Re: Which spam filter do you use?

2007-10-05 Thread M. Fioretti
On Fri, Oct 05, 2007 00:23:06 AM +0200, Eyolf Østrem
([EMAIL PROTECTED]) wrote:

 So, should I switch? I'm quite happy with bogo, especially with the
 current setup with some macros I borrowed from an article in linux
 journal (I think it was), but I would very much like to hear what
 your experiences are in this respect.

Besides some MTA-level filtering, I am using bogo instead of
spamassassin because of one simple reason: lack of maintenance.

I've read several times that SA rules must be constantly updated and
added, otherwise they get half-useless every few weeks. Is this still
true?

Bogofilter, in any case, _does_ let a few spam go through every
day. But once you have installed it and integrated it with Mutt,
procmail, whatever, it just works by itself. Upgrading SA can be
automated, of course, but is one more thing that must work fine all
the time, so I still prefer the other approach: 2/3 minutes every day
to retrain bogofilter on missed spam (via mutt macros and scripts, of
course) versus keeping a constant eye on what's happening in the SA
world.

Of course, if one has to set things up for many people is an
entirely different story.

One thing I plan to do some day, when I have time, is to try something
I've read online: install _another_ bayesian filter and run it from
procmail _only_ on those messages which bogofilter classified as
almost spam. This would minimize resource consumption, as the
lighter filter is enough to trap almost everything, but increase
accuracy while still being zero-maintenance or so.

HTH,
Marco
-- 
The one book on software and digital technologies that no parent or
teacher can ignore:  http://digifreedom.net/node/84


Re: Which spam filter do you use?

2007-10-05 Thread Wilkinson, Alex
0n Fri, Oct 05, 2007 at 08:06:55AM +0200, M. Fioretti wrote: 

On Fri, Oct 05, 2007 00:23:06 AM +0200, Eyolf Østrem
([EMAIL PROTECTED]) wrote:

 So, should I switch? I'm quite happy with bogo, especially with the
 current setup with some macros I borrowed from an article in linux
 journal (I think it was), but I would very much like to hear what
 your experiences are in this respect.

Besides some MTA-level filtering, I am using bogo instead of
spamassassin because of one simple reason: lack of maintenance.

Have you got any documents describing the process of integrating bogofilter 
with mutt ?

 -aW

IMPORTANT: This email remains the property of the Australian Defence 
Organisation and is subject to the jurisdiction of section 70 of the CRIMES ACT 
1914.  If you have received this email in error, you are requested to contact 
the sender and delete the email.



Re: Which spam filter do you use?

2007-10-05 Thread Kyle Wheeler
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Friday, October  5 at 08:06 AM, quoth M. Fioretti:
Besides some MTA-level filtering, I am using bogo instead of
spamassassin because of one simple reason: lack of maintenance.

I've read several times that SA rules must be constantly updated and
added, otherwise they get half-useless every few weeks. Is this still
true?

Bah; SpamAssassin is the swiss-army-knife of spam filters. It includes 
a bayesian filter (not to mention things like razor and dcc, which are 
constantly up-to-date), and as such, does not require updating the 
rules. Updating the rules can *help*, but it is not required to 
continue functioning at a reasonably high level.

The real criticism I'd level against SpamAssassin as compared to 
bogofilter is that SpamAssassin's bayesian classifier is relatively 
simple. To my knowledge, it doesn't tokenize word-pairs and phrases, 
but just single words; thus, something that uses more advanced 
bayesian techniques (I presume bogofilter fits this description?) may 
well beat it at that particular game---which is where updating the 
rules can help as a compensating factor. It's not like a virus-scanner 
where an out-of-date database is worthless.

~Kyle
- -- 
Just because you do not take an interest in politics doesn't mean 
politics won't take an interest in you.
   -- Pericles (430 BC)
-BEGIN PGP SIGNATURE-
Comment: Thank you for using encryption!

iD8DBQFHBeLqBkIOoMqOI14RAquVAJ9sG8NsvhJdHB1qCJdab6Xh2/fczwCggME0
FGk2+Rbq8wVY/Rleab56KhI=
=kjny
-END PGP SIGNATURE-


Re: Which spam filter do you use?

