9 comments

[ 2.8 ms ] story [ 16.7 ms ] thread
A quick skim reveals many words that are plainly innocuous in all but the most devious of contexts.
That may be true of words like "deposit", but I can't even come up with a dirty use for "k mart".

There are also a fair number of misspellings (which might make sense to catch intentional avoidance), but the weird part is the correct spelling is quite often missing - for example, several variations of masturbate, but not the actual word itself.

I was performing an assessment of a content filtering appliance once and we obtained the word list of terms that it filtered.

I'll never forget that one of them was "vertical bacon sandwich", which to this day I still think they made up (I've never heard it used in any context ever). We laughed about it for ages.

Maybe it was a test case that they used to spot check the filter.
Nice to see we have "juggalo" covered. Somewhere, Violent J is smiling.
"oui" is on there... wtf?
The list appears to have been copied directly from a list of words which the NFL Shop banned from appearing on customized jerseys. So directly, in fact, that the Pakistani list has a gap of three words between "UPTHEBUTT" and "URINATE" (1071 and 1075), right where the NFL list has a few blank lines:

http://outsports.com/nfl/2005/0301nflshopnaughtywords.htm

That probably explains most of the odd choices. :)

You can't text 'creamy' in Pakistan? Is this for real? It seems suspect.
Can someone explain the redundancies to me? If you can't text rot13("shpx") then why do they need to block rot13("shpx") + " you"? Or does that imply that rot13("shpx") + " u" would be accepted (it's not on the list)? Or did someone just not understand how longer strings match faster?