Tuesday, April 12, 2011

What's the opposite of porn?

I got together with a group of Xooglers recently and heard a story that was new to me. I found it pretty amusing.

At one point Google was refining its porn filter and and came up with a list of terms that gave a strong indication that a web page contained sexually explicit material. If these words appeared in any combination, the page was likely full of adult content and should be screened out.

There was one word on the list however, that had a very high negative correlation. That is, if this word appeared on a page in combination with a known porn flag, the page should still be considered safe to view for everyone.

That word was "county."

I didn't understand why this would be true until one of the Xooglers explained. Without making "county" a negative indicator, he told me, Google users would have a very hard time finding information about Beaver County, Pennsylvania.

Google has likely developed far more sophisticated filtering techniques since then, but I won't be surprised to hear that adult sites are adding the word "county" to their pages as a good luck talisman for the next few months. If it actually works, I owe spam-meister Matt Cutts an apology and a beer -- perhaps a nice I.Porter.A.

2 comments:

Geoff said...

Rather like the residents of Scunthorpe here in England been black listed :-)

Unknown said...

I was the receptionist when Jim and Larry decided to run this for the first time at this (I believe it was our first venture capital meeting) gathering of all these important men in suits. I assumed important because of the suits. Sergey was actually wearing pants that day instead of shorts. Regardless of what they wore, the words that started streaming down the computer screen *immediately* turned me a very very deep red and I'm pretty sure they were told to turn it off. I still have nightmares on a search I saw....