Re: A couple of comments on the open issues...

I used the term "interpreting" differently from the way Dan 
understood it.  But his interpretation is reasonable, so I 
apologize for any confusion I may have caused.

Let me restate my assertion: 821/822 and 2821/2822 specify the 
_syntax_ of the local part.  That syntax specification, to all 
intents and purposes, limits the local-part string to ASCII 
graphics and specifies the quoting conventions.

	(If you care about what I'm glossing over here, and
	don't know, I strongly suggest you go read the
	specifications rather than taking either my word or
	Dan's for it.  In reading them, remember that 821 and
	2821 specify the addresses that can be used to transport
	and deliver Internet mail.  Additional flexibilities
	permitted in message bodies are fine, but you may not be
	able to get the additional forms there unless you use a
	non-Internet (or non-conforming) mail transport. )

However, the definition in terms of a sequence of ASCII 
characters (with or without quoting) is a rather low-level piece 
of syntax.  There are no productions for disaggregating or 
organizing the local-part into specific subdivisions such as 
"user name", "subaddress", "embedded route", or the like.  And 
it was in the sense of being able to look at (or parse) a 
local-part and say, e.g., "ok, this is a user name and that is a 
subaddress" that I used the term "interpret".  And it is 
"interpretation" of the local-part in that sense, by anything 
but the delivery MTA, that is prohibited.  The _syntax_ of local 
parts consists of a string of ASCII characters and some quoting 
rules but, except for the delivery MTA, one cannot interpret 
those characters to suggest any particular meaning, nor can one 
subdivide them into units with particular meaning.

Except to the delivery MTA, they are just a character string, 
albeit one restricted to ASCII and with some rules about quoting.

     john



--On Saturday, 15 February, 2003 02:53 +0000 "D. J. Bernstein" 
<[email protected]> wrote:

>
> John C Klensin writes:
>> There are no standards at all for interpreting the local
>> part.
>
> False.
>
> RFC 822 clearly states that the local part consists of ASCII
> characters. The ASCII standard specifies interpretations of
> bytes 33-126 as glyphs.
>
> (If you're under the delusion that this interpretation is not
> actually standardized, or that users don't rely on it, please
> switch your MUA to EBCDIC header I/O for a month and let us
> know the results.)
>
> The question of which mailbox names are valid on a system, and
> how those mailboxes are handled, is (almost) entirely up to
> the system. But we have a global interpretation of ASCII
> mailbox names as glyphs.
>
> ---D. J. Bernstein, Associate Professor, Department of
> Mathematics, Statistics, and Computer Science, University of
> Illinois at Chicago