NEW MIME DRAFT

I have produced an updated version of the MIME document.  Unlike previous such
announcements, this time I am announcing a relatively minor set of changes,
based on the comments received since the version I announced at the beginning
of December.  It is my hope that this is the version that will become the
Proposed Standard.

Basically, the changes are almost all minor.  The
three areas in which I anticipate that controversy might continue are the
first three changes discussed in the list below.

There are also two
open questions, which appear at the bottom of the change list.

As
usual, you can ftp this anonymously from thumper.bellcore.com, where it can be
found as

pub/nsb/BodyFormats.ps
- or -
pub/nsb/BodyFormats.txt

Changes In December 30 Version Of RFC-XXXX

Changes that might be controversial

CHARACTER SETS: This is, as we all know, a mess. I have tried to effect a compromise between what I see as the two extremes represented by Dave Crocker and Keld Simonsen. The new draft removes the definition of the values "iso-10646" and "mnemonic", but retains the 8859- family, US-ASCII, and ISO-2022-jp. I think this makes a lot of sense as a "principled" solution to this dilemma -- it retains only the character sets that are already in use in email today, defining strings with which to label that usage. The new draft adds some new prose about "private" character set values, which must begin with "X-", and refers the reader to RFC-CHAR for a discussion of what other character sets there are and how to name them, i.e. as "X-foo" where "foo" is defined in RFC-CHAR.

EXTERNAL-BODY: More parameters are now defined, and a hyphen is added to the actual subtype name. More important, an access-type of "mail-server" has been defined, along with an example of how it works.

MIME-VERSION: Changed to 1.0, but retained. Changed syntax to RFC 822 "text".

Changes I don't expect to be seriously
controversial

QUOTED-PRINTABLE: Picked up Alain Fontaine's new definition of this encoding, which is a clarification but not a change.

RICHTEXT: Removed iso-646 and mnemonic keywords. Fixed a bug (!) in the richtext-stripper in the appendix, thanks to Guido van Rossum. Added "paragraph" support to the richtext-stripper, as suggested by Alain Fontaine. Added a little clarifying prose about white space.

HEADER REASSEMBLY RULES: The rules for reassembling the headers on message/partial and message/external-body have been simplified as per Marshall Rose.

MPEG: Replaced article reference with reference to draft standard.

DETAILS: The usual set of nits & typos in the draft, as pointed out by the ever-alert ietf-822 readers, have also taken place. Actually, there were a lot fewer than usual. Special thanks on this round go to Dave Crocker, Greg Vaudreuil, Alain Fontaine, Marshall Rose, and Guido van Rossum. I think the three non-ASCII are now reasonable in BOTH the postscript and ascii version of the acknowledgements.

Unchanged Issues To Be Resolved

Reconcile the quoted-printable encoding with the "Q" encoding in RFC-HDRS? I vote for removing the special interpretation of "_" in "Q".

The current document makes reference to the following internet drafts: RFC-HDRS, RFC-NETFAX, RFC-CHAR. Do we need to get rid of these references in order to satisfy the

IAB?