Re: [EAI] UTF-8 in Message-IDs

["Charles Lindsey" <[e-mail]>]

On Tue, 04 Oct 2011 02:42:57 +0100, John Levine <[e-mail]> wrote:

>>> If then, as is likely, they start to appear, then we will see how they
>>> cope and decide how best to regulate them in a future standard.
>>
>> Meanwhile, the propagation of such articles will be very bad.
>
> News servers aren't going to be thrilled about messages with
> unencoded UTF-8 in other headers, either.
>
> They'll have to upgrade to handle EAI, just like mail servers will.  I
> suppose we might want to invent some new NNTP verb that lets hosts
> tell each other whether they can accept EAI messages, but this is most
> definitely not the current EAI group's problem.

No they won't. User agents will have to be upgraded if people want to read  
the utf-8 stuff, or to reply by mail to articles From: people with utf-8  
addresses. But the existing servers will handle all this just fine except  
insofar as they might be rejecting (unnecessarily) utf-8 in some headers  
where it did not matter.

Servers will expect to be able to compare Message-IDs for equality by a  
simple octet-by-octet comparison, and they will expect Dates to be  
comparable (<,=,>), but I don't think EAI has changed that. Even Newsgroup  
names with utf-8 in them will work, as has alredy been demonstrated on the  
existing network, (so long as octet-by-octet comparison works, which means  
some severe normalization).

As to whether it is any of this WG's business, it is certainly NOT the  
business of this WG to put obstacles in the way of other mail-like  
protocols which might want to follow the EAI path.

-- 
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: [phone]                       
   Web: http://www.cs.man.ac.uk/~chl
Email: [e-mail]      Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9      Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5