Changing the script

Had the Internet been invented in China, you’d need some fluency in Chinese to type Web addresses. But as a U.S. invention, the Internet’s lingua franca is English.

The world is losing its patience with that state of affairs.

People around the world already have keyboards, software and Web sites in their own languages. Now they want Internet domain names, too, but the Internet’s key oversight agency allows that only in limited circumstances.

And that’s one of the reasons countries led by Pakistan, India and China have been pressuring the United States to cede control of the Internet’s addressing system. A late-night agreement during last week’s U.N. World Summit on the Information Society averted a showdown — for now.

“A lot of (Internet growth) is going to take place in developing countries where English is not the language,” said Nitin Desai, special adviser to the summit for U.N. Secretary-General Kofi Annan. “You really have to focus on how you get a multilingual domain name in scripts other than Latin.”

More than 30 percent of Internet users know English, according to InternetWorldStats.com, compared with 10 percent to 20 percent of the entire world’s population by some estimates. That means Internet growth opportunities lie largely with non-English speakers.

Native-language search engines can help some users get around language barriers, but local businesses and organizations still cannot easily advertise their brands.

“When I see an interesting advertisement on the train, I want to go back to its Web site later, but I usually forget the (address) when it’s only in English,” said Motoyuki Fujiwara, a consumer electronics marketing manager in Tokyo.

Non-English domains also would let multinational businesses reach potential customers in their own languages.

The Internet’s main traffic directories know only 37 characters: the 26 letters of the Latin script used in English, the 10 numerals and a hyphen.

Engineers have developed a system called Punycode to trick those directories into recognizing hundreds of other languages, but technical, linguistic and political hurdles prevent their more widespread use.

“Human beings didn’t design all their various alphabets and languages in any thoughtful engineering process,” said Cary Karp, who heads the group that runs the “.museum” domain.

For now, Karp’s organization is supporting several languages including French, Lithuanian, Spanish and Sami. But the suffix, “.museum,” remains in English. The same goes for all the other suffixes, including the Internet’s two most popular top-level domains, “.com” or “.net”.

Since 2001, “.com” and “.net” have been able to accept about 30 non-Latin scripts covering more than 350 tongues. And China’s “.cn,” Taiwan’s “.tw,” Japan’s “.jp” and South Korea’s “.kr” have supported their languages for a few years.

Others are more recent: Thailand’s “.th” began giving owners of English names the Thai equivalent in 2004, the same year “.info,” “.biz” and Germany’s “.de” started accepting the umlaut and other special German characters.

The “.org” domain added German, Hungarian, Latvian and others this year, while India’s “.in” will start offering Tamil and Malayalam scripts early next year. Others to follow include Kannada, a language spoken primarily in southern India.

“I browse the Internet to some extent, but I will use it even more if everything is in Kannada,” said Narasimha, a Bangalore office assistant who uses a single name.

Despite the stated demand, only 1.5 percent of all domains worldwide are in another language, according to VeriSign Inc., which runs “.com” and “.net.” In Thailand, fewer than 1 in 6 have claimed the Thai equivalent of English “.th” names — even though they are being given away for free.

Many businesses and organizations are waiting for Microsoft Corp. to support them in Internet Explorer, the world’s dominant browser, likely next year. For now users need Firefox, Opera or another browser or special software, called a plug-in, for IE.

“If you use two languages, you must switch using control-shift. It’s very inconvenient,” said Roman Shabaltas, webmaster at Russia’s Hermitage Museum. The museum has a Russian domain under “.museum” but doesn’t promote it.

It’s even more confusing for Arabic and other languages written right to left, as the English portion of the same Web address would run left to right.

Some aren’t waiting. China set up its own “.com” in Chinese within its borders, while an Arabic consortium is testing suffixes in Arabic. A Turkish group pushing a splinter Internet network has touted its potential to support addresses entirely in Turkish.

These unofficial efforts, however, risk fracturing the Internet.

People in China entering a Chinese “.com” address might get one site listed in the government’s directories while someone in Europe might reach a different one from the global, sanctioned directories.

Next month in Vancouver, Canada, the organization in charge of domains, the Internet Corporation for Assigned Names and Numbers, will conduct workshops to review some of these unofficial projects.

ICANN chief executive Paul Twomey, who is Australian, says his organization is well aware of the demand.

“We are committed to a global constituency, … but we are also committed to stability and security of the Internet,” Twomey said. “You can’t just turn on a dial.”