On Jun 26, 2010, at 10:36 AM, Tim wrote:
On Fri, 2010-06-25 at 14:54 -0700, JD wrote:
> I wonder how Google does it. only .01% of my google email is spam.
> The spam folder contains tons of spam, and it is automatically purged
> by google.
When you're a large mail host you have one big advantage in spam
killing: You will receive tons of identical messages, many
addressed to
bogus users, or honeypot addresses (addresses that you leak out,
somehow, that aren't for real mail use). When you receive large
numbers
of identical messages, especially to non-real addresses, you know that
they're spam, and you can mark every single one of them as being spam
with 100% confidence. You don't need to check for false positives, as
no real mail will be sent to such addresses. Whereas it is
possible for
lots of users to receive identical mail, if you have lots of people
subscribed to some popular lists.
I've done that (honeypotting) in the past, and it's a reliable
technique. Unlike many other anti-spam techniques which falsely
identify so many real messages as being spam that they make using
them a
waste of time (if you're having to keep on checking your spam box,
manually, there's no point in running anti-spam software). Not to
mention the problems caused when users have no idea that they must
check
for false detections, and simply never see some of their mail.
This is harder to do on an individual level, because most of your spam
messages are different from each other.
At the individual level, I keep telling myself I'm going to set up a
honeytrap, or maybe it should be called flypaper.
Deliberately leak trap addresses in places I tend to use my real
addresses, auto-blacklist anything that hits the trap addresses.
Haven't got around to it yet.
So you're left with trying to
look for *similarities* to prior spam. Though it is possible to make
use of other people's honeypotting data (the various anti-spam lists
that you've seen discussed in other messages in this thread).
--
[tim@localhost ~]$ uname -r
2.6.27.25-78.2.56.fc9.i686
Don't send private replies to my address, the mailbox is ignored. I
read messages from the public lists.
--
users mailing list
users(a)lists.fedoraproject.org
To unsubscribe or change subscription options:
https://admin.fedoraproject.org/mailman/listinfo/users
Guidelines:
http://fedoraproject.org/wiki/Mailing_list_guidelines