Re: [Exim] [exiscanusers] [other stuff] [we don't really wan…

Page principale
Supprimer ce message
Répondre à ce message
Auteur: David Woodhouse
Date:  
À: Nico Erfurth
CC: Tom Kistner, exiscanusers, Exim Users
Sujet: Re: [Exim] [exiscanusers] [other stuff] [we don't really want to see the original subject line anyway, do we?] [yes we do :)] exilist 0.2 - exim based listmanager
On Thu, 2003-03-13 at 16:12, Nico Erfurth wrote:
> I like subject-tags, they make filtering easy in simple mailclients, and
> they allow me to easily see a private reply to a listmail in my inbox.


So if I hadn't Cc'd the list -- if this was a private reply... would it
really have landed in your inbox? You're receiving two copies of this
mail. Does one end up in the list folder and one in your inbox?

If I know you're on vacation and likely to read the list with the 'D'
key when you get back, and I bounce a message to you or explicitly Cc
you in a reply to make sure you see it.... do you get the mail in your
inbox as I intended, or does your filter rule lose it in the mailing
list folder due to a false positive?

Filtering shouldn't be done in mailclients. Filtering should be done in
mail _delivery_ agents. And if you actually want it to be reliable, it
should be done on the SMTP reverse-path alone; since no other method
(that I know of) does not have false positives.

I'm aware that there are some people out there who are forced to use
broken setups where filtering cannot be done at the server, and where
they can't filter on sensible things even in the client.

But if people have such broken systems and the only thing they can
filter on is something unreliable, then they may as well be filtering
with the equally-unreliable method of looking at To: and Cc: headers,
which at least doesn't obscure the Subject: line for everyone else.

I wouldn't argue that a mailing list should pander to _those_ people at
the expense of obscuring the _real_ information in the Subject line for
everyone else. And I certainly wouldn't argue that any mailing list
software should enable such bogosity by _default_ for new lists.

--
dwmw2