On Wed, Jun 18, 2003 at 06:18:06AM +1000, Peter Kiem wrote: > OK that is fine but a large majority of spam is sent from dynamic IP > addresses which are NOT open relays but just used to spew out millions of > emails to the rest of us poor suckers. > In other words, ALL the ISP's you peer with are not doing what they need to do to keep spammers from using their access points to attack, penetrate and use other systems as mail relays.
Gee - what would happen if you actually verified who a person was before giving them an account? ISP's don't do anything to prevent thousands of spammers from getting access to internet, and then they claim they have to shut down normal rfc-correct practices because the spammers are causing you so much pain. They should just stop peering with any ISP who doesn't verify the real identity of their customers before giving them an account. The real reason spammers exist is the ISP's. The ISP's are so worried about getting every possible penny out of every possible client that they won't even use the identity verification tools ALREADY AVAILABLE from the credit card services to keep the spammers off the net. How about re-writing your peering arrangements so that any ISP who fails to do proper identity checks and cross checks against a database of spammers loses their peering privileges. This would, of course, cost money, so it won't happen. It would also, of course, mean most of the third world ISP's would lose peering for a while, until they got their houses in order. heh - So would AOL, I think. :-) Spammers would then (as they are already doing), pay other people to let them use their accounts to send spam with and those folks would also go in the database. Too bad. > Those of us that have to process THOUSANDS of emails every day know the > VALUE of rejected dynamic address SMTP for cutting down the junk. I agree it is a practical approach to reducing spam, but it is pretty much contrary to the original design and intent of the Internet. Switch to Bayesian filtering. Much smarter. You can even use it on the customer sign up process. :-) You can use Bayesian filters on just about anything. (Its similar, in a crude way, to the tools used to perform speech recognition.) Then once you've identified a customer as a repeat spammer, I think its very appropriate to use a tactical nuke. ;-) -- Jeff Kinz, Open-PC, Emergent Research, Hudson, MA. [EMAIL PROTECTED] copyright 2003. Use is restricted. Any use is an acceptance of the offer at http://www.kinz.org/policy.html. Don't forget to change your password often. -- redhat-list mailing list unsubscribe mailto:[EMAIL PROTECTED] https://www.redhat.com/mailman/listinfo/redhat-list
