Re: HELP!

[Erik Naggum <[e-mail]>]

Bob,

I think we're looking at user misunderstanding.  Forwarding messages like
this is very useful, and any comments that the forwarder might want to
include could go in a Resent-Comments field that nobody would read, or be
prepended to the forwarded message.  That somebody should get the brilliant
idea to use this for regular replies is not something I think we can do
anything about, except educate people that resending is not the same as
replying.

Yes, it's weird, I totally agree, but I think mailers that don't alert you
to Resent-headers are deficient, and more important to fix.  While we're on
deficient mailers: the mailers that do not provide a reply function that
keeps the subject, autogenerates the To and possibly Cc fields, and sets
In-Reply-To to the message-id and perhaps other identifying information,
and ideally includes a References header that actually keeps track of the
thread, is a lot more to worry about than this.  As you say, this has gone
unnoticed for years, while sub-minimal (a/k/a dysfunctional) mailers make
life harder for those of us who get lots of mail and try to link reply to
message sent.

<nostalgia>
I remember the days when the worst mailer we had available at the U of Oslo
was MM (MailManager), a truly great piece of software for TOPS-20.  These
days, the worst is that despicable Unix mailer that can't even handle RFC
822 syntax beyond user@domain, and some not even that.  I recently
installed Emacs 19.17 and discovered that the new rmail has new, improved
code to remove the name phrase in all headers.  And this is happening while
X.400 has made it de rigeur not to distinguish between the recipient and
his mailbox, or name and address.  And MIME has made it harder to send
8-bit messages because of the severely brain-damaged way 8-bit headers are
treated.  I thank my parents for not giving me a name with one of those
funny Norwegian letters in it.
</nostalgia>

The need to educate people who write mailers is mounting, and my very
personal opinion is that X.400 succeeds in getting people to implement it
better than we do because your average street-corner punk can't write an
X.400 mailer, but he can write what he believes to be an RFC 822 mailer.
This is both good and bad, but the amount of misguided creativity that goes
into RFC 822 mailers is truly scary, such as this "resend" feature that
comes with no instructions for proper use.

Is a "mail system requirements" RFC a useful thing to contemplate?  How
about an FYI or User Area RFC to explain this to people?  Perhaps a book
(like from O'Reilly) about Internet mail that explains what was intended
and how it should be used?  With MIME doing what it does to basic mail
service (that is, dividing people into at least two camps), I think this
should be a priority topic.  What do the other mail guys here think?

Best regards,
</Erik>
Sunnyvale, CA
--
Erik Naggum <[e-mail]> <[e-mail]>          ISO  8879 SGML
Chairman, SGML SIGhyper <[e-mail]>       ISO 10744 HyTime
"Memento, terrigena.  Memento, vita brevis."            ISO 10646 UCS