(2003-10-13) Bray Spam Pay Mail Relay

Tim Bray [on](http://tbray.org/ongoing/When/200x/2003/10/12/Spam Plan27) advancing the SpamWars by setting up relays as certification authorities, and you pay to send through them. The model is:

• I set my filters so that anything not in my people White List, or signed by a relay I trust, goes into a probable-spam hole. (Actually, Tim says every message needs to be signed. I guess geeks (senders) could get their own signatures; this would allow them to Route Around the relayers, but that would increase the hassle to each receiver having to accept more unique signatures).

  • Jon Udell has written about signing emails for a long time.

• I publish in my various spaces (blog, FOAF) that anyone (who I don't already know) who wants to contact me needs to use such a relayer (maybe I suggest a couple that I've already decided to trust) to sign their messages.

  • is there a mechanism where once I (the initially-stranger) establish contact/trust with you, I don't need to use the relayer any more? That could be a hassle, since you would have to switch signatures to associate with me.

Random thoughts:

• it's still mainly a client-side solution, since each user (recipient) needs to decide to start accepting messages signed by each given relayer.

  • This doesn't deal with the network congestion costs of spam, though maybe it does in the long term by making spam not worth sending

  • Will users just mindlessly accept every request, just like they click on attachments? Then every spammer could just run his own bogus relay.

  • So short of his side-point of having the government certify (some of) them, I wonder whether this would actually work.

• the point isn't the payment so much as the relayer having software/policies to ensure you can't send more than 100 emails in a day without an (email) exchange to verify that everything's all right. So, potentially, AOL/Yahoo/MSN (or any other group, like SlashDot) could run relays as a free service.

• his $0.01/per message is like a $10CPM...

Update: Tim has update the original page with comments and clarifications. He wants a very small number of relayers; this reduces the pattern of "training" users to accept any old signature.