Re: spam

[Eric Siegerman]

On Thu, May 16, 2002 at 06:15:07PM -0500, Skip Montanaro wrote:
> 
>     Dan> IMHO the mailing list should reject messages from nonsubscribers.
>     Dan> I thought it already did.
> 
> 
> I agree, however I think the current stable version of Mailman (2.0.x)
> currently only allows the list admin to hold messages, which then have to be
> reviewed by a human.

It's not a software limitation; it's FSF policy.  See:
        http://www.gnu.org/prep/mailinglists.html#Spam

It would seem that automated filtering is unacceptable to them,
whether content-based, open-relay detection, restricting posts to
subscribers, or whatever.  But moderation is ok.  So is providing
information to help subscribers to filter; they already add
headers like this to messages from open relays:
    X-RBL-Warning: (relays.ordb.org) This mail was handled by an open relay
        - please visit <http://ORDB.org/lookup/?host=211.152.181.131>

I guess their underlying goal is that posts may be rejected, but
only by a human, never by a program.

Info-cvs got really bad, probably a dozen spams a day.  Even
then, the only option the GNU folks would consider was to make it
a moderated list -- RMS himself said so.

This was done a little over a month ago, with a loose policy that
basically just filters out spam from non-subscribers (i.e. only
non-subscribers' posts are held for approval in the first place).
Things have been a *lot* better since this kicked in.

> That said, I know there is some spam control.  bug-automake sends me
> email periodically telling me to approve or reject some suspicious
> email.
>
> I don't have the password, though, so it is just accumulating.

Heh.  The new info-cvs moderator said there were 250-odd messages
in the queue when he took it over, dating back as far as 2000.

--

|  | /\
|-_|/  >   Eric Siegerman, Toronto, Ont.        address@hidden
|  |  /
Anyone who swims with the current will reach the big music steamship;
whoever swims against the current will perhaps reach the source.
        - Paul Schneider-Esleben