
Archive-Name: mail/mime-faq/part3
Version: $Id: mime3,v 3.6.0.2 1994/07/06 06:44:31 jsweet Exp $
Posting-Frequency: monthly


comp.mail.mime frequently asked questions list (FAQ) (3/3)

Part III -- Advanced Topics

This is part III of a Frequently Asked Questions document about MIME, the
multipurpose and multi-media standard for Internet mail.

Part I covers frequently asked questions.

Part II is a listing of MIME products.

Part III covers advanced topics.


10 Information

10.1 MIME-relevant RFCs and other standards

The RFCs mentioned here are mainly relevant to persons building MIME
software.  As an end user, if your mail system is nice to you, you
won't really have to know very much about these things.

RFC and Internet-Drafts are available by anonymous FTP from any decent
archive site.  If you're really stuck, try

FTP:      ds.internic.net:rfc/*
FTP:      ds.internic.net:internet-drafts/*

MIME is defined in RFC 1521 (MIME Mechanisms for Specifying and
Describing the Format of Internet Message Bodies) and RFC 1522
(Representation of Non-ASCII Text in Internet Message Headers).

These are Internet standards-track protocols.  For the full
implications of this, see RFC 1540 (IAB Official Protocol Standards).
Here is their current status.

    1521: Draft Elective Standard
    1522: Draft Elective Standard

These two RFCs do not fully define MIME.  For one thing, they are
based on RFC 822 (Standard for the format of ARPA Internet text
messages), as revised by RFC 1123 (Requirements for Internet hosts -
application and support) and must be read in conjunction with these.

For another, they are extensible.  See 10.2 for a complete list of
registered subtypes.

There are a whole lot of other RFCs that deal with e-mail, including
these.

IAB standards-track RFCs

    1502  X.400 Use of Extended Character Sets
    1496  Rules for Downgrading Messages from X.400(88) to X.400(84)
          when MIME Content-Types are Present in the Messages
    1495  Mapping between X.400 and RFC-822 Message Bodies
    1494  Equivalences between 1988 X.400 and RFC-922 Message Bodies
    1427  SMTP Service Extension for Message Size Declaration.
    1426  SMTP Service Extension for 8bit-MIMEtransport.
    1425  SMTP Service Extensions.
    1424  Privacy Enhancement for Internet Electronic Mail: Part IV.
    1423  Privacy Enhancement for Internet Electronic Mail: Part III.
    1422  Privacy Enhancement for Internet Electronic Mail: Part II.
    1421  Privacy Enhancement for Internet Electronic Mail: Part I.
    1327  Mapping between X.400(1988)/ISO 10021 and RFC 822.
    1314  File format for the exchange of images in the Internet.

Other RFCs (Informational, Experimental, or Historical)

    1489  Registration of a Cyrillic Character Set.
    1468  Japanese Character Encoding for Internet Messages.
    1456  Conventions for Encoding the Vietnamese Language.
    1428  Transition of Internet Mail from Just-Send-8 to 8bit-SMTP/MIME.
    1357  Format for emailing bibliographic records.
    1345  Character Mnemonics & Character Sets.
    1344  Implications of MIME for Internet mail gateways.
    1343  User agent configuration mechanism for multimedia mail format
          information.
    1339  Remote mail checking protocol.
    1321  MD5 Message-Digest algorithm.
    1225  Post Office Protocol: Version 3.
    1211  Problems with the maintenance of large mailing lists.
    1176  Interactive Mail Access Protocol: Version 2.
    1197  Using ODA for translating multimedia information.
    1154  Encoding header field for internet messages.
    1153  Digest message format.
    1049  Content-type header field for Internet messages.
    1036  Standard for interchange of USENET messages.
    934   Proposed standard for message encapsulation.
    807   Multimedia mail meeting notes.


10.2 MIME types

There are registered and unregistered MIME types.  Unregistered MIME
types begin with an "x-" and their meanings generally depend on
private agreements between senders and receivers.  This section lists
registered types and some known unregistered types.

10.2.1 List of registered MIME types

The latest list of registered MIME types is available from

FTP:      isi.edu:in-notes/media-types/media-types


Type            Subtype                Description          Reference
----            -------                -----------          ---------
text            plain                                       [169,NSB]
                richtext                                    [169,NSB]
                tab-separated-values                   [Paul Lindner]

multipart       mixed                                       [169,NSB]
                alternative                                 [169,NSB]
                digest                                      [169,NSB]
                parallel                                    [169,NSB]
                appledouble                [MacMime,Patrik Faltstrom]
                header-set                             [Dave Crocker]

message         rfc822                                      [169,NSB]
                partial                                     [169,NSB]
                external-body                               [169,NSB]
                news                        [RFC 1036, Henry Spencer]

