Re: idna folding (was Re: idna-bis and '゜')

Patrik Fältström patrik at
Tue Dec 18 11:20:49 CET 2007

The overall goal with my message was to explain why we (for some  
definition of "we") see it as Much More Important to define what can  
go in a U-label. And that is why IDNA200x is staying with creating  
that definition.

Other documents might follow. Registry policy, user interface  
guidelines, whatever... Specifically it is important to keep the  
various issues apart as the security implications (security  
considerations) are so different depending on what you look at, how  
and where.


On 18 dec 2007, at 00.36, Erik van der Poel wrote:

> I see. This sounds like the mappings that are performed in the UI (as
> opposed to the mappings that the browsers perform for <a href="...">
> and <img src="...">). I do see the need for interoperability at the UI
> level too. So we might have three different levels in the documents:
> (1) IDNA200X protocol spec (no mappings)
> (2) mapping spec (where user is not editing the host name;
> language-independent; e.g. HTML in search engines)
> (3) UI recommendations (where user is editing the host name;
> potentially language-dependent; e.g. user typing host name into
> browser)
> It would be nice if we only had one spec at level (1), and only one
> spec at level (2).
> We might have multiple documents at level (3), for various  
> languages, apps, etc.
> Erik
> On Dec 17, 2007 3:02 PM, Patrik Fältström <patrik at> wrote:
>> On 17 dec 2007, at 17.30, Erik van der Poel wrote:
>>> Would you please confirm whether you are talking about  
>>> registration or
>>> resolution here?
>> I have got requests from browser programmers whether it is possible
>> for them to know what equivalences the registries have used at time  
>> of
>> registration. So that the browser when accepting written input (on a
>> keyboard) can do similar mappings etc that is happening at time of
>> registration. Context dependent mapping. If nothing else to possibly
>> minimize phishing etc while still ensuring no false positives.
>> I.e more and more requests I have got regarding mapping at time of
>> lookup is to know how to ensure the mapping at resolution is the  
>> same/
>> similar as mapping at time of registration.
>>    Patrik

More information about the Idna-update mailing list