With cruise missiles and smart bombs dropping on Iraq, writing about spam seemed insipid until I did the math. What if Saddam Hussein, in a nerdy stroke of evil genius, decided to spam America into submission? What would it cost? Not as much as I thought, that's for sure.

Here is my calculation. I get e-mail associated with the PBS website for my show, "Nerds 2.01: A Brief History of the Internet." That show is a little dated and the last legitimate e-mail I received through the link came on April 19, 2002. Yet every day I get about 15 spam messages to this account. Since I have never sent a message from that address (I can't — I can only reply), the spam has to come from list-builders scraping that address off the screen. Where I find this account very useful is in gauging how much spam goes to average folks who don't write weekly columns — 15 messages per day.

There are supposed to be 170 million American online. Multiplying this number by 15 messages, with an average message size of 2K, yields 2.55 billion messages, totaling 510 gigabytes per day of spam. Experts suggest that spam is approximately 35 percent of all e-mail, so Americans probably receive about 1.5 terabytes of e-mail per day. And while there is some room for growth in this figure, we can guess that the system is a bit rickety and could be easily overwhelmed if the effort was concentrated.

So what is the commercial value of 510 gigabytes of spam? How much would it cost to send? The going rate, I'm told, for spam sent from someone else's server using their address list is about $400 per million messages, which means 2.55 billion messages would cost $1,020,000 without attempting to apply any volume discounts. One million dollars per day is the size of the U.S. spam industry, and since U.S. addresses get about 40 percent of the world's e-mail, that means the world spam industry is worth about $2.5 million per day or less than $1 billion per year.

As U.S. senator Everett Dirksen said during the Vietnam war, "A billion here, a billion there, and pretty soon you are talking about real money."

So Saddam (who doesn't have time to set this up now, trust me) could spend $1 million with foreign spamlords to buy half a billion messages that he'd send strictly to .gov and .mil addresses, which would surely be enough to overload those servers and turn George W. Bush into a Hotmail user and throw the Joint Chiefs into ICQ. Frightening, isn't it, that so little money could buy so much trouble?

All the more reason why we need to fix this spam problem for good.

Last week, I proposed that we make the spammers pay to spam us, only to have my very intelligent readership shoot that one down immediately. Nobody wanted to share their address book with the world, and who can blame them? PayPal, which I suggested as the logical host for the payment service, seems to be universally reviled. Not only wasn't the idea popular, but readers suggested any number of ways it could be compromised. Back to the drawing board.

Many alternative plans were suggested. Remember that we want to do this as simply as possible and without passing any laws, which would be pretty unenforceable anyway.

By the way, there apparently is already a pay-to-spam system much like the one I described called Cruelmail, which is a wonderful name. You'll find a link to it under the Links of the Week button on this page.

One popular suggestion was to rein-in rogue SMTP servers that are pretending to be other than what they really are. This is a good idea, and could probably be implemented through the IETF and applied globally. You'd need some kind of enforcement mechanism, probably through a DNS-like service that would reveal only legitimate mail servers with some further kind of identifier to link messages and servers. I can see some value to this, but implementing it is months or years away. Besides, there will continue to be clever exploits and the pain of blacklisting servers that really shouldn't be blacklisted (like cringely.com, which has been blacklisted in the past) would still be a problem. It is too darned hard.

Many people are happy with their spam filters. Some people are happy with their SPAM, claiming they really like to get that stuff. Some stoically admit that they clicked the box saying "send me e-mail updates" on a few dozen websites and so are resigned to getting spam forever.

Not me. I still like the pay-to-spam idea, but it needs refinement.

Microsoft, of all outfits, suggested a refinement of what it means to pay. What if, rather than paying in money, spammers had to pay in computer cycles? This is mentioned in an article in the April issue of the MIT Technology Review. This proposed Microsoft solution to spam: make people pay to send email by tying up their computer for a few seconds with some computation each time. Sending the occasional mail to your friends or co-workers isn't considered a problem by Microsoft, since we're only talking a few seconds of your computer's time. But for a spammer, sending thousands of messages per minute, obviously would have problems. The idea has some merit, especially if you could actually get people sending you email to do something useful for you with the computation they're asked to do. It would be an interesting way, for instance, to increase the power of GIMPS, or SETI at Home. But from where I sit, this is still too complex and requires modifying every e-mail program everywhere. Not only that, but it opens an entry point for hackers, who might turn that useful task into an exploit.

What we need is a scheme that doesn't require giving your address book to anyone, doesn't require a list that is updated from any external source, doesn't require messing with PayPal or any payment organization, yet still makes spammers pay. It has to be a system that can be implemented by a single person and doesn't require the world or the world's systems to be altered. Yet it also has to be something that individuals could have up and running by next week, putting an end to unwanted spam while keeping the wanted stuff.

There is an answer. I didn't come up with it. It came from a reader, and while I discounted it at first for the same reasons you will, it grew on me. After a couple days, I am now convinced it can work. If we all do it, spam will quickly die. But if only you and I do it, we will stop getting spam instantly.

The trick is to acknowledge that it is more important that the spammers pay than that they pay ME. We need to exact a price that doesn't compromise our address books or require the involvement of any outside service or system. And here it is. When e-mail arrives, your e-mail client should compare each message to your address book and allow through any messages from known and accepted addresses. In order to work, of course, this list can't include your own address since that is a common cheat. So your address book is involved, but it is never exposed or shared.

If a message comes from an unknown source, an automatic reply is sent saying, "I don't know who you are. If you want me to read your message, call me at 707-525-9519 and we can arrange to allow future messages to come straight through."

That's it. No spammer is going to ever call simply because it is far too expensive. They would have to pay, though not pay me, but the effect is the same. People who really want to reach you WILL call. If they don't, then that's their problem.

I already get enough mail as it is.