Impact of IDNs on web developers?

Question

Hi,

So, the BBC just released the story that ICANN is going to approve non-latin scripts for use in domain names (http://news.bbc.co.uk/1/hi/technology/8333194.stm).

I'm wondering what influence this will have on us web developers. Are we going to see errors when we're grabbing referral urls, or large numbers of unicode issues when creating links on a webpage? Does anyone know how IDNs are being implemented?

I would guess that platforms with somewhat patchy unicode support, like PHP, will be more affected than others.

Cheers

Answer 1

I imagine that it will be using Punycode, as specified in RFC3492. This should mean that current systems will continue to work just fine, but will not correctly display the non-latin characters.

For example, from the Wikipedia page, the URL http://tūdaliņ.lv/ can currently be encoded as http://xn--tdali-d8a8w.lv/ Systems which recognise this as non-Latin script have the option of displaying it with the non-Latin characters; systems which are not aware of this continue to be able to use it as a standard, if odd-looking hostname. The recent proposal is to extend this use to the top-level domain names.

Answer 2

Yes. This will make spoofing attacks a WHOLE lot easier (various encoded characters look nearly identical to the western ones). Filtering is going to be a pain. Some registries have stated that they will NOT support IDN (apparently .be and a few others), so it won't be universal. It's also not even clear if we're using UTF-8 or punycode or perhaps both (also depending upon which domain registrar you are dealing with, etc.). All in all it's going to be a gigantic mess, but not much more than the current one we live in.

Answer 3

It's been live for some years here in Korea, using what I presume is the same technology. Takeup seems to be very slow indeed so I think there will be time to fix problems as they arise (and there haven't been many).