Re: [EAI] Thinking about requirements / downgrade

["YAO Jiankang" <yaojk at cnnic.cn>]

----- Original Message ----- 
From: "Shawn Steele" <Shawn.Steele at microsoft.com>
To: "YAO Jiankang" <yaojk at cnnic.cn>; "Harald Alvestrand" <harald at alvestrand.no>
Cc: "Mark Davis ⌛" <mark at macchiato.com>; <ima at ietf.org>
Sent: Wednesday, September 16, 2009 12:42 AM
Subject: RE: [EAI] Thinking about requirements / downgrade


>> assumption is just a assumption. never comes true. if we move forward without considering the  "esoteric" one,
>> the protocol might not run properly. if you have a utf-8 address and produce a automaic ACE address which
>> happens to be there already, what will happen?
> 
> Worst case I'd imagine that it'd get converted to Unicode for display.  It should still "work"  I've never seen an address that looks anything like xn--, and while >that isn't saying much, it seems that it'd be worth the risk to whatever does use xn-- to make progress towards Unicode email.

maybe, xn-- isn't saying much for english.  but in Chinese Pingyin, XN can standard a city name "Xi Ning" which is a capital city of province.

If I am from Xi Ning, so I can name my mailbox such as Xn--yao, while Xn standards for the city name, -- as the separator, and yao as my family name.

so xn--yao at cnnic.cn  has a lot of meaning. you might not see such name in your life, but you can not prove that it might be there.



> 
>> That means that the different accounts but equalent will  belong to the different users.
> 
> That's doesn't seem possible.  If we used xn--, then it'd only be subject tothe different accounts condition if the legacy xn-- mail system were upgraded to >support EAI.  In that case the administrator would necessarily either prohibit new mailboxes that matched xn--, or change the existing xn-- addresses.
>

yes, the administrator of the email accounts can  prohibit new mailboxes that matched xn--. but you can not prohibit the email user client and other mail servers to do so. The implementor of the EAI protocols might implement the automatic algorithm to produce the alt-address. How does the implementors know the exception case? if you have a utf8 address which happens to map to the other xn-- address instead of  the xn-- address produced by automatic algorithm, how do the MUA and other SMTP servers know it?  these MUA or SMTP servers might still think that  xn-- address produced by automatic algorithm is still a right one for mapping to the uf8 address.


Yao Jiankang
CNNIC



> -Shawn