Solving the Spam Problem
Adam Kalsey has a piece on comment spam advocating attacking the tools used to send spam. His idea is that if you make spam tools harder to purchase or download, there will be less overall spamming done. This is akin to non-competitive enzyme inhibitors, which slow down the rate of enzyme reactions by knocking out reagent molecules. While taking down a few download and shop listings will make it slightly harder to get those tools, it’s just one small bump in the road for a spammer. He’ll either find them eventually, or write his own.
Six Apart suggests a number of end-user techniques, such as forced registration, capchas, and comment moderation. The problem with these solutions is that they require additional work to block spam. Ideally, anti-spam solutions operate transparently, without administrative or user attention. Only in extreme cases should spam require human attention. Spam Stopgap Extreme, a Wordpress Plugin I wrote to fight spam, does just that. It uses javascript to compute a value, and then checks it on the server. Clients never know that when they click “submit” anything is happening, and spammers don’t yet execute javascript on pages (it would be too costly).
Simon Willison advocates a number of heuristics techniques. Redirect links so that spam in comments won’t improve their pagerank. Keep a blacklist of known spammers’ domains and block them. Wordpress and the spaminator plugin embody heuristicism to its fullest, counting heuristic strikes against a comment until it is either processed or falls beneath the spam threshold. Unfortunately, no heuristic is perfect, or even good. Moreover, heuristics don’t evolve on their own with time, unlike comment spam. And counting words via baysian analysis will miss some spam, and catch others. It’s not a magic bullet either.
So for now, what can you do? How can we tell that a comment “belongs” on our blog? What we need is a secure method for authenticating users transparently. Make everyone responsible for their comments. Make users prove that they’re really humans, and not evil robots. Or just turn off the comments…
| This entry was posted on Saturday, December 18th, 2004 at 10:24 pm and is tagged with bump in the road, simon willison, human attention, server clients, enzyme inhibitors, spam problem, magic bullet, extreme cases, reagent, pagerank, spammer, spammers, moderation, molecules, wordpress plugin, threshold, strikes, blog. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback. |
This topic is quite trendy on the Internet right now. What do you pay the most attention to while choosing what to write ?
nice article! nice site. you're in my rss feed now
keep it up
Blog spam is dying and I couldn’t be happier.