That's usually referred to as a "white list" and I actually think it's a great idea for most people's personal email. If you have kids, it's the perfect solution. That way, you can email them and even their friends can email them, but they don't see the spam.

I don't know if there is a workable solution to spam, because it really comes down to the old "tragedy of the commons" thing, but I wonder if a combination of cryptographic signatures, public key infrastructure, a web of trust, and white lists, would get us a 99% solution.

I haven't entirely fleshed it all out, but it works something like this: we all establish identities for ourselves using public key cryptography. Now I can sign an email that I send to you, and you can know with absolute certainty that it came from me and that it hasn't been altered.

Next, we sign the public keys of people we personally know. So for example, I'd be happy to sign toseek's public key.

We set up white lists, and only accept email from people we trust. I'm on your white list, so you accept email from me. But one day, toseek wants to email you. No problem, you look up his public key in the PKI and find that it's been signed by me - and I'm someone you trust. So because you trust me, you also trust toseek. He is added to your whitelist and the email is delivered. All of that is automatic by the way. From the user's standpoint, the email just arrives in your box as it does now.

What this does is, it creates a very small barrier to entry for anyone who wants to send email. I can't just go out and create an anonymous account (tofu@myhost.com) anymore. Well, I can create that account but nobody will get the email I send from it. First, I have to make a friend and have them sign my public key. Then I'm inside the web of trust and I can send email.

Of course, spammers can do this too - this is why it isn't a 100% solution. It's only a 99% solution. So let's say that toseek turns out to be a spammer (something I've always suspected). He creates an account, alice@wonderland.com and uses his toseek account to sign alice's PK and admit her to the web. Now he starts sending spam from the alice account. Because I trust toseek (big mistake) I also trust alice and her spam makes it to my inbox. The spam also makes it into your inbox because you trust me.

So what we do is, we mark the messages from alice as spam. After a certain threshold, say 3 or 4 spams, alice is blacklisted from the PKI and removed from the web of trust. Of course, toseek can just create another account, right? Well, having someone you trusted become blacklisted also counts against you. So toseek gets a black mark too. After he has trusted some threshold of spammers, say 3 or 4 of them, we blacklist him too.

So, this wouldn't stop spam, it would just create a barrier to entry for the spammers. Today, a spammer can send a million emails with the click of a mouse. Under this system, then can't send very many at all before burning up an identity.

There's probably some little devil in the details that makes this scheme not work, or someone else would be doing it already.