Charter changes and a possible new direction

[Kenneth Whistler]

Patrik wrote:

> > Surely if that's the case (and I don't deny it), a system as
> > complicated as IDNA2008 is going to encounter some trouble.
> 
> Btw, one more thing...I must ask you one thing (and warn everyone else  
> that I will ask them the same question ;-) ):
> 
> You claim IDNA2008 is complicated. What do you include in that  
> statement? When I have implemented IDNA2008, the problems has been in  
> for example (if you look at the tables document) implementation of  
> this rule:
> 
> 2.2. Unstable (B)
> 
>       B: toNFKC(toCaseFold(toNFKC(cp))) != cp
> 
> *THAT* is not easy to implement, and very hard to get right (and  
> understand).

This is somewhat orthogonal to the evaluation of Paul's
suggested alternative or the charter questions per se.

But regarding this particular rule, while it is certainly
true that implementing Unicode normalization and Unicode
casefolding from scratch are complicated, difficult to get
right and to understand, I rather doubt that most implementations
of IDNA2008 would attempt that. There are perfectly good
independent libraries that implement Unicode normalization
and Unicode casefolding, and engineers would be well-advised
to simply use them.

It is not that Rule 2.2 per se is complex or that its intent
in IDNA2008 is complex. Rather it is that implementing an
API for toNFKC(cp) is complex. If you are just *using* such
an API, its function in IDNA2008 is rather straightforward.

--Ken