Fedora List content, guidelines and antispam

[Paul Howarth]

On Wed, 2005-05-25 at 09:39 +0800, John Summerfied wrote:
> I've been perusing my mail logs and I see mail, some from fedora-list, 
> being declined for these reasons:
> 1. Encoded bodies
> 2. Non-roman character sets
> 3. being called Peter Whalley
> 
> 1. As most experienced list members prefer plain text, I don't 
> understand the need for base64-encoding of bodies. The only purpose it 
> serves that I know about is to attempt to subvert mail filters. I figure 
> if you don't want me to filter on content, I don't want your email.

Fair enough. Base64 encoded plain text goes straight to the bitbucket on
my system too, except for mailing list messages, which are all
whitelisted.

> 2. Non-roman character sets implies the mail's not written in English or 
> any of the other European languages. If anyone writes to me it better be 
> in English because my French hasn't seen much practice since I completed 
> school 40 years ago, and I know no other languages, and English is in 
> any event the standard language for this list. Bouncing mail using 
> non-roman character sets means I get to not see lots of Chinese, 
> Japanese and Korean spam.

It also means you'll get not to see lots of mail in English from people
whose first language is something that needs a different encoding hence
their normal mail setting is for other character sets. They don't change
character sets to send messages in English because it's a hassle to do
so and there's no need, since the characters needed to communicate in
English are also present in their default character set.

> 3. I think we discussed this enough some time ago. The filter dropped 
> quite a deal of email.

I never saw a Peter Whalley bounce because my mailserver rejects
uol.com.br mails on the basis of that domain having no working
postmaster address.

> I'd like the folk who're compiling the guidelines to add the first two 
> points, and ask the list admin to enforce it. Along with any other good 
> ideas these suggestions trigger.

I'd hope that the guidelines are well peer-reviewed so that problematic
suggestions such as (2) can be weeded out before they're cast in stone.
Whatever anti-spam measures Red Hat have in place for this list already
seem to work *very well*, given the almost entire lack of spam on this
list (the "computer for sale" message earlier today was one of the very
very few that got through, and that wasn't a classic mail-to-all-and-
sundry spam either), and I don't see any urgent need to change that.

Paul.
-- 
Paul Howarth <paul at city-fan.org>