I don't think an analogue to Chilling Effects for harassment/threats would have the desired effect either. That kind of behavior needs to get shut down, not just shamed away. I think behavior that reaches the level of harassment/threats as you've described falls well outside the "fixable behavior" category where you can potentially get the person in question to stop. A repository of resolved cases might help, as a gallery of smoking craters meant to discourage such behavior in others.
I do understand the issue of free speech, and no ISP wants to present the appearance of violating a right that gets far more people riled up. And many of the rights we care deeply about get tested most when the most despicable people take advantage of those rights; hence why so many people hate organizations like the ACLU, and occasionally the EFF. However, as you said, the behavior you mentioned crosses far past the line. Threats actually cross the line from a legal perspective, so it seems like ISPs ought to accept that one pretty readily. And ISPs who have anti-harassment provisions in their terms of service ought to actually enforce them, at least against the less "borderline" cases. The lack of traction you described suggest that neither of those theories actually work in practice, which makes no sense to me.
(For the record, I do actually consider censorship a serious evil, and in general I favor technologies which make it difficult or impossible to censor, primarily because the same mechanisms used to apply censorship in cases I'd agree with can apply it in all cases. Communication tools remain one of the most important means we have of eliminating *other* evils in the world. All that said, as long as we have the ability to track down and stop evil, let's do so. And I really despise the concept that leads to "has more rights than they do"; that really takes the issue too far.)
On a vaguely related note, I wonder if a Bayesian filter could learn to classify harassment the way it classifies spam? And, could it tell the difference between harassment and discussion about harassment? (Ignoring the case of direct quotes.)
no subject
I do understand the issue of free speech, and no ISP wants to present the appearance of violating a right that gets far more people riled up. And many of the rights we care deeply about get tested most when the most despicable people take advantage of those rights; hence why so many people hate organizations like the ACLU, and occasionally the EFF. However, as you said, the behavior you mentioned crosses far past the line. Threats actually cross the line from a legal perspective, so it seems like ISPs ought to accept that one pretty readily. And ISPs who have anti-harassment provisions in their terms of service ought to actually enforce them, at least against the less "borderline" cases. The lack of traction you described suggest that neither of those theories actually work in practice, which makes no sense to me.
(For the record, I do actually consider censorship a serious evil, and in general I favor technologies which make it difficult or impossible to censor, primarily because the same mechanisms used to apply censorship in cases I'd agree with can apply it in all cases. Communication tools remain one of the most important means we have of eliminating *other* evils in the world. All that said, as long as we have the ability to track down and stop evil, let's do so. And I really despise the concept that leads to "has more rights than they do"; that really takes the issue too far.)
On a vaguely related note, I wonder if a Bayesian filter could learn to classify harassment the way it classifies spam? And, could it tell the difference between harassment and discussion about harassment? (Ignoring the case of direct quotes.)