2007-10-05 Thread M. Fioretti
On Fri, Oct 05, 2007 07:56:44 AM +0200, Christian Kuka ([EMAIL PROTECTED]) 
wrote:
 
 At the moment I'm using bogofilter, razor, pyzor, dcc, spamassassin 
 and clamassassin and the following procmail rules:

This (cascading several filters, lightest to heaviest) is what inmy
other reply I said I'd like to try someday. Do you have any _measure_
of how better it is wrt a bogofilter-only setup?

Something like bogo alone stops 80% of spam, razor 80% of bogo
missed, SA 80% of what is still left etc.

Thanks,
Marco
-- 
Your own civil rights and the quality of your life heavily depend on
how software is used *around* you:http://digifreedom.net/node/84


Re: Which spam filter do you use?

2007-10-05 Thread M. Fioretti
On Fri, Oct 05, 2007 14:25:43 PM +0800, Wilkinson, Alex wrote:
 Have you got any documents describing the process of integrating
 bogofilter with mutt ?

I just used these:
http://www.linuxjournal.com/article/6439
http://www.linuxjournal.com/article/7436

Another approach, valid also on Imap and compatible with webmail, is
to just move all the spam missed by bogofilter in a predefined with a
macro, and then have a cron shell script which runs the same
bogofilter command of those mutt macros on each file in that
directory and then remove the files.

HTH,
Marco
-- 
The Family Guide to Digital Freedom:  http://digifreedom.net/node/99


Re: Which spam filter do you use?

2007-10-05 Thread Michael Tatge
* On Fri, Oct 05, 2007 M. Fioretti ([EMAIL PROTECTED]) muttered:
 On Fri, Oct 05, 2007 14:25:43 PM +0800, Wilkinson, Alex wrote:
 Another approach, valid also on Imap and compatible with webmail, is
 to just move all the spam missed by bogofilter in a predefined with a
 macro,

When imap is involved sieve really is the way to go.
e.g.

# spam level 5 * or more
if header :contains X-Spam-Level * {
   fileinto INBOX/spam;
   stop;
}



HTH,

Michael
-- 
Excusing bad programming is a shooting offence, no matter _what_ the
circumstances.
-- Linus Torvalds, to the linux-kernel list

PGP-Key-ID: 0xDC1A44DD
Jabber: [EMAIL PROTECTED]


Re: Which spam filter do you use?

2007-10-05 Thread Eyolf Østrem
On 05.10.2007 (02:08), Kyle Wheeler wrote:
 On Friday, October  5 at 08:06 AM, quoth M. Fioretti:
 Besides some MTA-level filtering, I am using bogo instead of
 spamassassin because of one simple reason: lack of maintenance.
 
 I've read several times that SA rules must be constantly updated and
 added, otherwise they get half-useless every few weeks. Is this still
 true?
 
 Bah; SpamAssassin is the swiss-army-knife of spam filters. It includes 
 a bayesian filter (not to mention things like razor and dcc, which are 
 constantly up-to-date), and as such, does not require updating the 
 rules. Updating the rules can *help*, but it is not required to 
 continue functioning at a reasonably high level.
 
 The real criticism I'd level against SpamAssassin as compared to 
 bogofilter is that SpamAssassin's bayesian classifier is relatively 
 simple. To my knowledge, it doesn't tokenize word-pairs and phrases, 
 but just single words; thus, something that uses more advanced 
 bayesian techniques (I presume bogofilter fits this description?) may 
 well beat it at that particular game---which is where updating the 
 rules can help as a compensating factor. It's not like a virus-scanner 
 where an out-of-date database is worthless.

So this does in fact mean that without that extra time tending the
rules, SA may actually let more spam through? 

It's not that I'm dissatisfied with my current situation. My #1
concern is actually with false negatives; I've since long given up
browsing through the CapturedSpam folder to check for them before I
delete everything. I haven't had any complaints from people who think
I neglect them (unless the complaints also end up in the filter...),
but in one comparison I read, it seemed that SA had absolutely 0 of
that, whereas bogo might have one or two (out of I don't remember how
many thousand).

