5 ms·

The Hacker News crowd and the tech entrepreneurship ecosystem in general has a deserved reputation for a bias toward trying to fix problems that are prominent i

The Hacker News crowd and the tech entrepreneurship ecosystem in general has a deserved reputation for a bias toward trying to fix problems that are prominent in their own line of sight. In other words, a lot of startups trying to fix problems that vex prosperous, tech-savvy single people in San Francisco. But here's a geek-world problem that's staring us all right in the face. Despite the enormous potential benefit of the internet as a place for open, respectful discourse among people from all walks of life, it seems that every forum must eventually devolve into ad-hominem rancor and trolling, unless it's aggressively and time-consumingly moderated by dedicated human beings. I've been involved in trying to crack this particular nut since 1997 as publisher of OSNews, and I've been a participant in some relatively good forums such as HN over the years, but I can't help but think that there's an opportunity start with what Disqus has done and take it to the next level. In the article, Coates bemoans the limited moderation toolset that he got from Disqus, and I can't help but think that there's a potential big data play to fingerprint the most obvious kinds of trolling and give human moderators a quiver of time-saving tools to keep their forums on the straight and narrow. If anyone here has some ideas about how we might go about executing on this idea, let me know. I'd be interested in being involved.

panic 12y ago
If you haven't seen it already, you should watch Matt Haughey's "Lessons from 11 years of community" (link below), in which he describes some of the moderation tools and techniques MetaFilter has developed over the years. "Lessons from 11 years of community": https://vimeo.com/21043675 https://vimeo.com/21043675
- LeoPanthera 12y ago
  The problem with MetaFilter's moderation is that it is both heavy-handed and entirely opaque. If they decide a comment is unacceptable, for whatever reason, it is simply erased from history. The poster is not informed this has happened, nor told why. There are two enormous disadvantages to this model - it doesn't encourage commenters to improve the quality of what they post, and it ultimately creates an atmosphere of mistrust, from those who have been silently black-holed. The Reddit model of public votes is better, even when layered with admin moderation, because at least users can see what is happening, and why.
  - A_COMPUTER 12y ago
    In the past on Metafilter, if there were some unacceptable comments they were deleted but the moderator would add a comment that simply restated what was not allowed and asking people to stay civil. I think this can actually work really well, because it doesn't shame the people moderated and they can move forward without feeling personally slighted. This is in stark contrast to the person's name next to a big [DELETED] or Boingboing's mocking "disemvowelment".
    - wiml 12y ago
      I have to agree that one of the things that made Metafilter's moderation more successful was being able to moderate without alienating the users. Moderation often devolves into simply ejecting all but one "tribe" from a forum so that nobody's left to have a disagreement; MeFi avoided that trap far better than many places.
      - A_COMPUTER 12y ago
        Unfortunately I had to qualify with "in the past." I think it's an echo chamber today. I have heard more and more people talk about unfair moderation, which I never saw years ago. So invisible moderation does still have a downside. To be fair, I think most of the people still posting there are like old friends now.
ddw 12y ago
Mozilla OpenNews, NY Times and Wash Po are partnering to create a new solution. I'm not sure if its exactly what you mention here. https://blog.mozilla.org/blog/2014/06/19/mozilla-announces-a-partnership-with-the-new-york-times-the-washington-post-and-knight-foundation-for-a-new-online-community-platform/ https://blog.mozilla.org/blog/2014/06/19/mozilla-announces-a...
birdsareweird 12y ago
I can't help but see your comment as a symptom of the problem it's describing. Not trolling, but geek bias... specifically the blind spot of assuming everyone is as sincere as you. The big problem with comment sections isn't "trolling", it's that most people aren't there to have a fair, rational discussion in the first place. They are there to be entertained or distracted, that will be the main motivation of people doing the browsing. They want to see their opinions and choices validated, while uncomfortable and inconvenient facts are hidden or mocked. At minimum they will avoid anything that reflects badly on themselves and where they can't shift the blame easily. See for example the latest wave of "criticism = harassment" in news sections and on Twitter, where being called out for saying stupid things in public is frantically derided as "sea lioning" which must be stopped. So trying to fix comment sections behaviorally does not work because they stop being comments. At most you can do something like StackOverflow, which follows a strict Q&A format and bans repetition. It's not really a forum, it's a wiki of googleable questions, and like Wikipedia, is mostly unappealing and unwelcoming to outsiders with fresh ideas. Besides, the appropriate answer is not to censor and hide, but to