🚀 go-pugleaf

> If all else fails, a MIME message can use 'base-64' encoding, which uses
> the charset  A-Z, a-z, 0-9, and  the characters ".", "/",  and "=". This
> set  was  specifically  chosen  because  there  were  (after  exhaustive
> searching  no  known  gateways  that would  not  pass  these  characters
> unmolested. Combining  these with the MIME-defined  'charset=' parameter
> allows passing of any charset  registered with IANA through any gateway,
> as long as the receiving UA  knows how to display the charset requested.
>
> There   is  absolutely   *NOTHING*  in   MIME  that   places  additional
> requirements  on a  mail  gateway  over and  above  what current  RFC822
> requires. If the mail gateway will  pass a legal RFC822 message, it will
> pass  a legal  MIME message.  And in  fact, using  the defined  encoding
> schemes,  it's  possible  to  send MIME  through  gateways  that  RFC822
> cannot.

How do you propose to get everyone on the Internet to use base64
encoding when sending to BITNET users?  Also, base64 encoding is
unreadable to anyone who lacks a decoder, while quoted-printable is
mostly readable, if the proportion of non-ASCII characters is
relatively small.  Base64 decoders are pretty simple, but I suspect
they will be far from univerally available anytime soon.

The difference with MIME, is that data is being passed that is intended
to be processed by some program in order to be properly presented to
the user.  In such a situation, there is no opporunity for the user to
intervene and manually correct characters that did not come through
correctly.

If BITNET is going to be a full participant in the world-wide email
system in the future, it will need to be able to somehow handle
messages that use quoted-printable encoding.

> Now, can we get back to discussing *real* problems, such as the infamous
> line-length problem?
>
>                                   Valdis Kletnieks
>                                   Computer Systems Engineer
>                                   Virginia Polytechnic Institute

I agree that long lines are a real problem, one that I have been
complaining about for years.  I just don't agree that long lines are
the only real problem.  Incompatible EBCDICs are a real problem too.
It is desirable to solve more than one problem at a time, when it is
practical to do so.  However, I seem to be beating my head against a
brick wall here, so I might as well stop and ease the pain.  :-)

Roger Fajman
Computer Scientist
National Institutes of Health

(So there -- my title is as good as your title.  :-)