Some thoughts on MediaWiki spam
And then there's MediaWiki spam. Back in January, when I set up MediaWiki on my server for the Mondoglobo Wiki, I blogged a post that read, in its entirety, "I have seen the future and it is MediaWiki." This remains true, in large part -- Wiki is a powerful tool for organizing thoughts online, and allows collaboration in ways I find very compelling. (We'll look aside for the moment from my hatred for MW's performance, though...)

But MediaWiki's very power in allowing free-for-all editing means it's a great target for Webspam. And unlike a forum, a Wiki venue allows spammers to deface content at will -- and that sucks.

So something has to be done, of course, and I do believe that MW has some anti-spam extensions available, but I want to leverage the existing smarts of the modbot.

As I see it, there are two reasons MW spam is tricky. First, the new text is stored as a blob in the text table -- you have to do a diff yourself to see what changed, and if the diff tests positive for spam, then you need to roll it back. Second, you can't just delete the revision; you have to roll the page text back. Fortunately, rolling back looks like it's going to be relatively simple; you just retrieve the old version from the archive table, and write it into the current page table. Then you delete the revision record. Or so I presume, from some poking around yesterday. I haven't tested it yet.

I've got two spam changes waiting for modbotting, though. So I hope I'll be able to find some time this week to knock them out.

August 4, 2008 - Argh -- let me correct myself here by saying that MW spam has in the meantime become a serious problem. Since I don't have time right now to despam the Wiki, I've locked it, which is Not Good. But: No Choice. For now. There's just so long I can watch the server load hang above 20.

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.