Sorry to join this discussion so late.
Since I have to use my machine mostly on w2000 I am looking for a free solution
for mail filtering. Tried annoyance-filter using cygwin environment but there are
times it gets stuck and takes 98% CPU. I liked the proxy stuff though. Any
recommendation?
My bayesian filter in part of Celeste, which is the mail client embedded
in an extremely rich environment called Squeak. This environment
supports development in Smalltalk, which is also what it is written in,
and what Celeste is written in. It's open source, though not Open
Source.
1. Somewhere back on this thread there was a reference to a Bayesian
filter. Can someone supply a functional RTFM? (Not on Bayesian
statistics please!)
2. The same someone (I think) said that it was enabled in Mozilla 1.3a.
How so? RTFM?
STFW http://www.mozilla.org/mailnews/spam.html ...
Daniel Vainsencher wrote:
But really, in summary - who cares about retribution. Just get a
bayesian filter, they work VERY well, they subsume most other related
techniques in there (a white list is formed automatically, for example),
and it doesn't require much thought.
It's not perfect, so make
Hi,
Sorry for delaying my answer. I had tried to look at this in some other
ways, and I simply don't know how further to help you with this. You are
right about -l, and I haven't found time to look deeper into this.
The debug info is printed to stderr, but KMail may be capturing this into
DH M!--lvPVVE--y n!--lvPVVE--am!--lvPVVE--e i!--lvPVVE--s Rimma,BR
DH etc. etc.
DH
DH A dumb question. Do you take any measures against the sender of such a spam
DH message like complaining to authorities ?
No. As I mentioned, I have about 50 (low spam day) to 100 (spammer carnage
day) spams
DV OTOH, what we could do, is that after we confirm it's spam (my
DV MUA simply files it automatically away. The folder is emptied
DV with the trash folder), the mail could be forwarded to some
DV central authority, for further use in making life hard for the
DV originating spammers.
Well, if
Daniel Vainsencher wrote:
Are you kidding? this would -
Of course, the problem is someone would have to BE this
central authority, and we'd have to trust it not to misuse all of our
addresses...
What about spamcop: http://spamcop.net/
Daniel
PS - Not David Harel of Weizmann fame, I
DH Are you kidding? this would -
DH Of course, the problem is someone would have to BE this
DH central authority, and we'd have to trust it not to misuse all of our
DH addresses...
DH
DH What about spamcop: http://spamcop.net/
I once tried to understand how they are filtering the mail and
Stanislav Malyshev wrote:
DH Are you kidding? this would -
DH Of course, the problem is someone would have to BE this
DH central authority, and we'd have to trust it not to misuse all of our
DH addresses...
DH
DH What about spamcop: http://spamcop.net/
I once tried to understand how they
On Friday 24 January 2003 16:38, Amit Margalit wrote:
Hi,
I hate to ask the obvious things, but did you use 'razor-admin' to
register a user/pass pair with Razor2 ?
No, I didn't. If I correctly understood the man page, this is only necessary
if you want to report spam, and at least for the
/usr/bin/spamassassin -a -D -l /tmp/spam_assassin.log
Further to my previous post, I set up the log as Amit suggested, but after
getting 5 e-mails (2 spam and 3 from IGLU), I see that the log, in fact,
contains only a copy of the e-mails before processing by SpamAssassin. I
don't see
But really, in summary - who cares about retribution. Just get a
bayesian filter, they work VERY well, they subsume most other related
techniques in there (a white list is formed automatically, for example),
and it doesn't require much thought.
It's not perfect, so make sure you do go over all
what about that new statistical spam filter(using bassian alg)
did anyone try it out?
Ely Levy
System group
Hebrew University
Jerusalem Israel
On Fri, 24 Jan 2003, Amit Margalit wrote:
Hi,
On Fri, 24 Jan 2003, Vadim Vygonets wrote:
SpamAssassin is quite effective, indeed. I have no
On Thu, 23 Jan 2003, Daniel Vainsencher wrote:
I also use a bayesian spam filter, and am very happy with it. Also
google plan for spam/paul graham - the guy that recently revived the
interest in such filters.
I have also noticed quite a few spams recently that were made to pass such
filters:
On Friday 24 January 2003 04:15, Vadim Vygonets wrote:
Quoth Amit Margalit on Wed, Jan 22, 2003:
I use a combination of SpamAssassin and Razor. Very good.
SpamAssassin is quite effective, indeed. I have no experience
with Razor.
FWIW - I tried SpamAssassin and Razor, but the overhead
Hi,
On Fri, 24 Jan 2003, shlomo solomon wrote:
FWIW - I tried SpamAssassin and Razor, but the overhead caused by Razor was
too much. Instead of downloading 50 e-mails in a minute or so, it took almost
an hour! I checked the headings added by SpamAssassin and discovered that
I don't know
On Friday 24 January 2003 11:49, Amit Margalit wrote:
I don't know when that was. Razor2 appears to be quite fast, unless you
set it up to automatically submit all caught spam to razor.
