On 25 Jul 2005 at 15:00, Paul Karagianis <[log in to unmask]> wrote:

> Subscribers can't have reliable (list) mail delivery and use SpamCop.
> 
> Maelstrom was blocked 3 times in 4 days late last week by SpamCop, their
> web page claims that the block is due to something turning up in one of
> their double secret spam traps which they will neither disclose or
> discuss.  

Yup. I've complained to the main Spamcop guy about that. Spammers 
harvest "honeypot" addresses (among others), forge those addresses as 
the purported sender of the spam email. If the mail bounces, or is 
rejected because that address is not allowed to post to the list, or 
is sent a subscription confirmation (for an email sent to the 
listname-subscribe-request address), or if any kind of reply is sent 
back to that address, then you get listed by Spamcop. 

His response can be summarized as "too bad, you shouldn't respond to 
spam".

My attitude is that "if a honeypot has been compromised, it's no 
longer a valid honeypot and should be closed". He agreed with the 
premise but disagreed that spammers forging the honeypot address 
meant the honeypot was compromised.

The only solution is to find a very accurate and very fast anti-spam 
tool, set it up in a SPAM_EXIT, and set BOUNCE_SPAM=0 (the default). 
Unfortunately, I have yet to find a very accurate spam filter, let 
alone a very fast one. Setting up a spam filter will help, but with 
today's anti-spam technology, some spam will always get past and 
generate bounces or replies that will get you listed in SpamCop.

> In all cases where we've had contact with SpamCop we've seemingly found
> them to be polite well-intentioned people who responded quickly and
> corrected *their* problems.  

That's been my experience too, but it does get tiring to be 
constantly on your guard. 

-- 
Francoise Becker <[log in to unmask]>

Knowledge is just a click away: http://www.lsoft.com/optin.html