You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Justin Mason <jm...@jmason.org> on 2006/11/11 13:27:30 UTC

rule secrecy *again* (Re: Well, that didn't take very bloody long)

Loren Wilton writes:
> >         Ok, remember that "Name Wrote: :)" emails?  They've completely 
> > changed.  Now it's "hi username" instead.  Joy, oh joy.  Can anyone find 
> > any common elements in these emails because whoever this putz is, they're 
> > adapting a lot.  They hit us, we adapt, they immediately change tactics 
> > and come at us again.  Now with all the brilliant minds on this mailing 
> > list, we really should be able to find out who this putz is and nail all 
> > his stuff regardless of what tactic he switches to.
> 
> The reason they adapt is because there are detailed announcements on the 
> mailing list of the things that are easy to spot.  The guy sending these is 
> on the list too, so as soon as the oversight or excessive cleverness is 
> announced to the world, he knows what he has to fix.

ho hum... here we go again. :(

As I've noted several times recently -- these *are* being caught by rules
which were developed "in the open" -- namely RCVD_FORGED_WROTE, which has
been sitting in my sandbox for several weeks, was announced in a checkin
message (with diffs!), and is currently "live" in both trunk and 3.1.x
rule updates.

The rule has been visible since:

  r465179 | jm | 2006-10-18 10:11:15 +0100 (Wed, 18 Oct 2006) | 1 line

  add rule to catch 'Subject: foo wrote:' stock spam

Take a look at the graph of hit-rates over time in everyone's corpora:

http://ruleqa.spamassassin.org/last-night/RCVD_FORGED_WROTE?s_detail=on&s_g_over_time=1&s_zero=on&srcpath=#over_time_anchor

There's been no change in hitrates since 2006-10-18 -- in fact, in
cthielen and zmi's corpora, they rose *dramatically*.

Secrecy is *NOT* an essential element of rule development.  It seems
logical to think it is, but evidence repeatedly demonstrates otherwise.

For some spammers, it may _help_ -- but not for all, so it's by no means
essential.  On the other hand, secrecy damages collaborative development,
restricting rule refinement and improvement to a secret "cabal".  It's
antithetical to open source development.

--j.

Re: rule secrecy *again* (Re: Well, that didn't take very bloody long)

Posted by jdow <jd...@earthlink.net>.
From: "Justin Mason" <jm...@jmason.org>
>
> Loren Wilton writes:
>> >         Ok, remember that "Name Wrote: :)" emails?  They've completely
>> > changed.  Now it's "hi username" instead.  Joy, oh joy.  Can anyone find
>> > any common elements in these emails because whoever this putz is, they're
>> > adapting a lot.  They hit us, we adapt, they immediately change tactics
>> > and come at us again.  Now with all the brilliant minds on this mailing
>> > list, we really should be able to find out who this putz is and nail all
>> > his stuff regardless of what tactic he switches to.
>>
>> The reason they adapt is because there are detailed announcements on the
>> mailing list of the things that are easy to spot.  The guy sending these is
>> on the list too, so as soon as the oversight or excessive cleverness is
>> announced to the world, he knows what he has to fix.
>
> ho hum... here we go again. :(
>
> As I've noted several times recently -- these *are* being caught by rules
> which were developed "in the open" -- namely RCVD_FORGED_WROTE, which has
> been sitting in my sandbox for several weeks, was announced in a checkin
> message (with diffs!), and is currently "live" in both trunk and 3.1.x
> rule updates.
>
> The rule has been visible since:
>
>  r465179 | jm | 2006-10-18 10:11:15 +0100 (Wed, 18 Oct 2006) | 1 line
>
>  add rule to catch 'Subject: foo wrote:' stock spam
>
> Take a look at the graph of hit-rates over time in everyone's corpora:
>
> http://ruleqa.spamassassin.org/last-night/RCVD_FORGED_WROTE?s_detail=on&s_g_over_time=1&s_zero=on&srcpath=#over_time_anchor
>
> There's been no change in hitrates since 2006-10-18 -- in fact, in
> cthielen and zmi's corpora, they rose *dramatically*.
>
> Secrecy is *NOT* an essential element of rule development.  It seems
> logical to think it is, but evidence repeatedly demonstrates otherwise.

Indeed - if you have a rule that depends on secrecy then it is too
fragile to have a long life. Good rules have long usable lifetimes.

{^_^} 


Re: rule secrecy *again* (Re: Well, that didn't take very bloody long)

Posted by Steve Lake <st...@raiden.net>.
At 12:27 PM 11/11/2006 +0000, Justin Mason wrote:
>ho hum... here we go again. :(
>
>As I've noted several times recently -- these *are* being caught by rules
>which were developed "in the open" -- namely RCVD_FORGED_WROTE, which has
>been sitting in my sandbox for several weeks, was announced in a checkin
>message (with diffs!), and is currently "live" in both trunk and 3.1.x
>rule updates.

         Yeah, I pushed my updates for SA and now it seems that those spams 
aren't getting through anymore.  heh.  I can't wait for this spam war to 
end so I can go back to my more laid back 3 month cycle of updates instead 
of 3-4x's a day.  :(


Steven Lake
Owner/Technical Writer
Raiden's Realm
www.raiden.net
A friendly web community