ICANN's Tower of Babel Raises Questions

How do you spell “Yahoo” in Arabic or Google a soufflé in Mandarin? In 2010, the more than 5 billion non-English speakers in the world may get a chance to find out.

ICANN , the non-profit group that assigns Internet domain names, appears likely to approve a new technical standard to enable domain names in languages other than English and in character sets other than the Roman alphabet.

Part of this is a technical challenge -- one that ICANN characterizes as “fantastically complicated.”

Currently, it is possible to have portions of URLs in local languages, but the country code (.ru, .jp) must be rendered using Roman characters. So some of the world’s most commonly used orthographic systems, including Chinese, Arabic, Korean, Hindi, and Cyrillic (Russian/Eastern European) cannot currently be used in full email addresses.

This has created various levels of inconvenience, from having to print email addresses and URLs in English on business cards, to users with non-Western keyboards having to change character sets to fill in the address bars of their Web browsers.

When American consumers and businesses dominated the Web in the mid-90s, this was seen as merely a cost of entry; but now, over 50 percent of the Web’s 1.6 billion users come from regions that don’t use the Western alphabet, and they have been making their voices heard.

As non-English-speaking parts of the world account for more and more of the global online population, ICANN faces the possibility of the emergence of multiple, potentially incompatible, technical standards to accommodate the different character sets. To avoid this scenario of “continental drift,” the group is poised to embark on a formidable technical overhaul to afford linguistic parity to non-Roman alphabets.

These technical changes will ripple through the Internet’s current Domain Name System (DNS) to create a common framework for the interpretation of non-Roman characters. In effect, this will mean the implementation of a global, real-time translation layer between the semantic Web address (www.yahoo.com, for example) and the underlying IP address, potentially increasing the surface area for redirects and other mischief.

Time will tell whether the security problems created by the technology shift are greater than the problems of the current system, which often involve “unofficial” workarounds and local translation tools that are not part of the global Internet standard.

Despite the complexities, ICANN is pushing ahead with an aggressive implementation schedule. The group plans to take applications for the new Internationalized Domain Names (IDNs) by the second week of November, with the first batch coming online shortly thereafter.

One immediate result is likely to be a stampede of new registrations, as business owners rush to lock down their trade names as rendered in non-Western languages, and cyber-squatters stake their claims to the Tamil, Arabic, or Cantonese equivalents of “sex.com” and “invest.com.”

The longer-term impacts remain to be seen. Obviously, the move to INDs increases convenience and lowers an annoying access barrier for international users who do not speak English or understand Roman-alphabet characters.

The move also may increase the diversity of sites available in local languages (and only local languages). But that also increases the costs and complexity of localization for sites with global reach or aspirations to a global audience, while keeping other locally produced content out of the mainstream -- which, like it or not, is going to remain English for the foreseeable future.

— Rob Salkowitz is the author of Generation Blend: Managing Across the Technology Age Gap (2008) and co-author of Listening to the Future (2009). His next book is Young World Rising: How Youth, Technology and Entrepreneurship Are Transforming the Global Economy.