Home / Blogs

Getting a Handle on IDNs

R. Shawn Gunnarson

Internationalized Domain Names or IDNs are back in the news. ICANN recently released a document entitled "Proposed Final Implementation for IDN ccTLD Fast Track Process." It consists of a detailed discussion — 59 pages worth — of ICANN's proposed criteria to administer requests for internationalized country code top level domains (ccTLDs) from countries listed in ISO 3166-1, plus the European Union. In a nutshell, ICANN has now offered a path toward authorizing the adoption of ccTLDs in many countries' native languages. This marks a welcome advance for millions of Internet users who do not speak English or who do not use another language covered by ASCII. But with this advance comes some concerns. ICANN's proposal raises serious questions that must be resolved before IDNs are authorized by ICANN in the gTLD space. Let me explain.

Changing the characters in a domain name changes the domain name. Spoofers have long exploited this principle to create misleading websites, such as the one who cleverly registered a site under the domain name "paypal.com" but substituted a lower-case "а" from the Cyrillic alphabet in place of the normal ASCII "a." This trick redirects web traffic to mimicked sites for the purpose of defrauding users.

Introducing IDN gTLDs could create duplicate websites, too. Consider a Hindi version of the domain name "amazon.com." Second level domain names already can be transliterated into non-ASCII characters. A Hindi version of the site can appear as वीरबाला.com. That is, the domain name appears in Hindi characters followed by the familiar ASCII form of ".com." After IDN gTLDs are rolled out the entire domain name would be transliterated as वीरबाला. काम के. This change certainly would be more convenient for Hindi speakers, who would no longer have to shift from Hindi to ASCII text when typing in the domain name. But doing so also could create a new website. Introducing IDN gTLDs could create, in effect, two websites purporting to be "amazon.com."

This duplication of websites presents at least three serious problems.

Internet users will be confused and almost certainly frustrated as they attempt to navigate among websites that are no longer what they appear to be. Navigating to the IDN version of "amazon.com" could offer a marketplace for books and music and electronics, or it might offer something else entirely, from propaganda to pornography. Rather than having a uniform experience on the Internet, customers will have widely divergent experiences, even when trying to navigate to the same website, depending on where they access the Internet. Introducing IDN gTLDs will then have the perverse effect of discouraging Internet use, rather than expanding it.

Website owners will suffer injuries to their intellectual property. Trademarks like "amazon.com" or "nike.com," in which millions of dollars have been invested, will be diluted and tainted as customers in places as far-flung as Bangalore and Beijing navigate to the new IDNs, only to find that they are not dealing with the trademark owners or are dealing with them through a middle man. Consumer trust could be irreparably eroded.

Governments could use the transition to IDN gTLDs to seize control of local portions of the Internet to censor content. Blogs and articles critical of the government would be impossible to find if the government controlled websites at the registry level. Such a consequence would be a serious blow to the Internet as a tool of free inquiry and communication. Setting a precedent for the first time that a government may censor Internet content at the registry level can be expected to make censorship more widespread.

Introducing IDN gTLDs could fragment the Internet even further. Local IDNs have already been introduced, which has caused problems with interoperability and stability. Unless ICANN unifies the Internet under a single set of standards, the transition to IDNs could fragment the Internet even more seriously. A single, interoperable, international Internet will give way to an archipelago of local networks. The safety and stability of the entire DNS will be threatened in ways that are difficult to predict.

Avoiding these problems in the move to IDN gTLDs needs to be a priority. One suggestion is for ICANN to allow TLD registrars to reserve IDN gTLDs for current registrants in all available linguistic character sets. "Amazon.com" would then be reserved in every available language for the current registrants. Managing IDN gTLDs under a single standard will avoid the problems I have described. Registrars should be permitted to charge a registrant only when a new IDN gTLD is activated. Reserving the IDN space should cost nothing. In this way IDN gTLDs can be introduced without undermining the Internet's power as a means of communication, a forum of ideas, and a global marketplace.

By R. Shawn Gunnarson, Attorney at Law, Kirton & McConkie
Follow CircleID on
SHARE THIS POST

If you are pressed for time ...

... this is for you. More and more professionals are choosing to publish critical posts on CircleID from all corners of the Internet industry. If you find it hard to keep up daily, consider subscribing to our weekly digest. We will provide you a convenient summary report once a week sent directly to your inbox. It's a quick and easy read.

I make a point of reading CircleID. There is no getting around the utility of knowing what thoughtful people are thinking and saying about our industry.

Vinton Cerf, Co-designer of the TCP/IP Protocols & the Architecture of the Internet

Share your comments

Did you just wake up from a David Wrixon  –  Oct 22, 2009 5:06 AM PDT

*** (Removed as per CircleID Codes of Conduct)

ICANN have actively been pursing these matters for years, and have an ongoing process to iron out difficulties. However, introducing inane ideas at this stage is not helpful. Obviously Amazon is entitled to protect its trademarks but if you take name in ASCII that are not obvious trademarks or even generics it is going to get a hell of lot more complicated than you depict it. The alternative characters in many cases have been registered by separate registrants in dot com for nearly a decade. Some of those registrations will precede their dot com counter parts. Then you have problems determining which is the correct translation or even transliteration of any given term. *** (Removed as per CircleID Codes of Conduct)

My, my. Someone's in a foul R. Shawn Gunnarson  –  Oct 22, 2009 6:39 AM PDT

My, my.  Someone's in a foul temper this morning.  Setting aside your name-calling and abuse, let's think about what you've said rather than shout about it.

