3 ms·
The problem is that it's far more trivial to bypass the filters than it is to make the filters work well and not have false positives. Consider: sh1t, cu|\|7,
by InclinedPlane 13y ago
The problem is that it's far more trivial to bypass the filters than it is to make the filters work well and not have false positives.
Consider: sh1t, cu|\|7, a55, f00ker, and so on and so on.
Plus you have to constantly update to take into account new slang which can have double meaning. Some kid might say "she's just a thirsty ratchet" and there's no way in hell that a simplistic automated system is going to be able to flag that as being a use of expletives or filthy and derogatory language.
- greglindahl 13y agoI wasn't proposing fixing any of those problems, just the Scunthorpe problem! You are correct that there are more problems out there.
- vacri 13y agoI find it humourous looking at some oldschool grep commands that grep for 'username' and 'password'... but the first letter might be capitalised, so you grep for 'sername' and... 'ssword'. Because 'assword' is too rude for code, apparently...
- Datsundere 13y agoyou forgot "the N-word" they say in the media. It's white people getting away with saying the n-word.
- timje1 13y agoThis joke's got recursion![0] [0] https://news.ycombinator.com/item?id=6935041 https://news.ycombinator.com/item?id=6935041