confront and refute. You want bad ideas to be catalogued and paired with their antidote, not let people remain blissfully unexposed, their mental immune system ready to overreact to the slightest provocation. The only forums I know of where everyone gets along with everyone are echo chambers. In real discussions somebody's toes always get stepped on. Real names do not help there btw, people still make plenty of dumb comments with their face next to it, it only helps the offendees play the victim card. I think your "inevitable descent into trolling" is a bad description of how intellectual forums grow their audience: they become more superficial, more emotionally driven and more focused on presentation than merit. Like TED talks. But if you don't grow, you stagnate and turn into an echo chamber. Neither is desirable, you can only try to grow responsibly. You have to allow disruptive voices no matter what.
- ansible 12y ago
  Besides, the appropriate answer is not to censor and hide, but to confront and refute. You want bad ideas to be catalogued and paired with their antidote, not let people remain blissfully unexposed, their mental immune system ready to overreact to the slightest provocation. That is some of the ideas behind Marc Stiegler's "decision duel" idea in his book "David's Sling" [1]. He envisioned a system where two sides would lay out their arguments, and then the two duelists would link in arguments and counter-arguments to each point. This duel would occur with an audience watching, who would also contribute ideas and research to the duelists. It is actually the right time to be able to implement something like this now, with the advent of WebGL and other browser technology for the presentation. I was just looking at another HN link about the 'wikigalaxy' visualization system which just hints at what is now possible. [1] http://www.goodreads.com/book/show/3064877-david-s-sling http://www.goodreads.com/book/show/3064877-david-s-sling
A_COMPUTER 12y ago
I think the value proposition of Disqus is just comparatively low maintenance. You include a snippet of JS and you get a minimally functional commenting system that includes antispam and requires running no backend server software or security risks. In return Disqus gets analytics and a little bit of ad space. It's not imo going for discussion, it's meant to be flypaper for readers who get a good-feelings boost thinking they're being heard, so it's a monetization tool more than anything. The extreme end of this is Youtube comments, which accumulate so fast and so voluminously that actual discussion is near impossible, and the limited nesting actively fights replying. You can't follow anything. But millions of people pile their worthless comments into the void because it feels good to have your say. Old, crusty ugly forums are better for conversation than new, shiny JS comment systems because they were actually made for conversation, that was their raison d'etre. Have you tried Discourse? It's almost as good as the hype. So this is kind of related. I've been reading OSNews since the very beginning, and I really, really love it. It used to beat out Slashdot for me, which is about the highest praise I could give a site. In large part it's because it covered stuff I couldn't see anywhere else and the commenters were positive and knowledgeable. But in the last few years there are so many articles about non-OS political stuff, which in turn causes fighting in the comments. Don't underestimate that the types of articles published determines the type of commenters :-(
- smcnally 12y ago
  > It used to beat out Slashdot for me, which is about the highest praise I could give a site. Slashdot's meta-moderation contributed to discussion quality and high signal:noise. I'm still skeptical of parent comment's fingerprinting as a sole solution -- it takes humInt and humInvolvement. Arm willing human mods with tools like fingerprinting, provide incentives for many to moderate and for some to moderate the mods, and good discussions can happen.
yummyfajitas 12y ago
The difficulty is that the "trolling" folks like Coates want to eliminate is often simply a difference of opinion which is acceptable in one community but not in another. One of the few times I read Coates' blog was discussing a black/white gap in dating site response rates. He immediately banned people pointing out statistics showing that black women are disproportionately fat. The HN discussion on the same topic was heavily upvoted and rationally concluded concluded that explanation was probably incorrect. Building a system which will automatically detect that, e.g., statistics about race and obesity are trolling on Coates' site but useful contributions on HN is a difficult matter. Collaborative filtering is the best way to create a bubble, but that generally requires a centralized communication medium like Facebook, Twitter or Reddit.
bdr 12y ago
I wonder if rather than a pure software solution, there's an opportunity here for Disqus to provide a hybrid software-human moderation solution. It's clear that good moderation can revitalize a community (hi @dang), while bad comments are a vicious cycle that drive out the good ones. IOW, moderation is very valuable. But it's hard to hire for and do right if it's not already part of your org's thinking. Disqus could sell moderation-as-a-service, according to the customer's parameters, and take care of all those hard parts.

