However, for the more than two-thirds of the world who do not speak English, the DNS is anything but friendly.
"I have traveled throughout Asia and seen how Asians struggle with the usage of English," says James Seng, Chief Technology Officer of i-DNS.net. "Even myself, I make countless spelling and grammar mistakes. It is important for the Internet to be completely internationalized to help these individuals get onto the Internet."
If Seng and many other proponents have their way, this dream may soon become a reality. It's known as i-DNS, short for Internationalized Domain Name System. The i-DNS technology promises to give users the ability to use Web URLs, email addresses, and FTP in their native language, no matter what the language.
Given that most of the world uses languages that are currently not valid under the current DNS system, the implications of i-DNS are far-reaching.
Today, a Web user in China accesses Chinese Yahoo! by entering the English characters: chinese.yahoo.com. With i-DNS, this user could more easily input:.
Far less dramatic, but no less important, are European companies who will no longer have to rely on a variation of their corporate name as their domain name. For example, Citroën could actually use the domain name www.citroën.com.
The evolution of i-DNS
The DNS is a system of mapping readily recognizable names to numeric IP addresses and vice versa. For example, www.amazon.com maps to 126.96.36.199.
When the DNS was established, the limited character set (a subset of ASCII) was sufficient for the intended users - academics and programmers who already had a working knowledge of English. The ASCII subset was also a very practical solution in a world well before Unicode. As the Internet expanded and became more user-focused, the need for international domain names grew. But it was commonly accepted at the time that the DNS could not be internationalized.
Martin Dürst, at the time Research Associate at the University of Zurich, Switzerland, and now Internationalization Activity Lead at the World Wide Web Consortium, thought differently. "I had an idea of how to create a fake i-DNS and wrote it up as an Internet Draft just to prove that it could be done without changing anything inside DNS."
Dürst proposed a way of piggybacking i-DNS on top of the existing DNS by converting all non-ASCII addresses into ASCII. Instead of assuming the address is incorrect (because it falls outside the ASCII boundary) the client converts the address to UTF-5 (ASCII). This set can then be interpreted in the same fashion as the current DNS addresses.
The algorithm Dürst wrote was used in a testbed held by APNG in the second half of 1999. It supported Chinese, Japanese, Korean, and Thai encodings. Servers were reconfigured to accept (and resolve) international domain names and by most accounts the test was a success. But it is just one of several tests being conducted, and not the only encoding conversion being used. According to Dürst there are tests being conducted in UTF-5, UTF-8, as well as vendors using local Chinese and Thai encodings. To work towards establishing a standard, the IETF has established a Working Group devoted to this issue.
The domain grab is on Many organizations and individuals are not waiting idly by for standards to be agreed upon. They're staking their claim to multilingual domain names before someone beats them to it. Yahoo!, Citibank, Lycos, and Cathay Pacific have all registered i-DNS names.
There are a number of vendors willing and eager to accept these registrations. i-DNS.net has taken the lead in promoting its vision of i-DNS with it acting as the international registry, both distributing the technology that makes i-DNS possible (for free) and collecting a percentage of every registration issued. It has partnered up with a number of registrars and ISPs in China, Taiwan, India, Hong Kong, and Singapore (see below). It claims it can support up to 55 languages, and plans to offer email support as well.
Keep in mind that all of this is happening at a dizzying pace. i-DNS.net was founded in October, 1999, as a spinoff of the i-DNS research conducted at the National University of Singapore, under the auspices of Asia Pacific Networking Group (APNG). The first domain names weren't registered until November. Since then i-DNS.net claims its partners have registered more than 50,000 domain names, with a backlog of more than 100,000. People in Hong Kong, Taiwan, Thailand, and India are already using i-DNS services.
|i-DNS.net approved registrars:|
|LGA International (www.lga.net.sg)||Singapore||Chinese|
|3rd Generation NIC (www.3gdns.com)||Hong Kong||Chinese|
|CS Internet (www.cs.co.th)||Thailand||Thai|
|Softview Computers (www.tamilcinema.com)||India||Tamil|
The obstacles ahead
Rolling out a new technology that must be compatible with existing technologies on such a massive scale is a significant undertaking. "The trick is to create a system that doesn't impair the existing system," says Dürst.
Durst says that once a i-DNS standard is agreed upon, "it will be rather easy to deploy inside the DNS itself, and for the Web, but will take more time for email and other systems. But this may not be a problem, as the Web is where people care most about internationalized domain names".
In addition, users must have browsers and email software that are i-DNS-compatable. Users with English browsers won't be able to take advantage of Asian URLs so long as they don't have the software to input Asian characters or browsers that can accept Asian text.
So many players, so few standards Assuming that the technical issues can be overcome, there is the greater question of whose standard will win out.
Although Dürst wrote the algorithm for UTF-5, he recommends using UTF-8, "Now that now we are seriously approaching the problem, we need a long-lasting and cleaner solution that UTF-5."
Dürst says that UTF-5 has long-term limitations. He uses the analogy of cars trying to move large quantities of material down a narrow street. To increase the amount of material transported at any given time, there are two solutions - stack the cars on top of one another or widen the street.
"UTF-5 is the stacked car approach," says Dürst. It works, but it's not the most efficient method. "UTF-8 is the same as widening the street. If you know you're going to need to move more information over the long term, it makes sense to widen the streets now, even though it may take more time and effort."
No matter what solution is agreed upon, his primary concern is that it doesn't damage the current DNS system, "The greatest risk and threat to the Internet as a whole is an i-DNS where you don't know which machine you get to when you type in a domain name."
Seng agrees, "The last thing I want to see is incompatibility among the competitors, as it creates confusion and chaos for everyone." Seng says that i-DNS.net will work closely with the IETF to ensure that standards are implemented in a clearly defined manner.
What the consumer wants, the consumer gets "It is my wish that one day I will be able to use my Chinese name as my email address," says Seng - Seng may very well see this wish come true.
Today the Internet is a consumer-driven enterprise. If the consumer wants a Web site delivered in his/her native tongue, vendors are bending over backwards to make this happen. Web site localization is booming and it will not be very long before English represents a minority of the content on the Internet.
It only seems natural that domain names should also be localized. Most experts seem fairly confident i-DNS will happen. The only questions are how and when. One thing, however, is certain - there are registrars ready and waiting to provide you with a localized domain name, for a nominal fee.
Companies who assumed they had the world safely covered by registering every country top level domain name, such as ".jp," ".kr," and ".th," had better think again. There is a whole new range of domain name addresses up for the taking and the same land grab mentality has taken hold.