Lately I have been thinking about the subtleties/psychology that lives between-the-lines in answers and comments.

There are algorithms to try and warn posters of subjective questions, as well as a few token mechanical filters in comments for "problem cases"...such as preventing people from linking to "let me Google that for you" (it rarely represents a sincere attempt to be helpful, but prohibiting me from linking it to illustrate a point in this question would be foolish...yes?)

My hypothesis is that there should be no mechanical censorship, only warnings you may disregard (at risk of moderation). Further...I believe there would be broad impacts on the success of StackExchange if there was a warning system--similar to what runs against question titles--which checked your comments and answers prior to posting. If potential problems were detected you'd get an "Are you Sure?" with links to explanations of the issues.

The warning filter might start off simple enough. Basic example: it could look for words like "stupid". But the idea would be to imagine the warning step as ideally passing through someone who knew the person you were responding to really well, and knew their triggers. (Again: This shouldn't stop you from posting what you want to post, just to raise awareness...the question title warning works in the same way.)

I'm only suggesting the introduction of this step, and the importance of being able to ignore its feedback (even in the lmgtfy scenario). I bet some simple heuristics could remove roadblocks to communication.


1 Answer 1


While in general I agree with the sentiment, I think it will be incredibly difficult to implement, and has very little chance of actually having an impact IMHO. Warnings can't teach people to be polite or prevent them from being condescending, just like - and let's take some extreme examples - jail for grand theft auto isn't enough of a deterrent to have these idiots get caught on Bait Car every single week, or alleged child molesters continue to get busted on "To Catch a Predator" even though they are well aware of the consequences of their actions, that the behavior is not condoned by the community or by society in general, and that these sting operations continue to take place. In much the same way, people who talk down to others here are well aware that talking down is not condoned, and since there is no consequence to doing so (and wouldn't be in your proposal either, except an extra click), they have even less motivation to change than the car thieves or statutory rapists.

What I don't like about your suggestion goes back to the difficulty in implementation. We're going to start banning words, or making them taboo? So if I want to say "Microsoft made a stupid decision when they did x" I'm going to have to accept and ignore some warning?

If someone is behaving like a jerk, it isn't specific words that are causing a roadblock, but an overall attitude. I doubt there are many scenarios where such people aren't aware of it, in such a way that this warning will make them suddenly aware - learning to not be a jerk is quite different from learning to not include a tag or the word "question" in a title. As an example, I am quite capable of calling people stupid without using the word stupid.

  • Maybe the "are you sure" was a bit too strong. (The question filter doesn't do that.) My central goal here, like with the question title filter, is NOT to fix people who are purposefully trying to be antagonistic. It's rather to offer a "teachable moment" with some links/information for good-intentioned people who might not have thought through their words, but would if they had a hint available. It would be nice if the filter were smart enough to discern "doing what you suggest would be stupid" vs. "microsoft made a stupid decision", but even in the interim I think heuristics could help. Commented Jul 14, 2012 at 16:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .