Number of comments submitted to posts on this blog has gone through the roof recently as you can see on the graph below. Of course, practically all of these are spam. Unfortunately moderating this flood of crappy advertisements and link baits is now starting to take more of my time than I am willing to spend on it. Since I want to keep my little corner of the web a clean and friendly place I'm closing the comment submission until I come across some viable solution. In the mean time if you have a question or want to contribute something to one of my posts, feel free to sent me an email.
While Akismet has been doing a pretty good job of automatically filtering comment spam for me, it's been letting a non-trivial amount of it through in the recent months. Considering the increase in volume, that might not even be due to decreased accuracy.
The kind of spam I'm seeing is kind of surprising. Spammer fetches the blog post that contains the comment submission form, submits a comment and fetches the post again to verify that his comment is visible. After these three HTTP requests the originating IP is never seen hitting my server again, making me think this is done via a botnet or some other distributed operation like that. There is no obvious sign of crawling so I don't know how they get the URLs to spam. They use realistic looking user agent headers and the only obvious difference to a real browser is that they don't fetch any of the resources (images, CSS, ...) referenced in the HTML document.
The content varies, but a lot of the comments I've been removing manually these days look like bug reports ("the sidebar is not rendering correctly on my browser", "search doesn't work" and such) that are only given away by the obvious keyword stuffing in the author name and the URL (or when they are complaining about bugs in features this website doesn't have). They target both new and old blog posts, so just shutting down comments on old posts doesn't seem to be a solution.