I expressed a faint hope yesterday that the spam problem would be solved, and wouldn’t you know it, a proposal along those lines popped up in my RSS feeds. A couple of researchers have published a paper describing a way to use social networks as an anti-spam tool. In brief, the idea is to build e-mail cluster maps. As reported in Nature, the researchers:
…decided to tackle the problem by taking advantage of the fact that most people’s e-mail comes from a limited social network, and these networks tend to be clustered into clumps where everyone knows each other.
If I understand the concept correctly, those of you who decide, on a whim, to e-mail me with a question about CSS or to comment on something I’ve written would never get through if I were using such a system. If you’re nowhere near my cluster, then I don’t see how you’re going to get through. If the only criterion for being assumed a ‘real person’ is that you’re sending from a cluster, then all spammers would have to do is form their own clusters.
I’m quite interested in social networks these days, but I’m not sure that spam is one of the problems a social network can really fix. At this point, I’m coming to believe that e-mail delivery fees are the only possible solution, and I have grave doubts it would work. I’ve seen this idea described a few times, and here’s how it generally works.
- Everyone gets to set a cost for accepting mail. I could say, for example, any message has to have a 10-cent delivery fee paid for me to even accept it. You might set the threshold at five cents, or 50 cents.
- When sending a message, you authorize up to a certain amount to be paid for delivery. I might say that I’ll attach three cents to every outgoing message. For any account with that delivery fee (or lower), the message will reach the inbox, and I’ll be charged three cents. For any account with a higher delivery fee, the message is bounced back with a “needs more money to get through” error.
- Anyone can choose to refund the delivery fee, either one at a time or by creating a “free entry” whitelist. So I might set my delivery fee at 50 cents, but permanently give my friends a free pass into the Inbox. For random correspondents with legitimate inquiries, I could give their delivery fee back. For spam, I could read it and collect the delivery fee.
It sounds great, and the technology could probably be created without much difficulty. The general idea is that if you don’t want to see spam, you reject all messages with too low a delivery fee; if you want to stick it to the spammers, you read their messages and collect their money. I still see a few problems with the idea.
- If a spammer manages to fool my system into thinking the spam is coming from a friend, it gets in for free. If the mask is good enough, I never get a chance to collect the fee.
- Who’s going to volunteer to run the micropayment system that would have to underpin the whole setup? And if there are no volunteers, then where’s the business model that would be needed to get a company to do it?
- How does one keep the spammers from hacking, bypassing, or otherwise fooling the micropayment system, and wouldn’t any effective techniques to do so work just as well for the current mail system?
- Assuming there is a micropayment structure in place, what’s to keep large ISPs from charging everyone a cent to pass a message through their servers—thus making e-mail no longer free for anybody?
Maybe that last point would be an acceptable price to pay for ending spam. It would pretty much kill off listservs, though, and that would sadden me quite a bit. Even at a penny per message, every post to css-discuss would cost $35.67 to deliver to all the subscribers (as of this writing). On average, we get about 50 posts per day, so that’s $1,783.50 daily, or $650,977.50 annually. If I had that kind of money, I’m pretty sure that I wouldn’t spend it on a mailing list. I’d buy a Navy fighter-bomber instead.
On the other hand, with a delivery-fee e-mail system in place, I could very easily set up an account where people could send their CSS questions for a delivery fee of, say, $29.95. If I accepted delivery, that would get you a detailed answer to your question, or else a refund of the delivery fee if I didn’t answer (or there was no answer to be had). So that would be kind of cool. I suppose I could approximate the general idea today using PayPal or some such, but that would mean going to the effort of setting it up, which isn’t something I’m likely to do given that I have no evidence that there’s any real demand for its existence. Actually, the presence of css-discuss pretty much says that there isn’t, since it’s a whole community of people providing help for free.
So I got started on all that because of the idea of using social networks for spam countermeasures, and like I said I’m interested in social networking these days. In that vein, I was rather amused to see myself at #2 and #4 on rubhub’s new Top 10 lists, and not in the least bit surprised to find Zeldman sitting atop both lists.
I was also quite fascinated by Jonas’ ruminations on how XFN, VoteLinks, and related technologies can easily form the basis for rudimentary trust networks. Jonas, a sociologist by training, has been writing some very interesting things about the semantic web and social networking recently, which is why I’ve just added him to my blogroll and RSS aggregator. As he points out, combining XFN and VoteLinks would be a snap, and has the potential to enrich the semantics of the Web. Instead of just counting links to a page, a community assessment of that page could be tallied.
What interests me even more is the next step. What else can be done with link relationships, and how will the pieces fit together? How many small, modular metadata profiles would it take to begin semanticizing the Web? I suspect not too many. This seems like a clear case of emergent properties just waiting to happen, where every incremental addition dramatically increases the complexity of the whole. John Lennon once said that life is what happens while you’re making other plans. Meaningful technological advancement seems to be what happens while committees are making other plans. It could very well be that the Semantic Web will come to pass because the semantic web arose on the fringes and paved the way—that the latter will become the former, simply by force of evolution. That strikes me as rather poetic, since it means that the principles Tim Berners-Lee followed in creating and defining the Web would become the keys to where he wants to go next.