ietf-822
[Top] [All Lists]

Re: MIME implementation documentation

1996-08-19 08:24:28

Harald,

I am working on a library project called JEDDS which is about sending
ocuments in e-mail, using a library agreement called GEDI. (I can tell
ou all about any of this further if you'd like to know)


JEDDS is comissioning RLG, a company in MountainView California, to develop
a product; as part of the requirements specification we have designed a
mime message to carry the document. The product will include MUA to both
generate and receive the message, and of course we expect other applications
to use this mime format in time.

The current plan is to have a multipart/x-gedi-document, the first part
of which is a multipart/alternative body, with two alternative 
representations of the GEDI data (information accompanying the document to
be used by inter-library-loan software), one human readable, one machine.

The other body part(s) carry the document in well known formats,
mainly Tiff to start with.

So we will be using alternative and nesting. We intend to put the
body part forward to the GEDI group for adoption and when/if it gets
accepted in time to IANA for registration


I know these are only plans at the moment (and are thus flexible) so

i) may not count from your point of view

but

ii) we need to re-think our plans if MIME changes in ways significant
from our point of view.


JEDDS is an Australian based project, funded by the Australian Vice
Chancellors Committee, and has as members national libraries in Australia,
New Zealand. The British Library is also involved through it's participation
in the UK e-lib projects.

Ruth


 When in doubt, write a Web page....

 I have written up a Web page for the Last Call on

 http://domen.uninett.no/apps/last-call/mime-draft2.html

 and attached a text version to this document.
 Given the debate that has occured, I regard the question of 
 multipart/alternative as settled, while there is still missing evidence for 
 multipart/parallel, generation of nested body parts and handling of external 
 body parts.

 When new evidence arrives (on this list, please!), I will update the Web 
page.

      Harald T. Alvestrand
      Apps AD


 ------- =_aaaaaaaaaa0
 Content-Type: text/plain; charset="us-ascii"
 Content-ID: <23909(_dot_)840267369(_dot_)2(_at_)domen(_dot_)uninett(_dot_)no>


        LAST CALL FOR MIME DOCUMENTS TO RECYCLE AT DRAFT

    The following Last Call has been issued:

 The IESG has received a request to consider the following Protocol
 Actions:


 1. Multipurpose Internet Mail Extensions (MIME) Part One:  Format of
    Internet Message Bodies"  for the
    status of Draft Standard.

 2. Multipurpose Internet Mail Extensions (MIME) Part Two:  Media Types
     for the status of Draft
    Standard.

 3. MIME (Multipurpose Internet Mail Extensions) Part Three: Message
    Header Extensions for Non-ASCII Text
     for the
    status of Best Current Practice.

 5. Multipurpose Internet Mail Extensions (MIME) Part Five:  Conformance
    Criteria and Examples  for the
    status of Draft Standard



 The IESG plans to make a decision in the next few weeks, and solicits
 final comments on this action.  Please send any comments to the
 iesg(_at_)ietf(_dot_)org or ietf(_at_)ietf(_dot_)org mailing lists by August 
8, 1996.

    No objections to the Last Call were received before August 8.

    After the timeout date for the Last Call comments, the following
    issues with regard to implementation experience have been raised:
* Is multipart/alternative widely implemented for generation?
* Is multipart/alternative widely implemented for reception?
* Is multipart/parallel widely implemented?
* Are Nested Body Parts widely implemented?
* Are External Body Parts widely implemented?

    The issue here is if the "two implementations" rule (which
    traditionally have allowed two experimental implementations without
    terribly useful user interfaces) should be strengthened for a
    specification of this maturity to mean "two interworking, commercial
    implementations that can be used by someone not terribly skilled in
    the arcana of MIME".

    The sections below will give the arguments and evidence in each case.

 Is multipart/alternative widely implemented for generation?

    The question is if there exist MUAs that can generate a MIME
    multipart/alternative without excessive thinking on the part of the
    user.

    The following MIME UAs have been claimed to do so:
* Cyberdog for the Macintosh
* Ishmail
* An yet-unnamed product from Microsoft

    This AD regards the question as settled.

 Is multipart/alternative widely implemented for reception?

    The question is if there exist UAs that display the "best" body part
    (that being defined as the last part of the multipart/alternative that
    the UA is able to display).
    The following UAs have been claimed to do so:
* Pine
* Metamail
* Mac Eudora 3.0

    This AD regards the question as settled.

 Is multipart/parallel widely implemented?

    The question is if there are UAs that display all parts of the
    multipart/parallel "at the same time", as described in the drafts, and
    do not jsut treat it like multipart/mixed, and if there are generating
    UAs.

    To date, no generating UA has been identified.

    To date, only Metamail has been mentioned as an example of a
    displayer.

 Are Nested Body Parts widely implemented?

    Many MUAs work according to the "one message with attachments"
    metaphor, which is not the same as the MIME "bodyparts may nest to any
    level" structure. Forwarded messages are not part of this question.

    One may argue that the "inline/attachment" distinction that is carried
    in content-disposition is missing functionality in MIME. However,
    content-disposition is the subject of another ongoing action, and will
    have to catch up later.

    The question is if there exist MUAs that:
     1. Generate nested multiparts through a reasonable user interface
     2. Usefully display nested multiparts

    The following MUAs have reasonable support for generating nested
    multiparts:
*

    The following MUAs have reasonable support for displaying nested
    multiparts:
* EXMH 1.6.7

 Are External Body Parts widely implemented?

    The question is if there exist MUAs that:
     1. Generate External Body Parts through a reasonable user interface
     2. Usefully handle External Body Parts

    The IETF internet-drafts announcement uses multipart/alternative with
    external body parts to list the ways in which one can get the I-Ds.
    This is special purpose code, and only proves that such messages can
    be generated.

    The following MUAs have reasonable support for generating External
    Body Parts:
*

    The following MUAs have reasonable support for handling External Body
    Parts:
* MH 6.8.3


_________________________________________________________________

     Harald(_dot_)T(_dot_)Alvestrand(_at_)uninett(_dot_)no

    Last modified: Sat Aug 17 09:32:26 1996

 ------- =_aaaaaaaaaa0--

===============================================
Ruth Moulton            ruth(_at_)muswell(_dot_)demon(_dot_)co(_dot_)uk
Consultant              

65 Tetherdown, 
London N.10 1NH, UK     Tel:+44 181 883 5823

-