During the past month. And no, I didn't set it up to report spam to Razor.
As I said, this could be a case
EL what about that new statistical spam filter(using bassian alg)
EL did anyone try it out?
If you mean bayesian filters like bogofilter, I'm using it (bogofilter) at
home and at work and so do a number of my colleaugues and the results are
rather satisfying. The downside of the filter is that it
However, this kind of filter seems to be rather for personal use - because
if I would pass our sales/marketing communication through my database as
it is now, a lot of legitimate mails would end up in 'spam' bin, since
they talk exactly about the same things that many spams talk - prices,
I'd like to ask a related question: what about running a filer on the mail
server?
In the mail server I've set up, the only spam filter I put is denying by
connection address from black lists. This is because it requires
relatively little processing, and is done before the message hits the
queue
On Fri, Jan 24, 2003, Tzafrir Cohen wrote about Re: spam mail filter:
I have also noticed quite a few spams recently that were made to pass such
filters:
Every suspicious keyword was cut in the middle (som of them: a number of
times) by html comments.
Something like:
MA!-- sdfsdf --KE
Hi,
I hate to ask the obvious things, but did you use 'razor-admin' to
register a user/pass pair with Razor2 ?
Can you run SpamAssassin with Razor with some extra logging options and
send the results? For example, call SpamAssassin from .procmailrc like
this:
:0fw
| /usr/bin/spamassassin -a -D
be quite that rosy).
Daniel
Nadav Har'El [EMAIL PROTECTED] wrote:
On Fri, Jan 24, 2003, Tzafrir Cohen wrote about Re: spam mail filter:
I have also noticed quite a few spams recently that were made to pass such
filters:
Every suspicious keyword was cut in the middle (som of them: a number
This is a problem. It's a specific case of the rendering attack - in the
end, telling spam from information is a sort of turing test, and to even
try to make the decision, you have to judge the mail as it is rendered
(not at any of the lower levels at which it is transmitted). But so far,
I
DV you don't get non-spam html comments, eventually !-- and --! will be
DV considered strong indicators to spam, and that's it, they're filtered.
DV That the learning in the filter.
Here I have an example - a spam looking like:
M!--lvPVVE--y n!--lvPVVE--am!--lvPVVE--e i!--lvPVVE--s Rimma,BR
etc.
Stanislav Malyshev wrote:
DV you don't get non-spam html comments, eventually !-- and --! will be
DV considered strong indicators to spam, and that's it, they're filtered.
DV That the learning in the filter.
Here I have an example - a spam looking like:
M!--lvPVVE--y
Are you kidding? this would -
1. Take too much time
2. Require knowing who the sender really is, which is not trivial
3. Run the risk of reporting an innocent, which has a far higher cost
than me personally losing one email.
OTOH, what we could do, is that after we confirm it's spam (my MUA
IIRC mozilla 1.3a incorporates a bayesian mail filter
This will not help you immediately at first, but will quickly improve.
STW (or /., or better yet, LWN) for links to theory and discussion.
also lookup bogofilter .
[EMAIL PROTECTED] wrote:
hi
I am looking for a good mail filter to
I also use a bayesian spam filter, and am very happy with it. Also
google plan for spam/paul graham - the guy that recently revived the
interest in such filters.
Daniel
Guy Baruch [EMAIL PROTECTED] wrote:
IIRC mozilla 1.3a incorporates a bayesian mail filter
This will not help you
Hi,
On Fri, 24 Jan 2003, Vadim Vygonets wrote:
SpamAssassin is quite effective, indeed. I have no experience
with Razor.
First, they can be used together - SpamAssassin will query Razor for you
if both are installed, and having something listed in Razor gets the email
in question quite a few
Quoth Amit Margalit on Wed, Jan 22, 2003:
I use a combination of SpamAssassin and Razor. Very good.
SpamAssassin is quite effective, indeed. I have no experience
with Razor.
I might post a short intro on how to set this up if there is enough
demand.
Aren't the official docs enough? They at
By the way: about a year ago, I was looking for a milter-based library
or program to manipulate e-mails, including attachments. It's good not
only for spamming, but also to other purposes, like viruses. Instead of
blocking all the suspected e-mails, they can be converted to non-
dangerous
hi
I am looking for a good mail filter to filter all the spam mail i get
anyone has experiance with one ?
cheers,
erez.
=
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word unsubscribe in the message body, e.g., run the
Hi,
I use a combination of SpamAssassin and Razor. Very good.
SpamAssassin: http://spamassassin.org
Razor:http://www.vipul.net and
http://razor.sourceforge.net
I might post a short intro on how to set this up if there is enough
demand.
Amit
On Wed, 22 Jan
Hi,
I use a combination of SpamAssassin and Razor. Very good.
SpamAssassin: http://spamassassin.org
Razor:http://www.vipul.net and
http://razor.sourceforge.net
I might post a short intro on how to set this up if there is enough
demand.
Amit
Please
36 matches
Mail list logo