All in all, it may not be such a bad thing; if there IS some mail I
haven't answered but should have, I can always say sorry, never got
your mail, it must have been trapped in the filter. I'd hate to lose
that excuse... :-)

eyolf

-- 
And 1.1.81 is officially BugFree(tm), so if you receive any bug-reports
on it, you know they are just evil lies.
(By Linus Torvalds, [EMAIL PROTECTED])


Re: Which spam filter do you use?

2007-10-05 Thread Michal Vitecek
 hi,

Eyolf ?strem wrote:
I've been using bogofilter ever since I first installed KDE/Kmail and the
potentially hassle-free configuration of SpamAssassin led to constant
crashes. I belive it has been solved by now, and in any case Kmail is
ancient history, but I was wondering what experiences the list people
have with various filters.

 we're using qsf (http://www.ivarch.com/programs/qsf/) which i personaly
 find perfect. moreover it's fast so we can use it on our mail server
 without any problems (tried spamassassin, spambayes with bad
 excerience).

-- 
fuf ([EMAIL PROTECTED])


Re: Which spam filter do you use?

2007-10-05 Thread Holger Weiss
* Eyolf Østrem [EMAIL PROTECTED] [2007-10-05 13:22]:
 On 05.10.2007 (02:08), Kyle Wheeler wrote:
  The real criticism I'd level against SpamAssassin as compared to
  bogofilter is that SpamAssassin's bayesian classifier is relatively
  simple.

Yes.  Also, SpamAssassin is terribly slow.

  To my knowledge, it doesn't tokenize word-pairs and phrases,
  but just single words; thus, something that uses more advanced
  bayesian techniques (I presume bogofilter fits this description?) may
  well beat it at that particular game---which is where updating the
  rules can help as a compensating factor. It's not like a virus-scanner
  where an out-of-date database is worthless.

 So this does in fact mean that without that extra time tending the
 rules, SA may actually let more spam through? 

Yes, even if you do take that time :-)  From my experience, which
includes few benchmarks I did, Bogofilter's accuracy is way better than
SpamAssassin's, even if enabling SpamAssassin's bayesian classifier,
Razor, and a few other non-default modules.

Bogofilter's accuracy highly depends on good training, though.  It's
critical to not only train Bogofilter with misclassified messages but
also with messages it's unsure about.  Fine-tuning the configuration
might also increase Bogofilter's accuracy[*].

IMO, SpamAssassin is only useful if you don't want or cannot train your
spam filter for some reason (e.g., if you're an ISP, though in this case
SpamAssassin's bad performance can be a real drawback).

 It's not that I'm dissatisfied with my current situation. My #1
 concern is actually with false negatives; I've since long given up
 browsing through the CapturedSpam folder to check for them before I
 delete everything.

So you mean false positives :-)  In any case, any serious spam filter
allows for adjusting the spam/ham thresholds, so you can always buy
redrucing the number of false positives to almost zero at the cost of
increasing the number of false negatives.  SpamAssassin's default
configuration does just that (which makes sense, of course).

 it seemed that SA had absolutely 0 of that, whereas bogo might have
 one or two (out of I don't remember how many thousand).

Depends on the configuration.

Holger

[*] You could try bogotune(1) and/or increasing multi-token-count,
though this increases the database size and decreases Bogofilter's
performance.  It _should_ increase Bogofilter's accuracy, though for
me it had far less effect than described in the following posting
(mainly because I get better results with multi-token-count=1 than
described in the posting):

http://www.bogofilter.org/pipermail/bogofilter-dev/2006-August/003357.html


Re: Which spam filter do you use?

2007-10-05 Thread Holger Weiss
* M. Fioretti [EMAIL PROTECTED] [2007-10-05 08:19]:
 On Fri, Oct 05, 2007 07:56:44 AM +0200, Christian Kuka ([EMAIL PROTECTED]) 
 wrote:
  At the moment I'm using bogofilter, razor, pyzor, dcc, spamassassin 
  and clamassassin and the following procmail rules:

 This (cascading several filters, lightest to heaviest) is what inmy
 other reply I said I'd like to try someday. Do you have any _measure_
 of how better it is wrt a bogofilter-only setup?

