Re: Reporting spamming/protecting against spamming on USENET

["Ted Zlatanov" <tzz@lifelogs.com>]

On Tue, 14 Dec 2004, sb@dod.no wrote:

> Ted!

Hi!

> Can your package report spam to anyone on USENET?  Ie. is there a
> cancelling agency somewhere that takes tips?

The nature of Usenet is that it's decentralized and works by common
agreement.  Furthermore, cancelling any but your own articles is
considered more rude than spam itself, and you can't edit an article
without cancelling and reissuing it, so the answer basically is no.

Centralized places such as Gmane can be more aggressive.

> I'm trying to read gnu.emacs.help, and it would feel more satisfying
> if my `M-d's were to help others (like the gmane ones eventually
> will (again)).
> 
> Does the spam package work for USENET?  Or is it email only?

Reporting is done through spam-report.el.  If you can write a suitable
spam-report function, then people will be able to use it.  You could
submit the message to a spam clearing house, for example, and then
people can use that clearing house.  You could also submit the
originating server's IP to a DNSBL service and other people could use
that DNSBL service.  There is a newsgroup method (NOCeM IIRC) to do
this in a vaguely centralized fashion, but in my experience it's slow
and ineffective because Usenet is simply too large and there's too
much spam.

The spam.el package works for any Gnus group with the spam-autodetect
parameter (note it will work for nnrss, nntodo, whatever as well).
You can tell spam.el, by using this parameter as the manual sample
shows, to check all unseen articles in a group using the autodetect
methods.  Those articles found to be spam will be marked so.  That's
all spam-autodetect does.

When you exit the group, the usual spam.el backend processing is done
(this has nothing to do with spam-autodetect).  Here you can report
spam as per spam-report.el, or you can simply learn it for your own
bogofilter/whatever filters.  I use Bogofilter on the gnu.emacs.help
newsgroups and on my regular mail, and I train it with both newsgroup
spam and regular spam.  It works very well.  I should update the
manual example to my current setup one of these days, but it's really
not all that complicated and I keep tuning it.

Add spam-use-bogofilter to the spam-autodetect-methods from the
example and to the spam-process parameters for a newsgroup to get my
current setup in a nutshell.  FWIW, BBDB speeds up the spam check by
bypassing people I know and the blacklist is very ineffective.

> The manual has this from your sample file:
>      ;; all NNTP groups
>      ;; autodetect spam with the blacklist and ham with the BBDB
>      ((spam-autodetect-methods spam-use-blacklist spam-use-BBDB)
>      ;; send all spam to the training group
>       (spam-process-destination . "nnimap+mail.lifelogs.com:train"))
> 
>      ;; only some NNTP groups, where I want to autodetect spam
>      ((spam-autodetect . t))
> 
> But I don't understand how this releates to USENET or the NNTP
> backend...?  Unless `spam-autodetect-methods' is something that's NNTP
> only?  The manual entries for `spam-autodetect-methods',
> `spam-use-blacklist', and `spam-autodetect' doesn't seem to indicate
> that it is...?

It works in any Gnus group, and acts only on unseen articles when you
enter the group.  It's intended for read-only backends where you don't
have splitting, but nothing prevents its use in any other backends.

Ted