Re: The Case Against MIME

[Ian Brown <I.Brown@xxxxxxxxxxxx>]

Jeremey Barret wrote:

> Plain old non-MIME ASCII-armored PGP is one of the most 
> useful crypto tools out there, not to mention the sole basis of PGP
> integration into most mail readers that support it.

This actually misses one of the main benefits of MIME. In almost all
cases, a MIME-compliant mailer will do the decoding for you. If PGP/MIME
had a specific pgp-data type bodypart, *any* MIME-compliant mailer would
be configurable, upon receiving such a bodypart, to:

a) Remove the base64 (or whatever) encoding and
b) present the resulting binary data to any OP/PGP program

William Geiger wrote:

> As far as KISS what can be more simpler than rfc822 with a PGP Ascii
> Armor Block?

Binary data handed straight to your program?

This has further benefits:

Greg Vaudreuil wrote:

> One of the strongest reasons to use MIME encoding, even for plain text 
> messages is MIME's ability to support multiple languages and character 
> sets. While I am new to PGP, I do not believe that multi-lingual 
> support is part of PGP 2.6.X.

There is some support for using different codepages for text. But Greg
points to the main problem with continued use of armour - we have to
continue solving exactly the same problems as the MIME people. Different
character sets (Unicode would be very useful, for instance), multi-part
messages, and so on are issues armour and MIME both have to deal with.
Why not just leverage the work the MIME people are doing?

> Multilingual support is an IESG requirement for all IETF protocols 
> going forward.  At a minimum this requires identification of the 
> selected character set, and in the mail context, a way to encooding 
> these 8 or 16 bit characters into a 7 bit transfer format such as 
> Quoted Printable.

I don't think PGP has an easy method of encoding 16-bit systems such as
Unicode.

Ian.