IDNAbis Main Open Issues

[Harald Alvestrand]

Stephane Bortzmeyer wrote:
> On Sat, Jan 19, 2008 at 11:33:51AM -0800,
>  Mark Davis <mark.davis at icu-project.org> wrote 
>  a message of 193 lines which said:
>
>   
>> My recommendation is and has been: permit all characters from all
>> modern scripts.
>>     
If "permit" means the same as "ALWAYS", I am *strongly* against this 
approach. For all the reasons stated in -issues and in tons of previous 
email on this list.
>
> You explained very well the problems with the ALWAYS/MAY BE/NEVER
> approach and, for the record, I want to say that this entire approach
> is flawed and should probably be abandoned.
>
> In the case (I hope not) where IDNAbis is adopted by the IETF (I hope
> not), your proposal is a reasonable way to limit the problems it will
> bring.
>
>   
>> With any new version of Unicode, a company like Apple, Google, or
>> Microsoft updates its software to use that version, and characters
>> become acceptable that were not previously.
>>     
>
> Yes, the biggest problem with the IDNAbis change, it that restrictions
> which were in the registry are now in the resolver, which means that,
> in practice, it will be impossible to add new scripts or characters,
> because it is not feasible to upgrade all the programs.
Stephane,

are you talking about NEVER (which the resolvers SHOULD police), or 
about MAYBE (which the resolvers SHOULD NOT police)?

                  Harald