Mappings

[Gervase Markham]

Stephane Bortzmeyer wrote:
> I thought that such decisions were the responsability of the TLD. Now,
> do you mean that a browser's vendor decision (and a bad one)
> constrains TLD policy decisions?

TLDs have complete freedom about how they implement IDN. Firefox and 
other browsers have complete freedom about how and whether they present 
fully-rendered IDN domains to the user.

> Side note: how many TLD have a consistent policy about homographs? Not
> just an anti-Unicode policy but a policy to prevent the registration
> of bankofthewest.example and bankofthevvest.example to different
> entities?

I believe policy in that specific area varies. However, more generally, 
there are now 30+ TLDs who have policies which comply with our IDN 
guidelines and so have been whitelisted:
http://www.mozilla.org/projects/security/tld-idn-policy-list.html
That page also has a rather good summary of why we do things this way.

The guidelines are not exactly complex; I can quote them in their entirety:

> In order for us to display IDNs in a particular TLD, that registry
> concerned must have and keep a published policy stating which characters
> are permitted. If the set of characters contains pairs of homographic
> characters, the policy must specify a method to prevent two homographic
> domains being registered to different entities.

I would have thought any TLD which is issuing IDNs without going even 
that far would be doing its customers a great disservice.

Gerv