The LazyWeb Works! - solution to the comment-spam problem

Only two days after my lazyweb request for a way to combat comment-spam, Jay Allen posts this solution to the problem.

Jay's approach is exactly what I was looking for - he doesn't attack the problem based on how the spammer reaches your blog (IP blocking, hidden fields on the comment-form, etc), but rather he blocks based on the URL that the spammer is attempting to leave in your comment-system. His solution is also integrated in MT in that you won't need to edit the MT source and can add blocked URL's using MT's configuration interface. Very cool indeed.

There are a few minor downsides though:


  • Spam comments may not show up on the site, but are still stored in your database and therefore still need to be manually removed. To ease this process, I'll probably setup a private comment-spam rss feed to alert me when this is necessary. This feed will contain all entries that are filtered out of my weblog and regular comments-feed, and will have a button or link added to the description that I can click to remove the spam from my system.

  • Until manually removed, <MTEntryCommentCount> will still show the total number of comments - including the comment-spams

  • Spammers could theoretically still bypass the system by linking by IP or rewriting the URL using &#xx; characters. However, since Google will most likely be as confused by this as Jay's anti-spam solution, linking like that won't help spammers increase their Google ranking. Without that payoff, it's unlikely they will do this.

  • It does not only block comments containing a link to a blocked site, but also comments that just contain the name of a blocked site. So discussions about comment spam that use examples of spammer-urls in plain text will be filtered out as well.

Despite these minor shortcomings, I still think Jay's solution is the best one I've encountered so far, and hope it will keep this weblog spam-free for the foreseeable future. Thanks Jay!

TrackBack URL for this entry: http://www.hutteman.com/scgi-bin/mt/mt-tb.cgi/89
Comments

My blog is Domino based ,so I don't know about cross implementation, but I will soon do something very similar to the above.
In the comment I will compare any text, whether it is in a URL or not, to a list in a configurable document containing all manner of words and phrases that would make my mother blush.
I was originally going to just siphon them off to a different view (not display them as a live comment), but I like the idea of publishing them to a personal rss feed with a link to delete them from there.
Very cool. Thanks

Posted by Andrew at October 1, 2003 3:37 PM

To Spam or Anti-Spam, that is the question...
If there's one type of spam that gets on my nerves more than any other, it is spam that promotes anti-spam tools. Sure, anyone that reads this spam is clearly in need of some protection, but by doing this, the anti-spammer becomes a spammer. And if som...

Trackback from public virtual MemoryStream at November 7, 2003 11:31 PM
This discussion has been closed. If you wish to contact me about this post, you can do so by email.