On Mon, 26 May 2003, J. Noel Chiappa wrote: >... > > The *only* thing that's going to stop spam is charging for email. Everything > else is a waste of time, because you're going to run into impossible > arguments trying to define what's spam, and what's legitimate bulk email > (q.v. the recent message about IETF-Announce email). > > Needless to say, charging for email means changing the underlying protocols. > That's us. It's not going to be simple (we need an underlying micro-payments > system), and it's going to take a lot of attention to detail (q.v. > IETF-Announce again), but I think it's the only real cure. What I'd say is that the only way to reduce unsolicited email to a tolerable level is to change the economics of sending spam such that the folks paying to send spam will focus their emissions for much higher positive response rates. Sending physical advertising has many costs in addition to the transport cost so it probably costs more than the US 1st class mail rate per piece even though the advertising rate is much lower. While I get more junk mail than I'd prefer, 90-95% has some relationship to my interests. If .01% of the spam I receive has a relationship to my interests, I'd be surprised (getting rich quick isn't really an interest ;-:). There is an active proposal, perhaps already submited, in process by Rep. Zoe Lofgren to establish a penalties and a bounty for spam. This is an interesting variant of the enforcement paradigm, but I agree with those who have already suggested that our government too many larger crimes to consider. Instead, I believe some form of pay to send stamp based approach in parallel with the current 'free' system has the best chance of curtaining spam while preserving the 'free' nature of the current system. This will require new protocols as well as some social/legal infrastructure to accomplish (an effective payment system for example). I would expect that a portion of each sender fee would be credited to the receiver and some portion would support the new infrastructure. If we are able to design the protocols I imagine, the mail recipients will drive adoption fairly quickly once implementations are available. I would expect that an early filter criteria would be whether the mail arrived via the new protocols. No, don't relay to my blackberry, cell phone, etc. No, stick in a 2nd class inbox. Etc. Since part of this solution would have to be the identification/trust kind of mechanism Paul Vixie mentioned, easy to implement and utilzed filters could be based on the sender being identified. I will shut up now because I think the point of an IETF list discussion should be to build momentum for working the problem and not design complete solutions. Dave Morris