Domains to go non-English: will it cause internet havoc?

Angus Kidman11 October 2007, 1:38 AM

The biggest change ever to the way the internet works will be switched on for testing on October 15th. The question is: will the world patchwork of internet routers be able to handle it?


A plan to allow non-Roman characters in domain names will make the Internet more international than ever, but Australians look set to play a major role in determining its success.

ICANN, which is responsible for managing global domain policies, will on October 15 begin testing of whether allowing the character sets of 11 languages to be included in top-level domains (TLDs) causes widespread online chaos.

Currently, TLDs such as .com or .au  are limited to using the letters a to z from the Roman alphabet.

"This evaluation represents ICANN’s most important step so far towards the full implementation of Internationalised Domain Names," ICANN's president, Australian Dr Paul Twomey, boasted in a statement. "This will be one of the biggest changes to the Internet since it was created."

The initial test doesn't mean that you can sneakily go out and register any domain name you like in those languages using new characters. Initially, ICANN wants people to link to the domain example.test in each of 11 target language character sets -- Arabic, Chinese simplified, Chinese traditional, Greek, Hindi, Japanese, Korean, Persian, Russian, Tamil and Yiddish --  to see what kind of issues it creates in different applications and whether it impacts DNS stability.

The example.test site will host a wiki enabling the creation of pages within the tested target languages.

Australians are likely to have a large role to play in that testing. "Australians have been quite involved in ICANN -- much more than any GDP or per capita basis would suggest," Twomey noted during a local press briefing earlier this year.

Three of the languages being tested, Greek, Cantonese (in Chinese script) and Arabic, are amongst the top five non-English languages spoken in Australian, according to the Australian Bureau of Statistics.

ICANN had originally planned to set up versions of the .test site and domain in 20 languages, but is now concentrating on communities that expressed enthusiasm for the testing process and languages which have been used in earlier laboratory tests.

The .test TLD will not remain permanently in place.

As well as handling new character sets, the introduction of multi-language TLDs also requires that root servers can process addresses presented in right-to-left reading languages such as Arabic.


Post your comment



Comments

RSS feed Email alert

Adam:

I was about to say that it will help people execute phishing exploits, but a moth flew into my mouth.

29 February 2008, 8:32 PM (2 years ago)report abuse Send to a friend reply

Brendan:

I do believe that in the past UTF-8 characters were allowed in domains, such as www.äñz.com, would have been allowed, which would have allowed phishing. I don't see how Chinese characters could be made to look like anz.com :P

29 February 2008, 8:48 PM (2 years ago)report abuse Send to a friend reply

Tim:

But will foreign (woops, 'Non-Roman', I meant)scripts be displayed in Roman characters on an English (French, German, Latin, Spanish...) computer? It's one thing to have 'www.konichiwa.com', you can read that, but if you can't read the glyphs then how will you know what the web address is? Sure, users of non-Roman character sets have been obliged to tolerate Roman web addresses. For Greeks, Chinese and Arabs (et al) it will be more convenient to read a web address that has been registered in their mother tongue and displayed in their own script, but the flipside is that Greeks, Chinese and Arabs (et al) will also have to cope with Roman, Arabic, Chinese simplified, Chinese traditional, Greek, Hindi, Japanese, Korean, Persian, Russian, Tamil and Yiddish.

The internet user might find their online experience becoming analogous to having to find an unknown word in a dictionary which has been printed in Roman, Arabic, Chinese simplified, (etc.) scripts. Yikes!!


29 February 2008, 8:32 PM (2 years ago)report abuse Send to a friend reply

anonymous user Anonymous user


Tags