bank already does it right: "we will NEVER send you a link to Internet banking, type this in the location bar to access your account".
posted by Wrinkled Stumpskin at 5:13 AM on October 28, 2009


http://www.рayрal.com is not http://www.paypal.com. Yep, one of those URLs has some Cyrillic letters and the other doesn't. I'm looking forward to seeing what kind of fallout this is going to cause.
posted by crapmatic at 5:16 AM on October 28, 2009


My bank already does it right: "we will NEVER send you a link to Internet banking, type this in the location bar to access your account".

All this solves is the case where the href differs from the link text. It does not solve the case people in this thread are talking about.
posted by DU at 6:08 AM on October 28, 2009


I might be completely wrong on this, but it seems like forcing everything into latin characters creates security problems for the majority of the world because it users have play guessing games regarding inconsistent phonetic spellings. Is that Chinese character transliterated with a latin "d", or a "t"? Is that Russian name spelled with an "ov", an "of", or an "off"?

Electrius: So, this will create entire reams of domain names that most users in the world can't type? Great! I'm sure this will lead to no problems at all.

All three of the dominant operating systems have supported unicode and non-latin languages for over a decade now, either out of the box or as an additional download.
posted by KirkJobSluder at 6:19 AM on October 28, 2009 [1 favorite]


All three of the dominant operating systems

Windows XP, Windows Vista and Windows 7?????
posted by slogger at 7:02 AM on October 28, 2009


Actually, I was thinking MSWin, OSX, and Linux.
posted by KirkJobSluder at 7:06 AM on October 28, 2009


DU: I think it at least mitigates the other problem. I'm not going to type paypa1 into the url bar even if my eyesight is bad enough that it looks right to me. I can use Russian characters, but I would have to be trying pretty hard to type у instead of y.

People have never been able to trust link text anyway
posted by Wrinkled Stumpskin at 8:19 AM on October 28, 2009


My bank already does it right: "we will NEVER send you a link to Internet banking, type this in the location bar to access your account".

Well, it's settled, then—they are now completely immune to phishing.
posted by oaf at 8:29 AM on October 28, 2009


My bank already does it right: "we will NEVER send you a link to Internet banking, type this in the location bar to access your account".

Great, but how many people just copy and paste the text? Same problem.
posted by ChurchHatesTucker at 11:01 AM on October 28, 2009


From the BBC article: Some countries, such as China and Thailand, have already introduced workarounds that allow computer users to enter web addresses in their own language. However, these were not internationally approved and do not necessarily work on all computers.

This is a crucial point to me. If one talks about "Balkanization" of the Internet, then this is it. The sooner an international standard that takes into account the needs of all major user groups is implemented, the better. Otherwise more new (non)standards will be created and continued to be used. Phishing is just a side issue and shouldn't be over-emphasized.
posted by romanb at 11:54 AM on October 28, 2009


I'm aware that phishing isn't a solved problem. I know that this will make the phishing problem worse. I just think that this particular development doesn't make it significantly worse for a significant number of people, and I also think that it will have significant benefits for enough people to make it a good thing.

What romanb said, it's a side issue. It probably has a technological solution like highlighting sites with mixed character sets in the URL in red to scare off the unwary. Or using the existing "this appears to be a terrible site made by bad people" scare screen that I've seen. In any case, it's relatively unimportant compared to the benefits.
posted by Wrinkled Stumpskin at 12:10 PM on October 28, 2009


Forget phishing, I'm butthurt about another level of escaping in Javascript
posted by wcfields at 11:20 PM on October 29, 2009


« Older While his uncle hits the motivational speaker circ...  |  This is to teach you how to da... Newer »


This thread has been archived and is closed to new comments