 Something like bogo alone stops 80% of spam, razor 80% of bogo
 missed, SA 80% of what is still left etc.

I did that myself for quite some time, but note that while this of
course reduces the number of false negatives, it also increases the
number of false positives, so you won't necessarily increase the global
accuracy with such a setup.  Giving useful _measures_ is really hard
though as they will heavily depend on the configuration/training of the
filters and of course also on the actual mail corpus they operate on.

Holger


Re: Which spam filter do you use?

2007-10-05 Thread M. Fioretti
On Fri, Oct 05, 2007 02:08:26 AM -0500, Kyle Wheeler
([EMAIL PROTECTED]) wrote:

 Bah; SpamAssassin is the swiss-army-knife of spam filters. It includes 
 a bayesian filter (not to mention things like razor and dcc, which are 
 constantly up-to-date), and as such, does not require updating the 
 rules.

But razor and dcc consume bandwidth, so to figure out if SA is
worthwhile maybe one should ask:

0) Do I have a flat rate fast connection, where I wouldn't notice SA
   contactly doing network checks?

1) after the whole message has been downloaded anyway, does SA block
   a LOT more spam than bogofilter or qsf? if the answer is yes,
   (and that is quite a big if, judging from both online literature
   and other answers in this thread...) is the difference big enough
   to justify the extra CPU and/or bandwidth consumption, plus keeping
   the rules updated?

I mean sure, SA has tons of extra non-bayesian tricks to catch spam,
but if the bayesian algorithms in bogofilter or qsf catch almost all
of it anyway without those extra tricks, bandwidth, cpu cycles and
manual maintenance... do I need to bother (this *is* a serious
question, I'm really trying to figure out if the _possibility_ to go
from just 1/2 spam messages a day in my inbox to 0 is worth the extra
effort)??

Of course, the answer depend on one's needs, how much mail he or she
receives and much other stuff. And if one has full control of the MTA,
where lots of spam can and should be recognized and blocked before
ever starting SA or any other content filter.

Ciao,   
Marco
-- 
Help your relatives, friends and partners love Free Standards and Free
Software!   http://digifreedom.net/node/84


Re: Which spam filter do you use?

2007-10-05 Thread M. Fioretti
On Fri, Oct 05, 2007 14:48:09 PM +0200, Michal Vitecek ([EMAIL PROTECTED]) 
wrote:
  hi,
 
 Eyolf ?strem wrote:
 I've been using bogofilter ever since I first installed KDE/Kmail and the
 potentially hassle-free configuration of SpamAssassin led to constant
 crashes. I belive it has been solved by now, and in any case Kmail is
 ancient history, but I was wondering what experiences the list people
 have with various filters.
 
  we're using qsf (http://www.ivarch.com/programs/qsf/) which i personaly

Now I've remembered! The page I mentioned earlier in the thread, where
I had read of cascading different bayesian filters was centered just
around bogofilter and qsf, and is this one:

http://www.acme.com/mail_filtering/bayesian_frameset.html

where the author also justifies why he put the filters in that
particular order. It is an interesting read, and what I plan to try.

HTH
Marco
-- 
Help *everybody* love Free Standards and Free Software!
http://digifreedom.net


Re: Which spam filter do you use?

2007-10-05 Thread Eyolf Østrem
On 05.10.2007 (15:30), Holger Weiss wrote:
 * eyolf østrem [EMAIL PROTECTED] [2007-10-05 13:22]:
 
  It's not that I'm dissatisfied with my current situation. My #1
  concern is actually with false negatives; I've since long given up
  browsing through the CapturedSpam folder to check for them before I
  delete everything.
 
 So you mean false positives :-) 

Uhm... yes... positive, negative, who's to know what's what in the
long run?  

