Re: Registration of new charset BOCU-1 refreshed - UTF-8

[Markus Scherer <markus.scherer@jtcsv.com>]

Harald Alvestrand wrote:
 > ...

First things first:

 > And this is only one of many places where one uses charsets in protocols.
 > I think you should add Martin's warning to your registration - possibly
 > reformulated as follows (line 2 added):
 >
 > BOCU-1 is intended for limited use in special situations
 > where the use of this charset can be preconfigured or negotiated.
 > The preferred and most widely supported encoding for
 > Unicode/ISO 10646 on the Internet is UTF-8.
 >
 > OK?

OK.

> Remember that we have zero (none, nada, nil, zilch) generally supported 
> ways of figuring out what charsets the recipient of an email supports.


Well, if an SMTP email system does not support the Content-Type: header, then this is true for all charsets.
(Except that you can use heuristics for just about everything, but that's cheating.)


> Thus, the first email client that is capable of supporting BOCU-1 will 
> be capable of sending mail that no other email client in the world can 
> display legibly, and *has no way of knowing when they become capable of 
> doing so*.


True. I am aware of that.


markus