application     octet-stream                                [169,NSB]
                postscript                                  [169,NSB]
                oda                                         [169,NSB]
                atomicmail                           [atomicmail,NSB]
                andrew-inset                       [andrew-inset,NSB]
                slate                           [slate,terry crowley]
                wita              [Wang Info Transfer,Larry Campbell]
                dec-dx            [Digital Doc Trans, Larry Campbell]
                dca-rft        [IBM Doc Content Arch, Larry Campbell]
                activemessage                          [Ehud Shapiro]
                rtf                                    [Paul Lindner]
                applefile                  [MacMime,Patrik Faltstrom]
                mac-binhex40               [MacMime,Patrik Faltstrom]
                news-message-id             [RFC 1036, Henry Spencer]
                news-transmission           [RFC 1036, Henry Spencer]
                wordperfect5.1                         [Paul Lindner]
                pdf                                    [Paul Lindner]
                zip                                    [Paul Lindner]
                macwriteii                             [Paul Lindner]
                msword                                 [Paul Lindner]
                remote-printing                         [RFC1486,MTR]

image           jpeg                                        [169,NSB]
                gif                                         [169,NSB]
                ief             Image Exchange Format      [RFC-1314]
                tiff            Tag Image File Format           [MTR]

audio           basic                                       [169,NSB]

video           mpeg                                        [169,NSB]
                quicktime                              [Paul Lindner]

Each <type> has a directory

FTP:      isi.edu:in-notes/media-types/<type>

containing the definitions of its subtypes.


10.2.2 List of known unregistered MIME types

Here is a list of some known x-types, x-subtypes, and x-parameters.

The enumeration of these x-types here does not imply any kind of
standardization or open specification.  The meanings of x-types depend
on private agreements between senders and receivers.  Some x-types may
eventually become registered types; see sections 10.2.1 and 11.1.

Type         Subtype        Parameter     Description                  Reference
----         -------        --------      -----------                  ---------
application  octet-stream   x-conversions for compressed data     MH 6.8 viamail
       
audio        x-macaudio                   NOT sampled Macintosh audio      Iride
             x-next                       self-describing audio data      MH 6.8
                
image        x-macPICT                    Macintosh PICT resource          Iride


{ Please feel free to let us know about whatever whacko or
not-so-whacko x-types that your UAs may unleash on an unsuspecting
world. }


10.3 Internet Engineering Task Force (IETF) working groups

The IETF working group on Privacy-Enhanced Mail (PEM) has developed
extensions that permit confidentiality, authentication, and integrity to
be provided in a manner backwards compatible with RFC-821 and RFC-822
(see 3.4).  Work is underway to align PEM and MIME which will provide
real security to MIME e-mail.

The IETF MIME working group is not actively considering significant
changes to the specifications.  However the WG still exists as a forum
for MIME developers, as a home for interpretation questions, and to
handle any problems or ambiguities that might arise in MIME.


11 Developers' FAQs

11.1 How can I register a new MIME type?

The procedures for registering new content types, character set
values, access types, and conversions parameters with IANA (the
Internet Assigned Numbers Authority) are documented in RFC 1521.


11.2 What's ESMTP, and how does it affect MIME?

ESMTP (Extended Simple Mail Transfer Protocol) is a mechanism by which
extensions to "traditional" (RFC 821) SMTP can be negotiated by client
and server.  The mechanism (RFC 1425) is open-ended; so far two
extensions have been defined.

Message size declaration (RFC 1427) offers a graceful way for servers
to limit the size of message they are prepared to accept.  (With SMTP,
the only possibility is for the server to discard the message after it
has been sent in its entirety.  There is no way for the client to know
that it was the size of the message that caused the problem.)

When a message is returned to the user as being too large to deliver,
one possible approach might be to fragment the message using the MIME
Message/Partial mechanism, and resubmit it.

Depending on the exact reason for the "too large" rejection, this may
or may not be a good idea.  For example, the limitation may reflect
the recipient's disk quota, in which case the fragmented message will
not be fully deliverable either.

The possibility of fragmentation should, therefore, be left to the
user's discretion (not performed automatically by the SMTP client).

8bit-MIMEtransport (RFC 1426) opens up the possibility of sending 8bit
data in mail messages, without having to use base64, quoted-printable,
or another encoding, and without the breakage that can result from
sending 8bit data to an unsuspecting RFC 821 SMTP server.  RFC 1428
(Transition of Internet Mail from Just-Send-8 to 8bit-SMTP/MIME)
discusses some of the implications of this.


11.3 Where can I get some sample MIME messages?

FTP:      thumper.bellcore.com:pub/nsb/samples/*


11.4 Wouldn't MIME be better if it did <foo>?

This question is asked for various values of <foo>.  Perhaps the most
common is "multilevel encodings": see the next question.  There are
a couple general points that apply to all <foo>.

