Another milestone

« previous post | next post »

At some point around lunch time today, our filter nabbed its 2 millionth spam comment:

Since we logged the our millionth spam comment at some point on 9/1/2011, 152 days ago, we've been averaging about 1000000/152 = 6579 spam comments per day.

You'll understand, then, why I need to flush the spam trap periodically, without checking everything that has lodged there, even though I know that some small fraction of the total has been wrongly classified. But I do peek at the scum-covered surface from time to time, and I can tell you that the spamularity seems as far away as ever.

Unless, of course, all the comments like these are simply a smokescreen

Finally, an issue that I am passionate about. I have looked for information of this caliber for the last several hours. Your site is greatly appreciated.
I had to tell you this is a really great website, wonderful theme and it makes a change to see such a great page.
2HD68n Hi! Everyone who reads this blog – Happy Reconciliation and Accord..!
ygoblG Yet, much is unclear. Could you describe in more details!…
aVQt8q Yeah, it is clear now !… Just can not figure out how often do you update your blog?!…
Puppy mating is really a tricky subject matter ! the decision to type puppies is one thing that really must be well-planned along with completely regarded. There is a great deal which retreats into it, and also the …{Yankees dog collar|portable pet crate|Yellow dog collars|big dog cages|Anti barking dog collars|wooden dog kennels}
May ribbon Lingual, the so-called dog translator, truly translate your own dog's barks and also woofs into language discover about this the following.{Yankees dog collar|portable pet crate|Yellow dog collars|big dog cages|Anti barking dog collars|wooden dog kennels}
rEtirement is always that most crucial juncture within the lifetime of a practical individual through in which his life requires a wholly different study course, specifically in monetary concerns

…while the real AI programs succeed in fooling us all. To what nefarious end, we are apparently unable to imagine.



20 Comments

  1. neuromusic said,

    January 31, 2012 @ 2:54 pm

    I'm reminded of the XKCD take on the "Constructive" Spamularity: http://xkcd.com/810/

  2. Nathan said,

    January 31, 2012 @ 3:10 pm

    I finally passed a Turing test–w00t!

  3. Geoffrey K. Pullum said,

    January 31, 2012 @ 3:33 pm

    I'm suddenly worried that we can't tell whether the post above was really by Mark Liberman. What if the spam robots started writing Language Log, and everyone thought it was us, only they were wrong?

  4. Rube said,

    January 31, 2012 @ 3:37 pm

    And suppose it all took place in a Chinese room…

  5. David J. Peterson said,

    January 31, 2012 @ 3:59 pm

    If I may make a recommendation, the plugin Spam Free WordPress blocks 100% of automated spam comments to WordPress sites. To see an example of how it works, go to the comment form on my Dothraki blog here. So far the automated spammers haven't found a way around it. Since installing it on five blogs, I've only gotten one spam comment, and it was entered manually. It made my day knowing that they had to put in a little manual effort just to leave a link to a site selling genuine knock-off handbags.

  6. D.O. said,

    January 31, 2012 @ 4:02 pm

    Prof. Pullum, computers are unable to rage against misidentification of passives, I am sure of that.

  7. Rubrick said,

    January 31, 2012 @ 5:15 pm

    Spam or no, I think we can all agree that the decision to type puppies is one thing that really must be well-planned along with completely regarded.

  8. Amy West said,

    January 31, 2012 @ 7:17 pm

    I had never known that such a thing as spam comments even existed until I saw my husband going through the spam filter queue for his blog a couple months ago and read some to me. What a sheltered life I lead! However, it looks to me that many of these spammers use many of the same rhetorical techniques as many freshmen do to write a 500-word essay that says nothing.

  9. Colin said,

    January 31, 2012 @ 7:43 pm

    woof. translate that! woof.

  10. Faldone said,

    January 31, 2012 @ 7:56 pm

    @David J. Peterson: I think the problem is not with a spam filter's ability to filter out true positives, rather it's the filter's tendency to filter out false positives. Geoff Pullum's spam filter is 100% effective at filtering out spam comments.

  11. Trey Jones said,

    January 31, 2012 @ 8:08 pm

    Finally, an issue that I am passive about. Information of this caliber has been looked for for the last several hours. Your site is greatly appreciated. That this is a really great website had to be said; wonderful theme and to see such a great page makes a great change. Misuse of passives and "passives" makes me "angry". Puppies!

    The new SpecGram SpamBot 2000 is up and running!

  12. DMajor said,

    February 1, 2012 @ 2:14 am

    Fwiw, regarding "May ribbon Lingual, the so-called dog translator" — there really is a product named "Bow-Lingual" which purports to do just that, translate dog barks and whines for humans. The inventors received an Ig Nobel prize in 2002.

    The "bow" in "Bow-Lingual" is supposed to be for "bow-wow", but who-or-whatever translated that spam message must have confused it for an ornamental bow, a tied-up ribbon.

  13. Ginger Yellow said,

    February 1, 2012 @ 6:08 am

    May ribbon Lingual, the so-called dog translator, truly translate your own dog's barks and also woofs into language discover about this the following.{

    Maybe they should try to master human language translation first.

  14. Stan said,

    February 1, 2012 @ 7:51 am

    My favourite recent spam comment: "Hrm, Not the best post unfortunately. Sorry to be so blunt! You should try some Norwegian carrot cake to cheer you up."

    And a phrase from one I liked a couple of months ago: "both boys and girls feel the impact of just a moment's pleasure, for the rest of their lives".

  15. Frank said,

    February 1, 2012 @ 11:50 am

    I say we celebrate this milestone with some refreshments! I need everyone to chip in $10,000 to an escrow account I've set up for the festivities.

  16. Rube said,

    February 1, 2012 @ 12:40 pm

    @Frank:

    That's a capital idea. I happen to have a money order for $15,000. I will send it to you, and you can have the $10,000, as long as you promptly send me the surplus $5,000 by Western Union.

    Thanking you in advance for your most valued attention to this matter.

  17. Adam said,

    February 1, 2012 @ 4:31 pm

    @Faldone
    Geoff Pullum's spam filter is 100% effective at filtering out spam comments.

    Sure, and a similar filter would be 100% effective at filtering out bad Language Log articles.

  18. Adam said,

    February 1, 2012 @ 4:33 pm

    I'm suddenly worried that we can't tell whether the post above was really by Mark Liberman. What if the spam robots started writing Language Log, and everyone thought it was us, only they were wrong?

    Hmm, GKP could his approach to comments further, and just not let anyone but himself post LL articles.

  19. Sili said,

    February 2, 2012 @ 12:46 pm

    Hmm, GKP could his approach to comments further, and just not let anyone but himself post LL articles.

    Hmmm? It seems you accidentally the entire verb.

  20. Adam said,

    February 3, 2012 @ 4:54 am

    @Sill

    Oops, Skitt's law got me. I think the missing verb was "extend".

RSS feed for comments on this post