Gervase Markham gerv at
Tue Mar 11 21:37:46 CET 2008

Mark Davis wrote:
> Because that policy does not actually appear to be feasible. How does 
> TLD XX determine whether 2 strings are homographs or not? Is it up to 
> them? 

It's up to them; they publish a document saying how they are doing it, 
and we check to see whether we're happy with it.

Most blessed TLDs are allowing IDN for only a very limited number of 
characters, and so the question of homographs doesn't really arise. For 
others, e.g. .cn, there are established guidelines.

> If you try to register ball.XX and 
> ba11.XX is one of them blocked (for the "Firefox-blessed" TLD XX)?

We haven't required anti-spoofing algorithms to be applied within the 
ASCII character set. This is for historical reasons, really. 1 and l 
look pretty different in the font my Firefox (Ubuntu Linux 7.10) uses 
for the URL bar.


More information about the Idna-update mailing list