Re: Questions about auto-submitting from spam traps (and trust) (fwd)

[Justin S. <lis...@nu...>]

On Thu, 25 Sep 2003, Frank Tobin wrote:

> Simply reporting is fine.  If someone is looking at the count of reports 
> in the pyzor database, you will contribute to their knowledge.
>
> > When my spamtraps get spammed I tend to get hundreds or thousands of
> > copies of the same spam.  They come from different proxies all over the
> > world but they have the same body.  I don't know of a way for me to
> > filter out duplicates in an auto-submission process.  I don't know if
> > it's a possibility at all.
> 
> Don't worry about duplicates.  If you really wanted to avoid it, though, 
> it's easiy with formail (part of procmail); it has a built-in capability.

That makes sense.  One question I've always had but never asked is how 
much of the message does Pyzor use?  Is it just the body, subject + body, 
or everything?  Since I'm running numerous spamtraps on the same domains I 
tend to get hundreds or thousands of copies of the same spam from hundreds 
or thousands of open proxies globally.  If only the body is used then 
perhaps I should be using formail to weed out the dupes and save the load 
on your server(s) and my own.

> Pyzor really can't provide you with any such feedback.  I admit it is
> definitely a desirable charactistic of volunteer-based systems.  I'll 
> try to keep it in mind.

The only real feedback I need is to know that the message was accepted.  
I had trouble getting razor set up in the beginning.  Now I can see
message accepted entries in it's log.  That helped.  As long as I know the
messages are received then all is fine by me.  Of course it would also be
good to know if the submissions were rejected later though.  I wonder if
it would be possible to have an online stats system that would list by
userid that submitters stats such as total submitted, total spam/ham,
daily totals, messages rejected (if they are), etc..  Would that be a
possible future project?

Anyhow, thanks for the reply
 Justin