Spam and sieve vacation
janne.peltonen at helsinki.fi
Mon Aug 27 03:56:20 EDT 2007
On Fri, Aug 24, 2007 at 11:54:38AM -0400, Jorey Bump wrote:
> In my opinion, no amount of backscatter is acceptable, so I don't allow
> user-configurable autoresponders or forwarding. My antispam measures
> have reduced the amount that makes it to the user's inbox to about
> 5/week, so I will make a rare exception, but only if I configure it
5/week? Whee. We might achieve something remotely approaching that with
personalized bayesian filtering (a multi-discipline, internationaly
connected university receives quite a lot of ham that looks very much
like spam, so we are a bit paranoid about false positives).
> Forwarding has proven to be more risky than autoresponses,
> because agressive ESPs can create a temporary DoS to their sites for the
> entire server. This is particularly frustrating when the cause is your
> own user marking a forwarded message as spam.
> On systems that I use but don't manage, autoresponders and forwarding do
> cause problems, and servers get publicly blacklisted regularly. There is
> also an increase in volume caused by the backscatter from autoresponses,
> affecting both bandwidth and storage needs.
> That said, both features can be useful and even justifiable, but have
> fallen into disfavor due to the problems they cause. RFC 3834 compliance
> and continual evaluation of your antispam measures will help.
> Unfortunately, demand for these features often has a political component
> that can affect you professionally, so only you can decide what's best.
I'm trying to battle the political component. ;) I don't have much to
fear, though, since this appears to be one professor against the IT
department, university board, and so on. That is: my boss understands
that this is a trade-off between a useful feature and a risk of getting
Thanks for your insight again.
Janne Peltonen <janne.peltonen at helsinki.fi>
More information about the Info-cyrus