[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Unusual spam recently - hummm

Quoting Michael Stone (mstone@debian.org):

> On Thu, Jun 03, 2004 at 04:24:35PM -0700, Rick Moen wrote:
> >One can pretend that the matter's open for debate, but that would be a
> >waste of time:  It's happening.
> Sure it is. How do you manage to sleep, fixing all the email systems in
> the world *and* evangelizing at the same time? Must be tough.

(Troll duly noted and ignored.)

> >People who put significant stock in content-based filtering are pursuing
> >a losing antispam strategy.  
> All antispam strategies are a losing strategy, unless you're a zealot
> who believes that everyone else in the world (and human nature) will
> change. 

My, you _are_ a veritable fount of non-perspective, aren't you?

Some antispam strategies are by relative measures non-starters relative
to others.  In particular, doing all of your primary processing via
content-based filtering, particularly at the MDA/MUA level, has proven
over time to be particularly ineffective, and tempts people to do dumb
things like /dev/nulling all HTML mail.

> Every time you come up with a new fix someone else will come up
> with a new strategy to avoid it. What's your strategy to cope with
> perfectly valid smtp messages that just happen to contain spam?

DNSBLs, collaborative detection, and several measures that include
Bayesian recognition during SMTP time as a _secondary_ measure.  (Was
there a particular part of the word "primary" that you didn't get,

Ultimately, it would be good to collectively whitelist IPs that show
willingness to be responsible for mail they send, and bandwidth-throttle
others.  There are proposals being worked on for this.  See, for example: 

> >They'll probably figure that out by themselves, eventually.  In either
> >case, not my problem.
> Yet you can't seem to stop making it your problem. Curious.

What part of my describing strictly local measures, and insisting that
other people's misbehaviour is their problem, were you not quite

Reply to: