[thetabyte]

Especially after pg's response to the blog post about sexual assault at CodeMash, I wonder—why can't the flamewar detector just disable comments?

I tend to have a lot of respect for pg, and found his apology for what happened in that thread to be admirable. Whether or not preventing discussion of the issue on HN is positive or negative...I have very complex feelings on the issue, and see valid arguments on both sides.

What I do not have mixed feelings about, however, is that these issues need to be put front and center, so that people in our industry a) know they exist b) know how common they are c) are inspired to make personal effort to fix it. I would hope that pg agrees.

If he does, why not make such stories, when they set off the flamewar detector, maintain their ranking, but disable comments? That way, the issue is still raised, and people are still alerted to it, but it prevents the (some believe) "unproductive" discussion.

[jlgreco]

> I wonder—why can't the flamewar detector just disable comments?

In some cases it seems to, though by that time the discussion is long off the front page. I think this is a good thing, if something is on the front page then comments should be enabled, if only in case something in the article desperately needs to be corrected. If we get a post on here about a new study suggesting that vaccines may cause autism, that would almost certainly generate a flamewar but the absolute last thing we would want is for the post to remain on the front page and not permit anyone to post comments that may refute claims made by the study. In situations like those, it is better to kill the discussion and to take the post off the frontpage than to leave it there but disable commenting.

[pg]

The output of the flamewar detector isn't binary.

[Zak]

If the flamewar detector takes content in to account, could it consider the overall tone of the discussion and the content of each new comment and selectively kill new comments that seem flame-like?

I wrote a moderation bot to use on a political subreddit that used a text classifier to determine whether comments should be deleted. A goal was to delete flame comments rapidly. It worked, but reddit's reply notifications limited its effectiveness.

Of note: the phrase "you are a" ranked higher for the flame category than any particular insult.

[thetabyte]

Ahh, that makes sense. Can there be a sliding scale of delay for the ability to comment, or depth of thread allowed (before commenting is disabled for that thread), based on the output of the flamewar detector?

This may also have the side effect of allowing for some opinions to be shared, without being mired in flame-y back and forth. (And, importantly, still allows the story to stay more visible)