A Script for Every Surfer

The Washington Post - October 11, 2007
Catherine Rampell

Responding to criticism from the non-English- speaking world, the U.S. firm that oversees the Internet will test domains in foreign characters.

From washingtonpost.com

Is the World Wide Web truly worldwide? Depends on whom you ask.

Since the Internet came into widespread use, those among the 70 percent of the world that doesn't speak English have argued that the Web is inaccessible. So next week the nonprofit group contracted by the U.S. government to run the Internet will begin testing domain names in other alphabets.

On Monday, the Internet Corporation for Assigned Names and Numbers (ICANN) will conduct a test to see whether domains written entirely in foreign scripts can work without crashing the Net. For several years, the company has allowed domains that are half in foreign characters, such as [Chinese text].com or [Arabic text].org. For the test, domain names will look like [Korean text].[Korean text].

The long road to this stage, which comes nearly a decade after the technology for creating multilingual domains was invented, has left many in the non-English-speaking world impatient and angry. Questions of political and linguistic sovereignty, alongside accusations of American "digital colonialism," have motivated some countries to create their own Internets, effectively mounting a challenge to the World Wide Web.

Experts say the difficulties of typing in a foreign script have probably held back development of online economies abroad.

"Think of what it would be like if every time you typed out an e-mail address or visited a Web site you had to use Chinese characters or Sanskrit," said Michael Geist, who teaches Internet and e-commerce law at the University of Ottawa. "That's exactly like what people in other countries have to do."

Even the hybrid script names that ICANN allows haven't made things much easier. With this model, speakers of Hebrew, Arabic and any other language written from right to left must type half of the URL in one direction and the other half -- the .com, .net or .org postscript -- the opposite way.

Some advocates of internationalizing the Internet have accused ICANN of ignoring the needs of the developing world.

"Almost 10 years ago we went to the CEO of ICANN with the technology to make [multilingual domains] work," said S. Subbiah, co-inventor of the first multilingual domain technology, who estimates that 2 million of the 138 million domains registered worldwide contain non-English characters. "The response was basically, 'I'm too busy. Go learn English.' "

"There's . . . a little anti-American rock-throwing in that description," said Mike Roberts, the first president and chief executive of ICANN. "The engineers thought that trying to do the non-Roman alphabet thing with all this growth would destabilize the Internet and cause crashes."

The politically sensitive business of standardizing languages has also held up the process.

Countries with slightly different versions of the same script have fought over spelling. Debates have also raged over which corporate, sovereign and ethnic interests should control which domains. VeriSign, for example, is the U.S. registry that manages all the domains that end in .com, which represent half of all the domains in the world. Should it also be given control over multilingual domains that end in some translated version of .com? Or should countries have the right to control all domains in their own national languages? What about languages that cross borders, such as Arabic?

Even the technical tests have caused political flare-ups.

Next week's experiments use the domain name "example.test" translated into 11 languages. A previous model, however, used "hippopotamus" instead of "test." These plans went awry when an Israeli registrar realized the Hebrew word ICANN thought meant "hippopotamus" was an expletive and threatened to involve the Israeli government.

Some countries have taken matters into their own hands.

At least a dozen countries, including China and Saudi Arabia, have created their own domains in different alphabets and their own Internets to support these domains. A Russian newspaper article last July reported that President Vladimir Putin was commissioning the creation of a Cyrillic Internet. Users of Russia's Internet, like current users of China's and Saudi Arabia's, could surf the Web without going through U.S.-controlled ICANN servers.

"We have been told so many times it will be next year and next year and next year that ICANN will make" multilingual domains work, said Alexei Sozonov, chief executive of Regtime, a Russian domain registrar. "So countries now have their own deployments."

Without coordination, some experts say, these new networks will increasingly fragment and destabilize the Internet.

"The longer it takes for ICANN to introduce these domain names, the greater the amount of chaos there'll be," said Ram Mohan, chief technology officer at Afilias and the chair of ICANN's working group on multilingual domains and Internet stability.

Others say patching these countries' Internets together into a "federation" of Internets could preserve global interconnectivity.

"I don't think the sky is falling," said Milton L. Mueller, professor of information studies at Syracuse and a partner at the Internet Governance Project, a global Internet policy think tank. "There are strong economic incentives to maintain compatibility."

These economic incentives, however, may be outweighed by political interests. Independent Internets could, for example, give countries greater censorship power.

"If the Chinese can say you can't post the word 'democracy' in the title of a blog entry, then good luck registering 'democracy.com' in Chinese," said John Palfrey, executive director of Harvard's Berkman Center for Internet and Society.

Still, one country's censorship could be another's peacekeeping.

"There should be some restrictions on domain names that are culturally sensitive," said Yoav Keren, chief executive of Israeli domain registrar Domain The Net, which does not allow registrations of Nazism sites.

"Having a domain that insults a whole community for all eternity is not something you want," he said. "Look what happened with the Danish cartoons that insulted Muslims."

Security experts say coordination of multilingual domains is also important to protect consumers. The similarity of many language scripts makes Web users vulnerable to fraud. Eric Johanson, a security engineer, demonstrated this threat in 2005 with the creation of a paypal.com look-alike site where the first Latin letter 'a' was replaced with a Cyrillic letter 'a.' The two URLs look identical.

To prevent similar spoofing attacks, China, Korea and Japan recently developed standards for their overlapping scripts. Today, if you register a domain in Japanese kanji, for example, a similar version of the domain in Chinese characters is usually given, too.

"The Chinese and the Japanese were screaming and throwing shoes at each other at their first meeting," Subbiah said. "Then, when they got it right, they became the role model for how this should work."