Re: [idn] Canonicalization: [28] through [31]

[Paul Hoffman / IMC <phoffman@imc.org>]

At 11:06 PM +0800 6/26/00, James Seng wrote:
>Okay, I took for granted that we can do NFKC/NFD then case fold to do
>canonicalization.

That may become the implementation we chose, but I do not believe 
that is a requirement for IDN.

Further, as Kent says, UTR21 is (I believe purposely) separate from UTR15.

>Anyway, whats your suggestion on the rewording?

My original proposal was:

>  If the protocol requires canonicalization, it should be done before
>  the request enters the DNS service interface.

You are saying that, because we have consensus on canonicalization, 
we should take out the "if" clause. How about:

The protocol MUST specify canonicalization, and the canonicalization 
SHOULD be done before the request enters the DNS service interface.

--Paul Hoffman, Director
--Internet Mail Consortium