1. Please remember that MIME is the result of a lot of work by a lot
of persons, over a long time (look at the Acknowledgements section of
RFC 1521).  A great many ideas, probably including yours, were
considered.  In many cases, there were conflicting goals, such as
simplicity and interoperability on the one hand, and power and
flexibility on the other.

2. If you really think you've got an original idea which would improve
MIME, the correct place to pursue it is not this newsgroup, but the
working group mailing list (having first read the archives, to check
that it really is new).  Yes, this is going to be a lot more work than
posting a news article.


11.5 So what about multilevel encodings?

MIME uses a two-level encoding scheme.  The original object (for
example, a picture, or a text document) is encoded using a well
defined mechanism appropriate to that object (perhaps GIF for the
picture, and text/enriched for the document).  Then a second encoding
is used to ensure that the first encoding can be transmitted intact
(probably base64 for the GIF, and quoted printable for the
text/enriched document).  

Note that there is a very small number of the second encodings (five,
but three of these are simply indications of what kind of data an
unencoded body part contains), and it is not expected that there will
be many more in the foreseeable future.

The multilevel encodings idea is for a more generalized MIME-like
encoding mechanism that could indicate many arbitrary transformations
of the original object.  For example,

    Content-Type: application/tar; conversions="encrypt,compress,uuencode"

might indicate a UNIX tar file that had been encrypted, then
compressed, then uuencoded.  (This is a fictitious example of how MIME
might have worked; it's not legal MIME.  Don't worry if you've never
heard of some of these transformations.)

This may look like an attractive scheme at first, but it has a number
of problems.

1. If you've been brought up on UNIX and command pipelines, the
implementation of such a scheme seems trivial.  Surely any half-decent
machine can do something similar?  Unfortunately, this turns out to be
true only for a very restricted definition of "half-decent".  In
practice, it would be awfully difficult to implement this on a lot of
systems.  Probably even more systems would not allow new
transformations to be just "slotted in", and would require
recompilation or reshipping whenever a new one came along.

2. Each successive transformation reduces the size of the audience who
can successfully decode the message.  Every MIME mailer must be able
to decode base64 and quoted-printable, so it's guaranteed that you can
at least get back to the raw data.  What if, in the above example, I
have tar, decrypt, uudecode, but no uncompressor?

3. Such a scheme does not increase the scope of the framework defined
by MIME.  If uuencoded, compressed, encrypted tar files are useful
things to sling around, it is entirely possible to define a new MIME
type (presumably a subtype of application) to handle them.


11.6 Why doesn't MIME include a mechanism for compression?

Compression is a difficult area.  It was considered by the working
group, but no consensus was reached.  There is still work going on in
this area: there may someday be a compressed-64 encoding.

Most compression algorithms have one of more of these undesirable
properties: they are covered by patent, they require the ability to
treat the input as a stream of bits, they use a large data space.  The
chances of finding a truly interoperable compression algorithm are
therefore rather slim.

It is worth noting that most or all of the image and video subtypes
(including GIF, JPEG, TIFF, and MPEG) define their own compression
schemes.


12 Acknowledgements

Many persons have contributed to this document.

They include: 

Alan Robiette, Alec Henderson, Axel Boldt, Carlyn Lowery, Chris Pepper,
Christophe Wolfhugel, Christopher Davis, Craig Huckabee, Daniel
Fandrich, Daniel Glazman, Dave Curry, Dave Lacey, David Barr, David
Collier-Brown, David Miller, Douglas Boyce, Ed Anselmo, Ed Greshko,
Edward Vielmetti, Erik van der Poel, Gisle Hannemyr, Harald Alvestrand,
Ian Hoyle, James Ford, Jason Beyer, Jay Weber, Jerry Peek, Jerry Sweet,
Joe Ilacqua, Joergen Haegg, John Gardiner Myers, John Martin, John R
MacMillan, John Romine, Joyce Reynolds, Keith Moore, Larry Salomon Jr,
Lars-Gunnar Olsson, Luc Rooijakkers, Marc VanHeyningen, Mark Crispin,
Mark Grand, Marshall Rose, Martin Wendel, Masanobu Umeda, Michael
Parson, Michael Urban, Nathaniel Borenstein, Ned Freed, Niklas Agren,
Olle Jarnefors, Pat Farrell, Paul Eggert, Piero Serini, Quentin Smart,
Ran Atkinson, Ray Langford, Rich Ragan, Rick Troth, Ron Barak, Sascha
Wildner, Steve Dorner, Steve Hole, Stuart Lynne, Susan Straub, Syd
Weinstein, Tim Goodwin, Tim Kehres, Tommy Wallo, Yehavi Bourvine.

If we've left your name off, please accept our apologies.  Drop us a
note and we'll include it for next time.

Thanks also to the University of California, Irvine, and to Irvine
Compiler Corp. for providing the resources for maintaining this list;
and to Jonathan Kamens, for coordinating the *.answers groups,
and for his post_faq program which brought you this FAQ.


End of Part III