Anyway, thanks for all the feedback (I sortof knew when I sent it off
that I'm starting a long thread here). I think I'll start by
training mr bogo and ask him to be careful with those false
whatever.

e


-- 
The seven eyes of Ningauble the Wizard floated back to his hood as he
reported to Fafhrd: I have seen much, yet cannot explain all.  The Gray
Mouser is exactly twenty-five feet below the deepest cellar in the palace
of Gilpkerio Kistomerces.  Even though twenty-four parts in twenty-five of
him are dead, he is alive.
Now about Lankhmar.  She's been invaded, her walls breached
everywhere and desperate fighting is going on in the streets, by a fierce
host which out-numbers Lankhamar's inhabitants by fifty to one -- and
equipped with all modern weapons.  Yet you can save the city.
How? demanded Fafhrd.
Ningauble shrugged.  You're a hero.  You should know.
-- Fritz Leiber, The Swords of Lankhmar


Re: Which spam filter do you use?

2007-10-05 Thread Holger Weiss
* Christian Kuka [EMAIL PROTECTED] [2007-10-05 07:56]:
 I also read from a scanner called crm114 in the linux magazine that
 should be realy good, but never checked that.

I use OSBF-Lua[1] which is a port of CRM114.  It's a Bayesian classifier
which uses a more sophisticated algorithm than Bogofilter and others,
and which indeed performs even better than Bogofilter for me, regarding
both classification accuracy and speed.  It gives me a global accuracy
of about 99.9% (about 1 out of 1000 e-mails are misclassified).  Also,
it learns a lot quicker than Bogofilter, usually there's no need to do
any pre-training[2].  And it has far less configuration knobs to play
with, which is also good :-)  It just works out-of-the-box.

I put the Mutt (1.5.x) macros I use for training online:

ftp://ftp.in-berlin.de/pub/users/weiss/spamfilter/README.osbf4mutt
ftp://ftp.in-berlin.de/pub/users/weiss/spamfilter/osbf4mutt-0.2.tar.gz

Holger

[1] http://osbf-lua.luaforge.net/
[2] http://page.mi.fu-berlin.de/~siefkes/software/trainfilter/#section_1_4


Re: Which spam filter do you use?

2007-10-05 Thread Kyle Wheeler
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Friday, October  5 at 01:22 PM, quoth Eyolf Østrem:
 So this does in fact mean that without that extra time tending the 
 rules, SA may actually let more spam through? 

As in most cases, the answer is it depends. It *may* let more spam 
through. In my case, I use SA for everything, and I train it on 
everything I can lay hands on---and have for years. I didn't notice a 
difference in performance (for my personal mail) when I added 
sa-update to the crontab. On the other hand, not everyone on my server 
uses the training folders, and so their spamassassin configurations 
don't adapt as quickly to new spam trends. That is where, on my 
system, updating the rules files really helps with.

But everyone has different experiences.

I will say, though, it is true that SpamAssassin is really slow, but 
not for the reason many people claim. We have a pretty standard 
spamc/spamd setup, and if I turn off all the network tests, it flies. 
But, because I find the extra tests SA performs valuable, I have it 
crunch through all the various DNS, DKIM, and remote-spam-database 
tests, which adds significantly to its average processing time. 
According to my logs, it takes anywhere from 4 to 16 seconds to 
process each message, which is pretty hefty! Then again, it's 
processing mail at delivery time so most folks don't notice, and it's 
doing it all in parallel... but because of those network tests, it's 
unlikely that I could improve performance significantly by putting it 
on a faster machine. If my email server was under heavier load, I'd 
probably have to look into changing my filtering setup (put SA on a 
different machine, boost my dns cache size, maybe add a prefilter of 
some kind, etc.).

As for impact, it chunks through on the order of 5000 emails a day, of 
which over 86% is spam (according to my logs), and I think the last 
time I got a message that was incorrectly classified was probably... 
maybe a week ago, or so.

Some of my users have better experiences, some worse. Usually, if one  
of my users complains that they're getting too much spam, it's because 
they haven't been using the training features.

I think the real question you need to ask is: does my current spam 
system work sufficiently well for my taste, and what am I willing to 
pay to get better accuracy?

