We’ve been using Akismet to filter our blog post comments and eliminate the spam. It’s been working great, with virtually no false positives and only a few false negatives a day, which are easy for us to moderate. We stopped looking in the spam bucket long ago since I never found a false positive. Until now. And it’s me. Ouch!

A few days ago I noticed that a comment I made, while logged in to our WordPress based blog, didn’t show up. I assumed that I must have forgotten to actually hit the SUBMIT button so I made it again. And it still didn’t show up. Looking into the Akismet spam bucket, I found both comments. Neither one had any spam characteristics, at least not to my eye. I marked them as not spam, and assumed that this might help Akismet recognize future comments from me as not being spam.

But it happened again today — a comment I just made to my own post while logged in to our Blog was identified as spam.

Now I’m puzzled and wish I knew more about Akismet’s techniques. I might have made all three comments from my home, so maybe it’s my IP address (a Comcast address) that’s the feature marking my comments as spam. Does Akismet know that I’ve logged in to our blog? Maybe it shouldn’t matter — I think we allow anyone to create a low level account on our blog.

I’m at home now but about to go to UMBC. I’ll try adding a comment to this post from home and then when I get to work. That will test the theory that it’s the IP address that marks my comments as spam.

I wonder how many valid comments we’ve lost? The level of spam is too high to manually look through them every day.

Related posts: • 93% of blog comments are spam;  • Big spike in blog comment spam?;  • One billion spam comments;  

Posted on Monday, November 27th, 2006 at 1:00 pm in categories [Uncategorized]. Comments feed. Leave a response, or trackback from your own site. edit