One Meeelllion

Since it started just a bit over 2 months ago, Akismet has now blocked over one million spams! If each spam took people a second to deal with, then we have saved Akismet users 11 and a half days or about 277 hours of time collectively. I never would have expected it to grow this big this fast.

Over 55 thousand blogs are protected by Akismet right now, and as more and more people start taking advantage of the system (and using it for other platforms besides WordPress) the system will only continue to improve and become more effective.

Thanks to everyone who uses Akismet for helping us get this far — here’s to millions more spammers thwarted. :)

Yodel if you Hate Spam

As we just announced over on the WordPress blog, you can now do a “one-click” install of WordPress on Yahoo web hosting and they will even keep it upgraded as new releases come out.

What’s very cool about this deal from an anti-spam point of view is that Yahoo is auto-enabling Akismet for all their customers with a site-wide API key. This means Akismet is just automatically on without having to sign up for a WordPress.com API key, and every Yahoo customer has the equivilent of a Pro-blogger Akimset License. That’s a $5/mo value.

Whoa.

We think this really sets them apart from every other host out there, which invest huge amounts of money into email spam blocking but completely neglect comment and trackback spam, leaving you vulnerable and unprotected. Cheers to Yahoo for leading the pack on this one.

Best Comment Spam

Andy C. claim’s to have found the best comment spam in the world (caught by Akismet of course) but we’re not so sure of that.

Lately I’ve been seeing more pointing back to Dave Winer or Marc Canter’s blog, as well as some clever second-order spammers that use XSS on a third-party site like Sony.com to show their spam message. (That problem seems to have been fixed though.) Any other favorite spams?

Worm Stats

A blog called data mining has been mining the Akismet zeitgeist to find out what we all knew deep down inside: comment spam is wormy.

When the dots get far apart that’s when a comment spam attack is happening.

As the numbers keep rapidly rising (almost 700k spams blocked at the time of this blog post) I’ve been thinking it would be nice if we could turn them around somehow so the statistics aren’t so bleak. Maybe we should have some sort of party when it hits one million? There are already over 40 thousand blogs being protected by Akismet, I’m sure we can collectively think of something fun.

New Development Page

We’ve opened up an Akismet Development page which should be a good place to get started on your holiday hacking adventures, if you are so inclined.

The documentation is still a work in progress, so if you have any ideas or suggestions for it please share!

The development page also lists some of the awesome existing implementations and ports people have done, from Geeklog to Ruby to Python.

Algorithim Updated

The software behind Akismet underwent an upgrade tonight, which you should only notice if spam was getting through the filter before. As always, vote with your submit buttons. If you notice any mistakes mark the comment as spam or ham (not spam) and it will be submitted back to Akismet, which will learn and adapt.

This was the first major upgrade since Akismet launched. I also took the opportunity to upgrade some of the server software to the latest versions, which seems to have performance ever-so-slightly better.

Akismet in Python and API docs

David Lynch has written a cool library for using the Akismet API in Python. I expect over the holidays a few more implementations will come out as people have a little more time for hacking on fun projects. We’ve had a few more development rumblings on the spam-stopper mailing list, so if you have something finished be sure to let me know and I’ll be happy to blog about it.

One that I forgot to blog about earlier was this Spam Karma and Akismet integration. Spam Karma was one of the first plugins for WordPress to take a really comprehensive approach to addressing spam. The only other serious solution is Bad Behaviour, and Akismet complements both perfectly.

Finally the first draft of the API documentation is online, so if you’ve been waiting for that to try out your version be sure to check it out and leave comments if anything isn’t clear.

Update: Right after I published this someone wrote in about another Python implementation! “Mine is implemented differently – as a single class (using urllib2 rather than httplib).” The more choice the better.

Thanksgiving Plugin Update

There was one overwhelming piece of feedback from version 1.0 of the Akismet plugin for WordPress, that people wanted to be able to see the content of the comments marked as spam. This also manifested itself in a lot of “false false positives” being submitted to the system, where a comment would look legitimate except for hidden links or bad content at the end that Akismet would correctly catch but would look like a legit comment from the limited information on the Spam screen.

Well now with version 1.1 of the Akismet plugin (which you can download here) we’ve given it an interface more like the moderation screen in WordPress so you can review the whole comment before deciding to mark it as not spam.

This is an optional update, if you’re happy with the version of Akismet plugin you’re using now feel free to continue with it.

If you have any more great ideas, please contact us and we’ll try to work it in to a future version.