Re: [Q] why using spam.el ?

[Ted Zlatanov <tzz@lifelogs.com>]

On Wed, 21 Apr 2004, didier@lrde.epita.fr wrote:

> In my own experience, I use spamassassin via procmail (with a very
> simple spam detection scheme in gnus making the default (last)
> folder of nnmail-split-methods to a group named "spam"). After
> bayesian training over 6000 messages (3000 spam, 3000 non spam), my
> spamassassin undetected spam rate fell from 200 per day to 1 or 2
> per day.

If that's OK with you, then that's great.  spam.el supports the setup
you describe, by the way.

Where people find spam.el most useful is not in detecting spam, but
in catching and training on misclassified spam and ham.  I'll bet
that for you, it's not a simple process to do this.  At the very
least you can have misclassified spam and ham moved automatically
when you exit the summary buffer.

Also, spam.el supports many backends, and catches spam in newsgroups.

> Currently, the only thing I use in spam.el is the M-d key sequence,
> which is a bit overkill ;-)

OK, but why do you use it?  The spam-mark is not persistent, so it's
only good until you exit the group.  I'll bet you have something else
happen after you set spam-marks (or you'd like to have something else
happen).

Anyhow, I said the other day that spam.el is a *reactive* package, it
responds to what you need.  You don't say "now, process all spam."
Instead you exit the group summary buffer and things happen
automatically.  I think that works better, and it's less intrusive.

I hope that answers some of your questions.

Ted