February 11, 2004
The Idiocy Of Filtering Software

I don't know what brand internet filter my school uses, but it seems to do a very poor job, with numerous false positives and some false negatives, too. I didn't have anything to do during 7th period today, so I checked all the links on my blogroll, and no fewer than 20 were declared, in big red letters, to be "pornography". I hope the computer department isn't keeping track of my browsing. These sites flunked whatever test was applied:

  1. Anticipatory Retaliation
  2. RiShawn Biddle
  3. 2Blowhards
  4. Carraig Daire
  5. Cathy's World
  6. Classless Warfare
  7. Colby Cosh
  8. Cronaca
  9. Dustbury
  10. Electric Venom
  11. God of the Machine
  12. HobbsOnline (too much JadeGold?)
  13. Horologium (ditto)
  14. Ipse Dixit
  15. Isntapundit (he does mention "anal warts" in his most recent post -- 1/27)
  16. Reflections in D Minor
  17. A Small Victory
  18. The Sound and Fury
  19. Andrew Sullivan (too much gay stuff?)
  20. VodkaPundit (for the name, I assume)

I see nothing particularly offensive about any of them, though some (e.g. Cosh) use a fair number of four-letter words. What is particularly striking is the high percentage of relatively apolitical culture blogs (2Blowhards, Cronaca, God of the Machine, Reflections in D Minor) that are banned. Otherwise, there seems to be no pattern at all.

As for false negatives, the two sites on my blogroll that most frequently display pictures unsuitable for Catholic high school students were not blocked: I mean Curmudgeonly & Skeptical and Gweilo Diaries. All in all, the software might as well have flipped a coin, or rather plugged into www.random.org.

Of course, in some cases, the name of the site may have been the problem. Still, I would think a good filter would be able to tell that a site whose name includes the single word "blowhard" is far less likely to be pornographic than one where the "blow" and the "hard" are separate words. In some cases the problem may be words that aren't in any dictionary. If the software doesn't find 'horologium' in its internal word-list, does it assume that a name starting with 'ho' must be an obscentiy? Or does somebody somewhere think that a 'horologium' is something like an aquarium, only for ho's instead of fishes? Does someone think a 'cronaca' is some kind of marital aid or obscure anatomical term? And how did Buzzmachine make the cut? It sounds like a slang term for a vibrator to me. And Feces Flinging Monkey is also unblocked, despite its name.

Posted by Dr. Weevil at February 11, 2004 10:29 PM
Comments

They didn't ban me ... blast, I'm going to have to swipe more half-naked asian girls from Conrad.

Posted by: Robin Roberts on February 12, 2004 12:53 AM

If you can find out, I'd like to know what the filtering criteria are. I don't do pr0n, and the language on my site is generally no worse than PG-rated, although sometimes a bit vituperative. In fact, I make every effort to block pr0n references, as I don't like to show up in perverse search results.

The only thing I can think of that might cause me to be blocked is the recent discussion of gay marriage (and the links to two of my previous posts). I have no idea what would cause Lynn's site (or the other culture blogs) to get banned.

Posted by: timekeeper on February 12, 2004 10:59 AM

~sigh~ I guess I shouldn't have mentioned Janet Jackson's boobie.

On Amazon.com's books discussion forum (now closed) you couldn't talk about Philip K. Dick (which is sort of understandable) or Charles Dickens. (ridiculous!)

Posted by: Lynn S on February 12, 2004 01:29 PM

Yeah, the filters they have set up on our intranet at work don't let me go to Vodkapundit either. I haven't tried any of the others, because I really don't have time to play around on the internet, but -- I wondered what it is the bosses are trying to keep people from doing. "There will be no reading about drinking alcohol on company time!"

Posted by: Andrea Harris on February 12, 2004 09:42 PM

Hah!
I'm clean!

Posted by: Michael Tinkler on February 12, 2004 09:50 PM

In VP's case, it may be due to the inordinate amount of times he referenced nipples and boobies after the Janet Jackson show. (extrapolating from Lynn's comment above)

Posted by: timekeeper on February 13, 2004 07:45 AM

I can't post off-color pictures on my blog -- I've been informed my mother-in-law reads it.

Posted by: McGehee on February 13, 2004 10:47 AM

Years ago some women wanted to set up an AOL Breast Cancer chat group. AOLs software wouldn't allow it. They ended up with a Boob Cancer group.

Posted by: Dan on February 13, 2004 05:29 PM

I guess the only appropriate response to my inclusion would be, "Heh, heh. He said 'Dixit'!"

Posted by: Dodd on February 13, 2004 07:28 PM

Kevin,

does this you removed the Olson twins googlebomb? (grin)

Posted by: timekeeper on February 14, 2004 04:28 AM

My entire domain is banned at my workplace. I have no idea what caused it.

Posted by: Ricky on February 16, 2004 09:09 PM

To be fair, on my page (Anticipatory Retaliation) I did post the infamous list of 213 things that SPC Schwarz was forbidden to do. The list features such gems as:

#167: Not allowed to operate a business out of the barracks.
#168:Especially not a pornographic movie studio.
#169:Not even if they *are* 'especially patriotic films'

#200:My chain of command is not interested in why I 'just happen' to have a kilt, an inflatable sheep, and a box of rubber bands in the back of my car.

And so on. Tasteless? Perhaps. Pornographic? Not so much.

Posted by: Bravo Romeo Delta on February 17, 2004 04:59 PM

They didn't ban ME...
What good is software that won't ban a vicious right wing site?

LOL./..

Posted by: Bithead on February 23, 2004 10:56 PM