Infinite Monkeys and other BLs

[Richard Lynch]

Chris W. Parker wrote:

>Richard Lynch <mailto:rich at MAIL.WVNET.EDU>
>    on Wednesday, March 17, 2004 4:08 PM said:
>
>
>
>>What I want is
>>accurate denial of messages based solely on content rather than where
>>it comes from. Unfortunately, we're not there yet.
>>
>>
>
>what about bayesian filtering?
>
>i've got ms doing basic sa, and attachment type based filtering. this
>lets through more spam than i'd like but [1] i'm using an old version of
>ms and sa, and [2] i don't use any advanced features (bayesian, rbl,
>pyzor/razor, etc.). however, i've got a bayesian filter for outlook 2000
>installed and it catches everything else.
>
>i've never had a false negative* so i imagine that if i put bayesian at
>the gateway my spam count would go way down at the client level if not
>disappear.
>
>
I do use the Bayesian filtering in Spamassassin at the MS gateway and it
works reasonably well.  But, it's not so good that I can give up using
RBLs.  Filtering at the end user level certainly works better but that's
not practical for 60,000+ distributed users.  I'm stuck with a
multi-pronged approach for now (RBLs, Spamassassin, Bays, DCC, and
custom rules).

>chris.
>
>* i always get confused on this but.. false negative is when spam is not
>detected as being spam right? and a false positive is when non-spam *is*
>detected as spam?
>
>
Yes.

--


-------------- next part --------------
A non-text attachment was scrubbed...
Name: rich.vcf
Type: text/x-vcard
Size: 259 bytes
Desc: not available
Url : http://lists.mailscanner.info/pipermail/mailscanner/attachments/20040318/5aa29c5c/rich.vcf