panic 12y ago

If you haven't seen it already, you should watch Matt Haughey's "Lessons from 11 years of community" (link below), in which he describes some of the moderation tools and techniques MetaFilter has developed over the years. "Lessons from 11 years of community": https://vimeo.com/21043675 https://vimeo.com/21043675

LeoPanthera 12y ago
The problem with MetaFilter's moderation is that it is both heavy-handed and entirely opaque. If they decide a comment is unacceptable, for whatever reason, it is simply erased from history. The poster is not informed this has happened, nor told why. There are two enormous disadvantages to this model - it doesn't encourage commenters to improve the quality of what they post, and it ultimately creates an atmosphere of mistrust, from those who have been silently black-holed. The Reddit model of public votes is better, even when layered with admin moderation, because at least users can see what is happening, and why.
- A_COMPUTER 12y ago
  In the past on Metafilter, if there were some unacceptable comments they were deleted but the moderator would add a comment that simply restated what was not allowed and asking people to stay civil. I think this can actually work really well, because it doesn't shame the people moderated and they can move forward without feeling personally slighted. This is in stark contrast to the person's name next to a big [DELETED] or Boingboing's mocking "disemvowelment".
  - wiml 12y ago
    I have to agree that one of the things that made Metafilter's moderation more successful was being able to moderate without alienating the users. Moderation often devolves into simply ejecting all but one "tribe" from a forum so that nobody's left to have a disagreement; MeFi avoided that trap far better than many places.
    - A_COMPUTER 12y ago
      Unfortunately I had to qualify with "in the past." I think it's an echo chamber today. I have heard more and more people talk about unfair moderation, which I never saw years ago. So invisible moderation does still have a downside. To be fair, I think most of the people still posting there are like old friends now.

ddw 12y ago

Mozilla OpenNews, NY Times and Wash Po are partnering to create a new solution. I'm not sure if its exactly what you mention here. https://blog.mozilla.org/blog/2014/06/19/mozilla-announces-a-partnership-with-the-new-york-times-the-washington-post-and-knight-foundation-for-a-new-online-community-platform/ https://blog.mozilla.org/blog/2014/06/19/mozilla-announces-a...

birdsareweird 12y ago

I can't help but see your comment as a symptom of the problem it's describing. Not trolling, but geek bias... specifically the blind spot of assuming everyone is as sincere as you. The big problem with comment sections isn't "trolling", it's that most people aren't there to have a fair, rational discussion in the first place. They are there to be entertained or distracted, that will be the main motivation of people doing the browsing. They want to see their opinions and choices validated, while uncomfortable and inconvenient facts are hidden or mocked. At minimum they will avoid anything that reflects badly on themselves and where they can't shift the blame easily. See for example the latest wave of "criticism = harassment" in news sections and on Twitter, where being called out for saying stupid things in public is frantically derided as "sea lioning" which must be stopped. So trying to fix comment sections behaviorally does not work because they stop being comments. At most you can do something like StackOverflow, which follows a strict Q&A format and bans repetition. It's not really a forum, it's a wiki of googleable questions, and like Wikipedia, is mostly unappealing and unwelcoming to outsiders with fresh ideas. Besides, the appropriate answer is not to censor and hide, but to confront and refute. You want bad ideas to be catalogued and paired with their antidote, not let people remain blissfully unexposed, their mental immune system ready to overreact to the slightest provocation. The only forums I know of where everyone gets along with everyone are echo chambers. In real discussions somebody's toes always get stepped on. Real names do not help there btw, people still make plenty of dumb comments with their face next to it, it only helps the offendees play the victim card. I think your "inevitable descent into trolling" is a bad description of how intellectual forums grow their audience: they become more superficial, more emotionally driven and more focused on presentation than merit. Like TED talks. But if you don't grow, you stagnate and turn into an echo chamber. Neither is desirable, you can only try to grow responsibly. You have to allow disruptive voices no matter what.