~Kyle
- -- 
To believe in God is impossible---to not believe in him is absurd.
-- Voltaire
-BEGIN PGP SIGNATURE-
Comment: Thank you for using encryption!

iD8DBQFHBkxSBkIOoMqOI14RAqP4AKCDDa6efM5yi54+HFsZHCxn1atlcQCfcHui
8QCYRTaN6F3hnUPsXpYAhj0=
=211r
-END PGP SIGNATURE-


Re: Which spam filter do you use?

2007-10-05 Thread Kyle Wheeler
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Friday, October  5 at 04:02 PM, quoth M. Fioretti:
 0) Do I have a flat rate fast connection, where I wouldn't notice SA
   contactly doing network checks?

Indeed! Not all solutions are perfect for all situations.

 1) after the whole message has been downloaded anyway, does SA block
   a LOT more spam than bogofilter or qsf? if the answer is yes,
   (and that is quite a big if, judging from both online literature
   and other answers in this thread...) is the difference big enough
   to justify the extra CPU and/or bandwidth consumption, plus keeping
   the rules updated?

Unfortunately, that's a tough one to answer. How much CPU is a spam 
worth? Does the answer change depending on whether you get 5 spams a 
day, or 5000? There's no right answer to that one, it's all personal 
preference.

 And if one has full control of the MTA, where lots of spam can and 
 should be recognized and blocked before ever starting SA or any 
 other content filter.

That depends on what you're willing to put up with. For example, many 
of my users have learned to distrust things like DNS blacklists, and 
by extension any completely blocking spam mechanism. ANY anti-spam 
technique will have false-positives, and sometimes users aren't 
willing to put up with that, which changes the requirements of your 
antispam solution.

~Kyle
- -- 
Every gun that is made, every warship launched, every rocket fired 
signifies, in the final sense, a theft from those who hunger and are 
not fed, those who are cold and are not clothed.
-- Dwight D. Eisenhower
-BEGIN PGP SIGNATURE-
Comment: Thank you for using encryption!

iD8DBQFHBk5+BkIOoMqOI14RAsN2AKDJS8mF2Fi6dMZqzB1UVJPsUdb6nwCguz69
6DlFQihv3TkWUgUxX+oJKSw=
=8blO
-END PGP SIGNATURE-


Re: Which spam filter do you use?

2007-10-05 Thread Holger Weiss
* Kyle Wheeler [EMAIL PROTECTED] [2007-10-05 09:38]:
 I will say, though, it is true that SpamAssassin is really slow, but
 not for the reason many people claim.

My claim is that spamd requires an order of magnitude more CPU and
memory ressources than Bogofilter and others do.  Of course, that's
usually not an issue on single-user systems.

 We have a pretty standard spamc/spamd setup, and if I turn off all the
 network tests, it flies.

At my workplace, for about 40.000 users, we need six dedicated servers
for spamd and one which does Bogofilter and other stuff.

 As for impact, it chunks through on the order of 5000 emails a day, of 
 which over 86% is spam (according to my logs), and I think the last 
 time I got a message that was incorrectly classified was probably... 
 maybe a week ago, or so.

So you get about 1 out of 35.000 messages misclassified (which would be
an accuracy of about 99.997%)?  I cannot quite believe that :-)  In any
case, it does sound as if you do get significantly better results out of
SpamAssassin than I ever did, despite a lot of tweaking I tried.

 I think the real question you need to ask is: does my current spam
 system work sufficiently well for my taste, and what am I willing to
 pay to get better accuracy?

Yes.

Holger


Re: Which spam filter do you use?

2007-10-05 Thread Kyle Wheeler
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Friday, October  5 at 05:14 PM, quoth Holger Weiss:
 As for impact, it chunks through on the order of 5000 emails a day, 
 of which over 86% is spam (according to my logs), and I think the 
 last time I got a message that was incorrectly classified was 
 probably... maybe a week ago, or so.

