The language of the internet is English. This is a result both of its origin and of the dominance of the United States in the world's information technology industries.
In the search for new markets there has been a major investment by software companies such as Microsoft to provide multilingual versions of their end-user software. In addition it is possible to implement non-Latin script alphabets on computer screens and keyboards. But English language dominance remains visible in many ways:
- Applications are written in English and then converted to other languages, creating an English structure centred approach.
- New technology words are created from English sources, then taken directly into other languages.
- Manuals, programming tools and core coding are created and remain in English, even when the user interface is presented in other languages.
- The naming conventions of the internet, described in its Domain Name System (DNS), are in English.
This last example is now the subject of considerable investigation. A web location such as <www.greenleft.org.au> tells the reader that the site has some connection to Australia, is an organisation of some sort, and is something to do with "green left". But what if the name of your organisation depended on the use of particular accents to be meaningful, or could only be written in a different script entirely?
The issue of expanding the character set used for such a location isn't simple. Each character has to be recognised by the DNS software in the devices that send packets backwards and forwards over the internet, and only standard English alphanumerics and a couple of other characters are currently recognised.
Should the system simply be based on the shape of the characters? "H" for example is an "h" in Latin script but an "n" in Cyrillic script. On the other hand, one vowel which is considered by Swedish and Danish to be identical is actually represented differently in each script.
Would the same name in two different scripts point to the same web site or not? And a name in a local language has to mean something rather than just being an arbitrary group of characters.
There are many challenges that face the creation of a truly multilingual internet. Some of the most knowledgeable people in the field such as John Klensin, chairperson of the Internet Society's internet architecture board, are of the opinion that this problem points to a fundamental flaw in the DNS architecture, and that this now needs to be fixed (a huge undertaking). Possibly in five to 10 years we will see serious progress.
There is another problem, however. Right now there is money to be made in selling multi-lingual domain names. Corporations in this area, such as VeriSign, are keen to sell anything they can, regardless of whether it will work or not. At stake is the creation of effective or fundamentally flawed multilingual internet support.
BY GREG HARRIS (gregharris_greenleft@hotmail.com)