ietf-822
[Top] [All Lists]

Re: Default Dispositions

1999-10-21 12:25:27
I agree with you, John... besides I already have proof of misuses of this 
kind of thing when totally unnecesary.

It seems there is some version (or configuration or something) of Lotus 
Notes' SMTP gateway that does the following:
When it gets a Notes mail to relay to Internet (SMTP/MIME world) it 
assumes it's 7bit USASCII, if later on it realizes there is an accented 
character, it generates a second part of a multipart message with contat-
disposition: inline where the second part is quoted-printable (or 8bit) 
iso-8859-1.

Note that USASCII is a proper subset of ISO-8859-1 and I think it's clear 
for any reader of any version of the MIME specification that what should 
be done here is use a single part with quoted-printable iso-8859-1. 
Result: since content-disposition is not part of the basic MIME 
compliance requirement there are lots of MIME readers that don't 
understand this (#3) and you end up seeing an awful #1 or #2 splitted in 
the middle of a word.

I would vote for either #1 or #2 going as default just to discourage its 
misuse (and would like to see a comment in the RFC stating: "don't use it 
for this kind of thing since you should'n").

El 19 Oct 99, a las 10:29, John C Klensin escribió:

--On Tuesday, 19 October, 1999, 08:46 +0000 Charles Lindsey
<chl(_at_)clw(_dot_)cs(_dot_)man(_dot_)ac(_dot_)uk> wrote:

Option #3 seems like the most reasonable interpretation to
me.  If you have >a multipart/mixed, and a part doesn't have a
content-disposition, it makes >the most sense to me to display
the part inline if you can and as an >attachment if you can't.

I am surprised (but a couple of other responses that have not
reached the list yet also support that view).

My chief concern with #3 is that the reader get no indication
that the message was a multipart. He just sees the texts
following on with no apparent break (even on the same line, if
the author was careful to arrange it that way). Are you happy
with that?

My initial impression was "silly", for just that reason, but
don't have time right now to compose an in-depth explanation.
The only strong argument I can think of for it is that it would
force a lot more use of content-disposition, which may or may
not be desireable.  The notion of using multiparts to represent
different charsets/ languages into a single message flow was
discussed during the MIME development process and rejected ---
that sort of thing really needs to be done by markup, not
multipart structures and the hope that the receiving MUA will do
something very specific.

    john



--
Mariano Absatz - El Baby
mailto:baby(_at_)baby(_dot_)com(_dot_)ar
http://www.baby.com.ar/
    PGP KEYS: http://www.baby.com.ar/datos/personales.html#claves_pgp
  |\  _
  _\\/'>     Powered by Pegasus Mail
 /|__)       http://www.pmail.com
  ) )\
-----------------------------------------------------------
If at first you don't succeed, call it version 1.0 


<Prev in Thread] Current Thread [Next in Thread>