 So you get about 1 out of 35.000 messages misclassified (which would be 
 an accuracy of about 99.997%)?  I cannot quite believe that :-)  In any 
 case, it does sound as if you do get significantly better results out of 
 SpamAssassin than I ever did, despite a lot of tweaking I tried.

Well, the twist to that statistic is that not all 5000 of those emails 
are for *me*. I'm only one of the users on my system that uses SA 
(though I'm certainly the biggest user of email on the system). Based 
on yesterday's logs, I personally get around 10,521 emails (including 
spam) a week. Assuming an oversized fudge factor of maybe ten 
misclassified messages that I either didn't notice or don't remember, 
that puts my accuracy around 99.9%.

I think this really does, though, boil down to: different filters do 
better for different people.

~Kyle
- -- 
I beseech you, in the bowels of Christ, think it possible you may be 
mistaken.
 -- Oliver Cromwell
-BEGIN PGP SIGNATURE-
Comment: Thank you for using encryption!

iD8DBQFHBly8BkIOoMqOI14RAsAUAJ9Ub6P2lQdUHXWK0aTq4EeaasKeSwCgxFfz
y7saTUN2IJvuttxUNpnPWE8=
=4sAa
-END PGP SIGNATURE-


Re: Which spam filter do you use?

2007-10-05 Thread M. Fioretti
On Fri, Oct 05, 2007 09:47:26 AM -0500, Kyle Wheeler
([EMAIL PROTECTED]) wrote:

 On Friday, October  5 at 04:02 PM, quoth M. Fioretti:
  0) Do I have a flat rate fast connection, where I wouldn't notice SA
contactly doing network checks?
 
 Indeed! Not all solutions are perfect for all situations.

Yes, that's the same thing I had said, which it applies to all the pieces of
the puzzle, including this:

is the difference big enough
to justify the extra CPU and/or bandwidth consumption, plus keeping
the rules updated?
 
  And if one has full control of the MTA, where lots of spam can and 
  should be recognized and blocked before ever starting SA or any 
  other content filter.
 
 That depends on what you're willing to put up with.

Of course. The MTA can block many surely spammish messages (those
pretending to come from inside your network, for example). At the same
time, DNS blacklists as a completely blocking mechanism make sense
only _if_ their maintainer is inhumanly perfect. Otherwise, it comes
too often too close to censorship (when who decides what you will not
receive is somebody ELSE, of course: any individual filtering
exclusively his or her own mail must remain free to shoot himself in
the feet).

Ciao,
Marco
-- 
Your own civil rights and the quality of your life heavily depends on
how software is used *around* you: http://digifreedom.net/node/84


Which spam filter do you use?

2007-10-04 Thread Eyolf Østrem
I've been using bogofilter ever since I first installed KDE/Kmail and the
potentially hassle-free configuration of SpamAssassin led to constant
crashes. I belive it has been solved by now, and in any case Kmail is
ancient history, but I was wondering what experiences the list people
have with various filters.

From what I've read, bogo is quicker than the other contenders, but
lets more spam through. While the speed was a concern in Kmail, since
the filtering was done in the app itself, which meant that it was
unresponsive for a while while the filtering was going on, that is not
so much of a concern now, when that is taken care of by procmail. That
leaves me with c. 10-15 spam mails a day that slip through (out of c.
150-200). 

So, should I switch? I'm quite happy with bogo, especially with the
current setup with some macros I borrowed from an article in linux
journal (I think it was), but I would very much like to hear what your
experiences are in this respect.

Eyolf

-- 
Unceasing warfare gives rise to its own social conditions which have been 
similar in all epochs. People enter a permanent state of alertness to ward
off attacks. You seethe absolute rule of the autocrat. All new things become 
dangerous frontier districts-new planets, new economic areas to exploit, new 
ideas or new devices, visitors-everything suspect. Feudalism takes firm hold, 
sometimes disguised as a politbureau or similar structure, but always present. 
Hereditary succession follows the lines of power. The blood of the powerful 
dominates. The vice regents of heaven or their equivalent apportion the wealth. 
And their know they must control inheritance or slowly let the power melt away. 
Now, do you understand Leto's Peace?

