At times I’ve been really impressed with the quality of spam messages I’ve seen appearing on blogs. Spam is a hard problem to solve; if it wasn’t, we’d be rid of it by now.
One of the popular defenses against spam is using a captcha. The basic problem that a captcha is trying to solve is, “Are you human?” The way it does it is by trying to present a commenter (could be human or machine) with a problem that a human can solve, but a computer cannot.
The only reason that squiggly lines and letters are used is that humans are relatively good at them and computers are not.
But captchas aren’t perfect, and they’re only going to get less perfect over time. Why? Because they fly in the face of the very problem that computer science is focused on solving:
“How can we get computers to do stuff for us?”
We want computers to be able to do the kinds of things we can do, so that they can do them for us. (We also want computers to do things we can’t do, but that’s slightly different issue. Actually, anything a computer can do, a human can do, just really, really slowly ; – )
So… at the same time some of us are desperately trying to design great captchas, more of us are desperately trying to design great captcha breakers. Computer scientists around the world are dreaming of a machine as capable as a human, and bloggers around the world are screaming out, “Please, no!”
However, the real problem with a spam message is not that a computer posted it. The real problem is that you don’t want it on your blog, because it’s unhelpful and/or malicious.
The question of “Do I want this?” is a more interesting question to solve. It’s a social question, albeit one in a technical domain. It’s the question that modern e-mail spam filters try and answer, using statistical analysis of word usage to identify stuff you don’t want.
Now… can I think of a computer generated comment that I’d welcome on my blog? The first thing that jumps to mind are related links. If a computer posted a genuine, relevant link (especially if it was a link from a friend’s blog), that might make my post better, not worse, and I’d welcome that. A link to a spammer’s site? Not so much.
But, hypothetically speaking, if a computer could get ‘smart’ enough to post the exact same comment that a friend would have… what’s the difference? The bits are just bits… they don’t care who posted them.