Charles Arthur On Technology

We can't always rely on a hurricane to protect us from the irritations of spam

It's an ill wind, as they say. If you believe the word on the anti-spam lists, Hurricanes Charley, Frances and Ivan have all disrupted one of Florida's major internet businesses: spamming. Some people reckon that spam volumes have dropped because those generators of junk e-mail working out of locations like Boca Raton have been without power, and perhaps homes.

It's an ill wind, as they say. If you believe the word on the anti-spam lists, Hurricanes Charley, Frances and Ivan have all disrupted one of Florida's major internet businesses: spamming. Some people reckon that spam volumes have dropped because those generators of junk e-mail working out of locations like Boca Raton have been without power, and perhaps homes.

It's hard not to feel a shiver of schadenfreude - tempered by the realisation that a million other people have been uprooted. The collateral damage of that small reduction in spam (if indeed it's happened, which some doubt) is too great to bear.

And that is the sort of situation we find ourselves in with the wider war on spam, a parasite on the net that just keeps growing. The suggestions for how to get rid of it appear to people to be worse than simply coping with it.

And where are fingers pointing? Rather emphatically at Microsoft, with organisations as diverse as the Internet Engineering Task Force (IETF), which is the main standards organisation for the internet, and AOL, still the world's biggest ISP, rejecting Bill Gates's ideas on getting rid of this annoying, destructive menace. And spam keeps growing: the checking service Postini reckons it comprises 79 per cent of all e-mail, and Commtouch Software, which sells anti-spam services, estimates that the number of unique new unsolicited e-mail messages has risen 42 per cent, from 350,000 per day at the end of 2003 to 500,000 a day by the end of June. (Why "unique new messages"? Because each message gets sent by the multi-million. The challenge for spam filters is to deal with differing messages; if all spam was a single e-mail, we could just block that.)

It didn't have to be this way. In January, Gates promised that spam "will soon be a thing of the past". Speaking to the World Economic Forum in Davos, he gave it two years. I can understand his reasons for giving that timeline. Two years is a very long time in technology: give Microsoft's smartest programmers that sort of deadline and they'll easily come up with something that will catch all spam at the server - because catching spam has to be done by verifying identity, not the arcane methods generally used to examine content for hypertext links and "meaningful" words. (The reason why spam these days contains strange snippets of news stories or juxtaposed words from dictionaries is that spam filters think those must be sensible e-mails from real people.)

But two years is also long enough to work such a technological solution through the tortuous process of approval by standards bodies such as the IETF - because there's no point in solving spam just for one corner of the internet, especially not when you're Microsoft, and hence everywhere. So Microsoft drafted its plan, and took it to the IETF and other internet bodies.

Called Sender ID, it is a complex method in which any e-mail would have to be authenticated as coming from the domain, or person, who claimed to be sending it before the recipient's mail server would pass it on. It's a complex process, which Microsoft essentially rolls into a black box in its diagram at www.microsoft.com/mscorp/twc/privacy/spam_senderid.mspx. There is some explanation in the documents on that page, and a 3MB PowerPoint presentation on how it might work.

Sender ID sounded great - on the surface. "It's royalty-free!" said Microsoft. To which everyone else replied, "Who said anything about royalties?"

It turns out that Microsoft has patented parts of the Sender ID framework. Royalty-free it might be, but that's not the point. When you have something like the internet, which relies on openness and sharing, to have a dominant company trying to get everyone to adopt its own patented scheme jars. Advocates of open-source systems rejected it last week - joining AOL, which has already said it will implement a smaller, non-Microsoft part of the scheme called the Sender Policy Framework.

Microsoft argues that the patents, and (free) licences it was also demanding from anyone who adopted Sender ID were necessary to forestall any future lawsuits about who owned the intellectual property of the scheme.

Microsoft's argument has its points - if you don't take the wider good into account. Unfortunately, in this instance the company looks hidebound, unable to adapt to the idea of simply creating something in order to benefit others. This is strange, because Gates's charitable work is well-recorded (and many Microsoft personnel do plenty for charity, if less visibly). You might think that giving away software would be even easier - after all, it's the day job.

But that's not the Microsoft way. Which is a pity. I don't think we can, or should, keep counting on hurricanes to cut spam for us.

www.charlesarthur.com/blog