How engineers tamed the Internet's Tower of Babel

At a United Nations meeting last month, a bespectacled Swede made a small, barely noticed announcement that nevertheless represented a pivotal moment in the history of the Internet. "Regarding the technical implementation for the world wide web, we are done," Patrik Fältström told the Internet Governance Forum. By "we are done", he meant that following a decade of hard work by a global consortium of engineers and linguists, they had finally decided on a document that will enable all the world's languages to be fully represented on the internet. People will be able to type in addresses in their own language, search in their own language and move around the Internet in their own language. The challenge was every bit as immense as it sounds. The Internet was designed to work with the English alphabet - a to z, and the numbers 1 through 9. Useful symbols rapidly made their way into the system - plus, minus, dash, and so on - each represented with a particular code (or, as Internet engineers insist on calling it, an "identifier"). Agreeing on identifiers was easy at first, but as Internet use spread across the globe, people started asking for more to be added to fit other languages, whether an accent on a letter, or an entirely different alphabet. Global balancing act As languages have spread and developed, some elements have changed and some stayed the same. Some have grown to have different meanings. Some look identical and are anything but. One thing is for certain: everyone is unshakeable in their belief that their language is as valid as any other. No matter how wonderful the Internet is, it does not override culture and history. The result has been a very careful balance. "No script and no person will be happy with the definition of identifiers," explained Fältström. "Everyone will be unhappy. We just have to find a standard that makes people the least unhappy as possible." It can be difficult for an English speaker to grasp the problem. For example, the small dots over the "a" and "o" in Fältström's surname carry significance and meaning. Because it is a western language, we are able to view it as an "a" and an "o" with some dots. Not so with different alphabets. Fortunately, there is a real-world example that makes this global balancing act more understandable. Richard Haigh is a web designer from the exotic climes of Nottingham and the proud owner of "£.com". He has decided he wants to use the site to cover the debate over Britain's possible future adoption of the euro. "When it does kick off, I want to provide somewhere where people can voice their concerns," he explains. Despite having "no personal belief either way", he thinks that he's on to something unique with his pound-symbol domain name. But Haigh doesn't actually own "£.com". He owns "" - the identifier used to represent the pound symbol. In fact, £.com doesn't (strictly speaking) exist. Why? Ask John Klensin, who is, along with Fältström, the person most responsible for unusual additions to the Internet's domain name system. He is blunt about Haigh's web address: "The £.com domain shouldn't exist - it has been prohibited all along," he explains. When told it clearly does exist, he is unremitting: "If [the web address] resolves, it is probably another bug. Somehow it has been sneaked through." Haphazard nature In fact, when you type £.com into your browser you are in fact entering the £ sign's identifier, xn--9a. Your browser translates the £ sign as its identifier, goes to and is then redirected to the more palatable How? Because "£" is just another identifier. For computers, £ is dealt with in exactly the same way as ö. But since £ isn't a part of a language, it is one of the identifiers that has not made it through the process. There is a good reason why, Klemsin explains: "For some, 'pound' means the hash symbol [#], for others it is the pound sterling symbol [£], and others refer to the hash/pound symbol as the 'number sign'. It is a communications nightmare." But £.com does exist, thanks to the haphazard way the Internet was created. If something takes off, it rapidly becomes part of the system; if it doesn't, it is left to wither on the vine. But no one is discourteous enough to remove the failed experiments on the way. There is a long and often complex procedure that arrives at a set of standards, recognised officially, that ensures something will be accessible right across the Internet. After a very long, often difficult process, that standard for including the world's languages has now been put forward to the one organisation that can formally enter it into the Internet - the Internet Corporation for Assigned Names and Numbers (Icann), based in California. Icann has put out a formal request for comments as a final precaution. And once that process is finished, then everyone has to figure out how to actually make the standards work with the existing Internet infrastructure - another daunting task already under way. The truly global Internet is on its way. • What's in a name? English-speaking web users have traditionally been very blase about the non-English Internet. That changed last year when researchers set up a fake Paypal website using a Cyrillic ? to replace the first ‘a’ of To the average user, the URL looked the same, except the second site said simply "Meeow". It didn't take a genius to realise the criminal possibilities and suddenly a previously unnoticed area of Internet development was given some much-needed attention.

print this article

Return to internet news headlines
View Internet News Archive

Share with: