Over at Ajaxian.com, we frequently have to spend several minutes a day manually parsing through comment spam (not all of it, thankfully; just the close calls). Starting with the past week or two, I’ve noticed a huge jump in the difficulty level of determining whether something is spam or not. Check out the following spams:

1. On a WordPress 2.x story: “Progress is normally a great thing but ever since WP 2.X IBP stopped working.”
2. On a Harry Potter story: “dis is kul!!!!!!!!! harry potter rocks!!!! harry potter rulz!!!!!!!!! cn’t w8 4 the nxt buk!!!!”
3. On an Ajax chess website story: Are you interested in learning chess? Do you already know how to play and want to advance your game to a new level? I can help. My name is Karen and I have seven years of experience working with students of all ages and abilities, from grade one through twelve. You can contact me at karen@celll-phone-accessory.org thanks

All of these comments appears to be legitimate (if inane); the only clue is the website associated with the commenter (not shown) and in the last example here of course the email address is a clue.

These are by no means the best examples I’ve seen, they’re just the ones I have on-hand right now. I’ve seen some very suspicious comments that seem to fit right in to the technical discussion being had — until you look at the website. And even then, it’s not always clear at all… and what if the commenter just happens to work for an e-tailer? Ugh… when humans have a hard time figuring out whether a comment is spam, I’m guessing it’ll take a while for automated comment processors to catch up.

Perhaps it’s time to strongly encourage login accounts to comment on stories…