Stability of valid IDN labels

[John C Klensin]

--On Tuesday, 22 April, 2008 18:06 +0200 Patrik Fältström
<patrik at frobbit.se> wrote:

> 
> On 22 apr 2008, at 16.28, Andrew Sullivan wrote:
> 
>> If I understood that correctly, then it seems possible that
>> the Unicode classification of a code point could get changed
>> between Unicode versions, and therefore that the derived
>> property could change, too.
> 
> No, rather that is why we have a special category for
> "backward compatibility". Of course, to have a codepoint added
> to that category will require a revision of the "tables"
> document, which imply a decision can be made at that point in
> time whether it should be added or not.

To say this a little differently, such changes in Unicode are
not made in secret, but involve opportunities for public
discussion.  It is reasonable to assume that the negative
effects on IDNs of making such changes would be part of the
relevant discussions.  If the Unicode decision were to make a
change, there would be adequate time to review whether the
codepoint should be added to the exception list to keep its IDN
behavior stable.

I personally believe that it would be desirable to have a
procedure for making additions to the exception list for that
type of situation that was less heavyweight than spinning up a
WG.  I hope the WG can devise such a procedure.    On the other
hand, if we cannot, or the conclusion is that the likelihood is
sufficiently low that we should not change the trouble, I'm
happy to rely on the ability to revise the "tables" document as
Patrik suggests. 

    john