When I send a message via e-mail with the following subject:
Subject: "Der Führer" was originally the German word for "The leader"
It gets by my mail program converted to:
Subject: =?iso-8859-1?Q?=22Der_F=FChrer=22?= was originally the German word for
"The leader"
My mail program then notes that this line is longer than 60
characters and splits it as follows:
Subject: =?iso-8859-1?Q?=22Der_F=FChrer=22?= was originally the
German word for "The leader"
The recipient of my e-mail then gets the following subject:
Subject: "Der Führer" was originally the
German word for "The leader"
When someone using the same mail program writes a reply to
this message, the mailer will first convert the Subject to
Subject: Re: =?iso-8859-1?Q?=22Der_F=FChrer=22?= was originally the
German word for "The leader"
Then the mailer finds that the first line again is longer
than 60 characters, and splits the Subject to be sent as
follows
Subject: Re: =?iso-8859-1?Q?=22Der_F=FChrer=22?= was originally
the German word for "The leader"
This means that the final recipient will receive reply
looking as follows:
Subject: Re: "Der Führer" was originally
the German word for "The leader"
Not very nice! Note that increasing the limit from 60 to 78
characters will not solve the problem, just reduce how often
it occurs a little.
An even worse case occurs for a message with the following Subject:
Subject: My attempt at German: Öffentlig ist über und Non-Öffentlig is unter
Will in a reply be shown to the human recipient as
Subject: Re: My attempt at German: Öffentlig ist
über und
Non-Öffentlig is unter
The encoded format of the above is:
Subject: Re: My attempt at German: =?iso-8859-1?Q?=D6ffentlig?= is
=?iso-8859-1?Q?=FCber?= und
=?iso-8859-1?Q?Non=2D=D6ffentlig?= is unter
Question 1: Is this correct behaviour of a standards-compliant
mailer?
Question 2: If the answer to question 1 is "yes", then is there
something wrong with the standards? Should the next edition
of MIME include something for headers similar to "=" at the
end of the line in Quoted-Printable lines to indicate soft
linebreaks?
A detail: Note that my mailer converts the doubleblip character
to =22 in a word containing a non-ascii character, but not in
a word not containing a non-ascii character. This is a detail
and not related to the main problem. And note that the problem
gets worsended because every word with non-ascii characters
will get its own =?iso-8859-1?Q?, is this needed, would it
not be enough to just have one =?iso-8859-1?Q? at the beginning
of the subject and one ?= at the end? Not that this would
solve the main problem presented by this e-mail, just a note.
--
Jacob Palme <jpalme(_at_)dsv(_dot_)su(_dot_)se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/jpalme/