I did not suggest that ICANN hasn't been pursuing these matters.  I merely raised a few issues that a transition to IDN gTLDs will present and suggested a solution to some of them.  Naturally, the solution of letting registrars reserve IDN space has complications that would have to be worked out in practice.  But I notice that you didn't propose an alternative solution of your own.  Do you have one?

You raise legitimate issues about non-obvious names and correct translation or transliteration that I did not talk about in an 800-word article.  After all, it's a blog, not a white paper.

You also raise an interesting and difficult issue about already registered characters.  Unless ICANN is prepared to allow a form of domain name squatting, in which trademark owners and other interested parties are required to pay for the privilege of not having their names misused in foreign languages, another criterion besides first-come, first-served might have to be considered when deciding how to authorize particular IDN gTLDs.

Your response tells us nothing more. Of David Wrixon  –  Oct 22, 2009 7:09 AM PDT

Your response tells us nothing more. Of course Trademarks need to be protected and there is a mechanism for that, but your solution seems to want to disregard the intellectual property of just about every other nation and culture. Also Trademark law limits rights by geography so extending them across the globe and assuming marks that have never been used is nonsense. It is nearly always the case that more than one company has some claim on a name. If you start trying to arbitrarily extrapolate in this way the arguments become very complicated very quickly. If you look up a word in a dictionary you may be presented with half a dozen pseudonyms each of which may provide a valid translation for a word. If you knew anything about languages you would know that words do not map, equivalents will always have wider and narrower scopes of meaning. You cannot, even if you agree on a mapping of characters simply do that. Cyrillic and Arabic the two main alphabetic languages have different numbers of letters. The Arabs do not write any short vowels, and many constants do not have exact equivalents. There are also only three basic vowels in Arabic, so simple transliteration cannot work. It gets even more complicated when you are trying to transliterate in to Chinese which does not have an alphabet at all. There you are reliant on Tone to differentiate meaning a concept entirely missing in English. Furthermore, Chinese is not even a language; it is a script that serves many related and a few non-related languages. Moreover, Trademarks by their nature relate to the visual appearance rather than the sound or the meaning. Even trying to agree how Amazon should be properly represent in each of the languages could take decades to determine without trying to do it for every conceivable registration configuration. *** (Removed as per CircleID Codes of Conduct)

And I'm sorry that a self-appointed expert R. Shawn Gunnarson  –  Oct 22, 2009 7:39 AM PDT

And I'm sorry that a self-appointed expert like you feels the need to flame anyone who dares to speak on his favorite subject.

And I am sorry that a practicing David Wrixon  –  Oct 22, 2009 8:34 AM PDT

And I am sorry that a practicing Attorney cannot appraise himself of the basic legal background before opining in a manner that is only going to confuse his potential clients.

And then we have the little issue David Wrixon  –  Oct 22, 2009 7:37 AM PDT

And then we have the little issue about how much right a late comer like Amazon Bookstore really has to the name which derives from Greek Mythology. I suggest your ponder the Wikipedia disambiguation page before asserting too many ludicrous claims:

http://en.wikipedia.org/wiki/Amazon

And I think the Greeks themselves might even give you some more perspective, but I bet you didn't know that the Mary Celeste was originally christened Amazon?

There's tons of safety baked in to the current testbed process... Jothan Frakes  –  Oct 22, 2009 11:13 PM PDT

@RSG ... While avoiding the fish slapping fight that you and David have going, I have carefully read this article of yours…

While you raise some interesting points, there's tons of policy, process, and law already in place that protects existing brands from the types of confusion that you're alleging might take place.

Help me understand your reasoning… How could someone confuse वीरबाला. काम के with amazon.com? Not saying it couldn't happen, but it just really seems an edge case more than a prominent issue.  Were it to happen there are a number of remedies available now today to get it solved.

That said, the current testbed for introducing IDN ccTLDs is quite restrictive in that it allows only the country name and only if it does not contain any roman A-z letters in it.  So वीरबाला. <IDN for .INDIA in Hindi> might be possible as a registration, but it really seems like the likelihood of confusion drops massively, and the remedies exist now today for alleviating infringement.

So what's the problem here, exactly?

I actually do speak and read hindi so let me point something out .. Suresh Ramasubramanian  –  Oct 23, 2009 6:19 PM PDT

"amazon.com." Second level domain names already can be transliterated into non-ASCII characters. A Hindi version of the site can appear as वीरबाला.com. That is, the domain name appears in Hindi characters followed by the familiar ASCII form of ".com." After IDN gTLDs are rolled out the entire domain name would be transliterated as वीरबाला. काम के

The hindi characters there are for a word that well .. very vaguely translates to the original concept behind amazon .. a brave woman, or in this case, a brave girl वीर "veer" and बाला "baala".  "काम" is sort of the transliteration of "com" - at least its pronounced similarly "kaam" (which is, by the way, the hindi word for "work").  The last के which snuck in there is a hindi word meaning "of".

If you want it transliterated - where in this case too, Jothan's argument applies 100%, you'd get अमेज़ान.काम .. doesnt look either from the glyphs or from a visual inspection, anywhere identical to "amazon.com"

To post comments, please login or create an account.

Related

Topics

DNS Security

Sponsored byAfilias

Cybersecurity

Sponsored byVerisign

Domain Names

Sponsored byVerisign

New TLDs

Sponsored byAfilias

IP Addressing

Sponsored byAvenue4 LLC

Whois

Sponsored byWhoisXML API