Stability of valid IDN labels

[Eric Brunner-Williams]

Mark,

I have the impression that there are alpha-2 code points that are absent 
from the root zone, e.g., "SU", and that 3166, as absurd as it may be, 
defines the syntax for the valid subset of all octet pairs(*), or 
alpha-2 code points.

However, your larger point is that there is a well-defined requirement 
for "/syntactic /validity forever", and you can share that with me, 
because I don't know where the requirement originated.

Eric