1 |
* Alan <alan@×××××.org> [021108 21:24]: |
2 |
|
3 |
>Just my $0.02 about bogofilter.... |
4 |
|
5 |
|
6 |
>The two limitations of bogofilter that need to be addressed someone are: |
7 |
> 1 can only "interact" with it through mutt or some custom hacking |
8 |
> through other programs (I think that emacs vm can notify |
9 |
> spam/non-spam with bogofilter if you know the right spell) |
10 |
> 2 needs enough spam and non-spam to properly set up the word lists. |
11 |
|
12 |
>However, if 2 is done right, then 1 isn't a real problem. |
13 |
|
14 |
>Anyway, that's my $0.02/testimonial :) Spread the love! |
15 |
|
16 |
You're right on your observations, but as you have already seen |
17 |
yourself, the power which comes with Bayesian filters has its drawback, |
18 |
that being that you need to train them. The more spam and good mail you |
19 |
feed to the filter, the better it'll work. |
20 |
That also means that each person will have its own words database suited |
21 |
to the kind of spam and good mail each one receives. |
22 |
As to the need for mutt, it really makes things nicer, but I0m confident |
23 |
that other MUA might use it somehow too. I use Mutt, though, so I cannot |
24 |
help with other apps. |
25 |
|
26 |
|
27 |
-- |
28 |
Javier Marcet <jmarcet@×××××.com> |