As everyone with WordPress knows, Akismet is the best plugin for it. Akismet catches all of the probable spam comments and holds them in a queue until they are approved. Originally I noticed all of the blog spam comments I received were complete gibberish and contained links to who-knows-where. Lately, their messages have begun to get more interesting. For instance, take a look at these comments:
- In truth, immediately i didn’t understand the essence. But after re-reading all at once became clear.
- How much money does the government take out of online contest prize money?
- How much money does the Treasury typically print?
- Really enjoyed this! Well done!
- How can i use money on my paypal balance instead of using money out of my bank account?
All these comments could work had they not been found connected to a post that had nothing to do with online contests, the Treasury and PayPal. The other two, just didn’t make sense within the context of the post either. Nothing I’ve written about Mount Bonnell was particularly deep and therefore, no real essence. Had I seen this posted under a post detailing some interesting subject, I’d almost let it go (but then you’d see who it was from and that would tell you right away it was spam).
Anyway, with these comments becoming more and more probable as true comments to your article how do you think the spam filters are going to work? I could see them integrating a reasoning-type procedure that does some analysis of the original post versus the content of the spam comment a decision could be made based upon some relevance calculation. I suppose the other thing to do would be to force all commentators to register for your website and then the admin can accept / ban / remove the users as they see fit. However, it doesn’t mean that a spam bot still couldn’t get through — it just might take longer for them to be accepted. Unfortunately, I just don’t see a failproof way to remove all spam comments from a blog without some sort of human interaction. I’m sure we can come very, very close, but the last 5 to 10% of the spam is probably going to need to be determined by an intelligent human.