[EAI] MIME questions (was: draft-hurtta-eai-messagestore-00.txt)

[Frank Ellermann <nobody at xyzzy.claranet.de>]

Kari Hurtta wrote:

 ["what is MIME"] 
> Was that discussed on meeting?

Don't think so, I only know the jabber log:
http://www3.ietf.org/meetings/ietf-logs/eai/2007-03-20.html

> You was suggested to ask from MIME experts that
>        Is UTF-8 possible in MIME version 1.0 header fields ?
> What was result?

Nothing, unless "[10:39:54] --- nsb has joined" 
indicates something that's not visible in the log.

> My feel is that
>    - If UTF-8 can be added to address and subject
>      header fields without label, then UTF-8 can be added also
>      to MIME header fields without label.

Yes, that's where we disagree, because it makes the
identification of a message/utf-8 too difficult:

Looking into the header is one thing, but parsing
the body for MIME part headers is another thing:

Looking for somebody else's trouble, none of our
business, what if the MIME structure is broken,
should that affect message/utf-8 transport, etc.

> MIME header field syntax is already updated
> without changing MIME-Version header field.

In an IETF theory styling itself as "RFC 2231".

It's IMO possible to implement 2049 ignoring
2231, and treat 2231 "features" as garbage.

Of course that software then couldn't claim to 
support the (approved) NetNews message format,
but that RFC won't be published before another
RFC is ready... <sigh />

> So changing MIME header field syntax yet one
> more time is not different.

2231 is in a certain sense "compatible" with
MIME 1.0, while raw UTF-8 in Content-* header
fields directly _violates_ a requirement in 
2045 (or whereever it was, we quoted it here
often enough).

Show me that Keith, Ned, and "nsb" (or maybe
Bruce overruling them all :-) say "no problem",
and I might consider an "it's only MIME, we're
free to screw with it" approach... <shudder />

[[ I'd still claim that it's unnecessary, and
   that it should be done in a proper 2045bis
   after EAI turned out to work as expected ]]
  
> Label
>   MIME-Version: 2.0
> may be usefull to indicate that downgrade is
> required also for header fields on MIME
> structure.

Yes, and it would warn old MIME parsers that
they might stumble over some surprises (in the
form of raw UTF-8) if they try to parse "2.0"
as "1.0".  We could also formally require 2231
support for MIME-Version 2.0, just to be sure.

Maybe the result would be more like a 2049bis
than a 2045bis.  

Frank



_______________________________________________
IMA mailing list
IMA at ietf.org
https://www1.ietf.org/mailman/listinfo/ima