Touchstones for "Mapping"

[Harald Tveit Alvestrand]

Andrew Sullivan skrev:
> On Wed, Mar 25, 2009 at 09:08:36PM +0100, Harald Alvestrand wrote:
>   
>> terminology quibble:
>> if it's an U-label, then it's already in canonical form....
>>     
>
> Yes, but my view still stands: it's way better to have the A-Label
> form, because you can't be sure that whatever looks at the U-Label
> later won't be an IDNA2003 client.
>   
I agree with you for all cases where the "whatever looks at the label" 
won't be a human with a text display who expects to find human readable 
strings.

Among all the good things said about A-labels, "human readable" is not 
one of them.

                  Harald