  -- The Stolen Journals


Re: Which spam filter do you use?

2007-10-04 Thread Kyle Wheeler
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Friday, October  5 at 12:23 AM, quoth Eyolf Østrem:
So, should I switch? I'm quite happy with bogo

If you're satisfied with your current spam prevention technique, 
there's absolutely no reason to switch. At best, you'll get fewer 
spams. If the spam level you currently receive is acceptable: count 
your blessings.

For what it's worth, all the domains I administer have used 
spamassassin for several years, and the accuracy is simply stunning.

~Kyle
- -- 
Of course it's the same old story. Truth usually is the same old 
story.
   -- Margaret Thatcher
-BEGIN PGP SIGNATURE-
Comment: Thank you for using encryption!

iD8DBQFHBZHMBkIOoMqOI14RAtUbAJ91DOTtzyh1sI8hagsTxF0Y5Y0eawCgwIrI
yAw6/xg2FPtkzaN8yqiijxs=
=nnJ7
-END PGP SIGNATURE-


Re: Which spam filter do you use?

2007-10-04 Thread Christian Kuka
Hi,

Thus wrote Eyolf Østrem ([EMAIL PROTECTED]) [07.10.05 06:54]:
 I've been using bogofilter ever since I first installed KDE/Kmail and the
 potentially hassle-free configuration of SpamAssassin led to constant
 crashes. I belive it has been solved by now, and in any case Kmail is
 ancient history, but I was wondering what experiences the list people
 have with various filters.

At the moment I'm using bogofilter, razor, pyzor, dcc, spamassassin 
and clamassassin and the following procmail rules:

-snip--

SPAM=$MAILDIR/.Junk/
VIRUS=$MAILDIR/.Virus/

DCCPROC=/usr/bin/dccproc
BOGOFILTER=/usr/bin/bogofilter
PYZOR=/usr/bin/pyzor
RAZOR=/usr/bin/razor-check
SPAMC=/usr/bin/spamc
CLAM=/usr/bin/clamassassin

:0 fw
| $BOGOFILTER -u -e -p

:0 e
{ EXITCODE=75 HOST }

:0:
* ^X-Bogosity:.(Yes|Spam)
$SPAM

# Razor
:0 Wc
| $RAZOR -conf=$HOME/.razor/razor-agent.conf

:0 Wa:
$SPAM

# Pyzor
:0 Wc
| $PYZOR check

:0 Wa:
$SPAM

# DCC
:0 fw 
| $DCCPROC  -ERw whiteclnt -ccmn,10

:0 e:
$SPAM

# Spamassassin
:0 fw: $PMDIR/spamassassin.db
| $SPAMC

:0:
* ^X-Spam-Status: Yes
$SPAM

:0 fw: $PMDIR/clamassassin.db
| $CLAM 

:0:
* ^X-Virus-Status: Yes
$VIRUS

-snip--

At the moment I don't get any spam/junk mail but sometimes some mails
from mailinglists (especially from the debian list) are in the junk folder.
But I also have to say that it realy takes some time till I get a mail.


 From what I've read, bogo is quicker than the other contenders, but
 lets more spam through. While the speed was a concern in Kmail, since
 the filtering was done in the app itself, which meant that it was
 unresponsive for a while while the filtering was going on, that is not
 so much of a concern now, when that is taken care of by procmail. That
 leaves me with c. 10-15 spam mails a day that slip through (out of c.
 150-200). 
 
 So, should I switch? I'm quite happy with bogo, especially with the
 current setup with some macros I borrowed from an article in linux
 journal (I think it was), but I would very much like to hear what your
 experiences are in this respect.

I also read from a scanner called crm114 in the linux magazine that
should be realy good, but never checked that.

 Eyolf
 

Christian
-- 
---
PGP/OpenPGP/GnuPG encrypted mail preferred in all private communication.
Key ID: 0x61E7150B - 4EFC 3FA6 FB8E 2BD5 CA11  6F15 F557 6B5D 61E7 150B

Christian Kuka
[EMAIL PROTECTED] 


signature.asc
Description: Digital signature