ansible 12y ago
Besides, the appropriate answer is not to censor and hide, but to confront and refute. You want bad ideas to be catalogued and paired with their antidote, not let people remain blissfully unexposed, their mental immune system ready to overreact to the slightest provocation. That is some of the ideas behind Marc Stiegler's "decision duel" idea in his book "David's Sling" [1]. He envisioned a system where two sides would lay out their arguments, and then the two duelists would link in arguments and counter-arguments to each point. This duel would occur with an audience watching, who would also contribute ideas and research to the duelists. It is actually the right time to be able to implement something like this now, with the advent of WebGL and other browser technology for the presentation. I was just looking at another HN link about the 'wikigalaxy' visualization system which just hints at what is now possible. [1] http://www.goodreads.com/book/show/3064877-david-s-sling http://www.goodreads.com/book/show/3064877-david-s-sling

A_COMPUTER 12y ago

I think the value proposition of Disqus is just comparatively low maintenance. You include a snippet of JS and you get a minimally functional commenting system that includes antispam and requires running no backend server software or security risks. In return Disqus gets analytics and a little bit of ad space. It's not imo going for discussion, it's meant to be flypaper for readers who get a good-feelings boost thinking they're being heard, so it's a monetization tool more than anything. The extreme end of this is Youtube comments, which accumulate so fast and so voluminously that actual discussion is near impossible, and the limited nesting actively fights replying. You can't follow anything. But millions of people pile their worthless comments into the void because it feels good to have your say. Old, crusty ugly forums are better for conversation than new, shiny JS comment systems because they were actually made for conversation, that was their raison d'etre. Have you tried Discourse? It's almost as good as the hype. So this is kind of related. I've been reading OSNews since the very beginning, and I really, really love it. It used to beat out Slashdot for me, which is about the highest praise I could give a site. In large part it's because it covered stuff I couldn't see anywhere else and the commenters were positive and knowledgeable. But in the last few years there are so many articles about non-OS political stuff, which in turn causes fighting in the comments. Don't underestimate that the types of articles published determines the type of commenters :-(

smcnally 12y ago
> It used to beat out Slashdot for me, which is about the highest praise I could give a site. Slashdot's meta-moderation contributed to discussion quality and high signal:noise. I'm still skeptical of parent comment's fingerprinting as a sole solution -- it takes humInt and humInvolvement. Arm willing human mods with tools like fingerprinting, provide incentives for many to moderate and for some to moderate the mods, and good discussions can happen.

yummyfajitas 12y ago

The difficulty is that the "trolling" folks like Coates want to eliminate is often simply a difference of opinion which is acceptable in one community but not in another. One of the few times I read Coates' blog was discussing a black/white gap in dating site response rates. He immediately banned people pointing out statistics showing that black women are disproportionately fat. The HN discussion on the same topic was heavily upvoted and rationally concluded concluded that explanation was probably incorrect. Building a system which will automatically detect that, e.g., statistics about race and obesity are trolling on Coates' site but useful contributions on HN is a difficult matter. Collaborative filtering is the best way to create a bubble, but that generally requires a centralized communication medium like Facebook, Twitter or Reddit.

bdr 12y ago

I wonder if rather than a pure software solution, there's an opportunity here for Disqus to provide a hybrid software-human moderation solution. It's clear that good moderation can revitalize a community (hi @dang), while bad comments are a vicious cycle that drive out the good ones. IOW, moderation is very valuable. But it's hard to hire for and do right if it's not already part of your org's thinking. Disqus could sell moderation-as-a-service, according to the customer's parameters, and take care of all those hard parts.