Hearthpwn

This may work if you already have the rude/polite flags up for people. The problem is getting the flags up in the first place.

For example, if a person is being rude then alnog with the rudeness, they are going to go flag you, assuming this is day 1 of the patch and, thus, no one has been marked 'rude' yet.  Given that this person is already going out of the way to cuss you out an additional step to mark you 'rude' doesn't seem rather silly.  

As such you could potentially be marked as 'rude' by every rude person you meet while they get marked back by yourself.  If, say, after 10 people you get the rude flag, that potentially means after 10 cussings you'll be automarked as rude by the system.  Those rude people will eventually be marked as 'rude' as well of course.

But eventually you get a case ofa whole lot of innocent folks being dragged down to 'rude hell' not being able to talk to 'polite' folks or block the 'rude' folks.  

 

That's just one key problem that many companies need to get a handle of: automation doesn't work for stuff like this.  Any time you make an automated rule based system with set results you'll result in exceptions that fall between the cracks and the ability to manipulate the system.  A great system can probably handle 60-80% of cases successfully but you need something flexible like a human brain reviewing the situation, to avoid the false flags and dirty tricksters.  But if we had that then a simple report system would work out with a quick reviewer to catch the oddballs.  

To be honest, I think the systems that work in MMOs and other games will probably be best: good tools to find friends and similarly minded folks so that your brain can decide who you spend your time with. 

Though if I WOULD add something new I'd add a filter...a RECEIVING filter.  As in NOT a filter on what YOU can say but a filter on what you receive.  It doesn't affect the sender: they get to send whatever they want at you and get no feedback against it.  BUT if you, say, set the filter to "strict" then a typical public chat request would be "you **** **** ** *** ** and ***".  Could even have a HYPERSTRICT setting where any sentence with a rude word gets flat out eliminated, so you'd see them send a blank message.  Again, to them they sent the message properly and get nothing telling them otherwise, but you get the filtered version.

It's a similar mentality to why squelching emotes work.  It's the least freedom stomping while also offering the least amount of troll enabling (they get to say what they want but won't realize you aren't bothered by it).