Browser IDN display policy: opinions sought

Gervase Markham gerv at
Mon Dec 12 13:06:27 CET 2011

On 10/12/11 16:35, Patrik Fältström wrote:
> Let me emphasize that what John writes here is extremely important.
> If you have the slightest opinion of what "confusing" implies, and
> what implications approval of "too similar" TLDs might have,
> specifically cross scripts, you should let ICANN know.
> Now.

How? Hopefully this is the right channel.

Mozilla is of the very strong opinion that once it has entered the
territory of creating non-ASCII TLDs and opening up the TLD process to
applicants on a wide scale, the "confusables" issue becomes a matter of
serious ICANN concern. If I pay my money, can I register
".cyrillic_c-om"? If not, why not? What about ".cöm"? The implications
of having no restrictions in this area are the risk of opening up
user-unfriendly scenarios ranging from the shady (typo interception to
drive search engine traffic) to the downright criminal (phishing).

(IMO, the definition of 'confusable' for TLDs needs to be wider than
just visually confusable; I don't think anyone should be able to get
".cmo" either, for pretty much the same reasons - but that's out of
scope for my comments here.)

For this purpose, ICANN should be setting out best practice in the area
of algorithmically determining visual confusability and, once it has
done so, it should be contractually obliging all registries with whom it
has a contract to follow that best practice in regard to end-user

Registries should be free to implement whatever technical measures work
best for them - ICANN should define a required outcome ("no two domains
which are 'confusable' under these rules should be registered to two
distinct entities"), not a mechanism.

As a browser maker, we would prefer not to have to put any restrictions
on the display of validly-registered IDNs, by being able to rely on
registries to do the right thing (and accept the blame if it goes
wrong). We want the registry community to solve this problem, rather
than having to impose a solution from outside.

Does that help? :-)


More information about the Idna-update mailing list