Archive for 2006-05-08
Crusade against trackback spam
2006-05-08The trackback spam is starting to become annoyance for me. Many blog sites are said to turn off trackback due to this trouble. Though my blog is safeguarded with comment authorization plugin (thus no spam can escape my eye
), I still have to manually click, click, click to mark them as spam. Not as troublesome as deleting spam, but still.
As somebody might have already dealt with it, I started to look for existing solutions. There has been various hacks available, and quite a large part of them are dealing with wp-trackback.php.
- This “WordPress Trackback Spam Killer” use a keyword based blacklist. I don’t even try it — I know in the guts that keyword blocking can never be effective. Everybody knows the successful rate (or lack thereof) of spamassassin when encountering non-latin characters. (I don’t mean I’m against spamassassin, it really helped a lot when blocking spam from western countries.)
- Another hack also modifies wp-trackback.php, attempting to apply comment spam filter against trackback message as well. Doesn’t worth my time to try.
- This patch by Elliott Back checks if IP address of some URL is equal to the trackback sender; I didn’t try it, but from its comments, quite some people are complaining this patch blocks many legitimate trackbacks.
Since I want to avoid using patches whenever possible (it is quite difficult to remember which patch has been applied, and it may or may not work with later Wordpress), I try to look for plugins instead.
- Scott Buchanan wrote this plugin. Haven’t tried it yet, but from its description, it seems to block non-permalink trackback linking. Sadly, a few of the spams really do use permalink. But anyway, I’ll give it a try if the following one doesn’t work.
- The one under my probation (Trackback Validator Plugin) is written by Computer Security Lab of Rice University. It determines a trackback is a spam or not based on the assumption that legitimate trackback contains link back to my site. Somehow that looks more logical to me, that’s why I’m trying. Let’s see.










