Re: Subaddressing

--On Saturday, 15 February, 2003 22:55 +0100 Marc Mutz 
<[email protected]> wrote:

> On Saturday 15 February 2003 22:23, Marc Mutz wrote:
>> So the requirements to balance are
>> 1. keeping complexity out of the server for the reasons
>> stated above 2. keeping the concept of opaque local-parts in
>> all it's strictness.
>
> In the light of John's recent mail, I should add that I'm
> talking about  keeping or not keeping the concept of opaque
> local-parts /for IMAs/,  iow, whether or not to _extend_ the
> concept to ILPs.
>
> I'm specifically not trying to start a discussion whether or
> not opaque  local-parts as currently used are a bad idea -
> they're not. Apologies  if my mail came across like that.

I think we are getting close to figuring out where we agree or 
disagree, which is progress even if we can't agree.

When I read the above, my reaction is "MTAs don't read minds". 
You presumably don't intend that they do, so I probably don't 
understand your suggestion.

Unless there is an external negotiation, or something more 
drastic, both traditional local parts and ones containing i18n 
strings are going to enter and leave MTAs in Forward-path and 
Reverse-path fields containing a Mailbox.  If the presence of 
i18n material is not identified to the MTA, than any information 
that identifies that a particular address is an IMA must be 
embedded in the Local-part (obviously, an ACE prefix is such an 
embedded indication).  However, if the Local-part of the 
relevant Mailbox is opaque, then the MTA can't tell whether what 
is in it is a traditional (opaque) local part or an 
internationalized one.  In the limiting case, if the MTA peeks, 
it still can't be guaranteed that IESG--string is an IMA rather 
some sort of traditional address or instruction.  After all, it 
is just an ASCII string that obeys the quoting rules.

So, I don't see how you can have non-opaque IMAs without making 
traditional local-parts non-opaque.  Unless, again, there is 
some sort of handshake at the SMTP level that tells the MTA what 
is an IMA and what isn't.

What am I missing?

       john