Re: [idn] Fw: Cyrillics - Latin

["Soobok Lee" <lsb@postel.co.kr>]

(sorry for my previous posting in bad html settings, i repost this in pure text)
----- Original Message ----- 
From: "Dave Crocker" <dhc@dcrocker.net>
> 
> Domain names are character strings, they are not language.   
>

Domain names are not pure character strings. 
Rather, Domain names are a subset of entire character strings  
*with imposed rules and constraints and requirements*
to serve as identifiers to *name* uniquely and unambigously   
domains in the DNS    for international user communities.

Your claim seems to be insisting that IDN is for *machines*, not for human beings.
machines don't complain about the ambiguity between <latin A>.com and <Cyrilic A>.com. 

Moreover, the problem of IDN itself may become out-of-scope of this WG by introducing the 
out-of-scope *policy* rules. 

Any sound address systems or identifier systems should be self-contained and
self-complete.  Deadly needs for per-zone basis compensating solutions like
policy-based multiple registrations prove and admit that IDN must not be pure strings.
The "policy" cannot be applied or coded or enforced further into numerous independant applications, 
but only into per zone  authoritative servers. 
Such inconsistency due to two-layered (stringprep and then policy) rules will cause 
confusions and troubles.

This WG is not qualified to discuss the success of such "policy", and its discussion is
surely out of scope of this WG. If current IDN solutions rely on such 
out-of-scope and still-unresolved and not-existing and not-verified "policy",
how can we declare this IDN solution is okay and justify last calls for it?



Soobok Lee