5 ms·

A version of social discussion site optimized for continuous discussion around a topic. One of my biggest complaints about HN is that it is very reactionary an

by GavinB 17y ago

A version of social discussion site optimized for continuous discussion around a topic. One of my biggest complaints about HN is that it is very reactionary and every discussion ends in 24-48 hours. One of the reasons we have less "Hacker News" is that we've exhausted superficial discussions of a lot of topics and don't have a way to go more in-depth. I have some ideas about how to implement this but I don't want to spam the comments with a long essay.

jimmybot 17y ago
I totally agree; it's a problem I've thought about as well, and it's not just HN either. People scan news headlines (whether on news media outlets or on news aggregators) always looking for something shocking or moving or hilarious or despicable. But after the initial reaction, often it fizzes out and there's no follow up. I would say though comments fairly quickly, the discussion can and does continue at times through a second or third link posted that gets voted up. But yeah, not enough times, and it's not even obvious often what it was posted in response to. Well, I'd be interested in looking at that long essay if you end up posting it somewhere.
philwelch 17y ago
Along the same lines--a wiki that contains all documented arguments and counterarguments on a given topic instead of rehashing them all the time over and over again on social discussion sites.
derefr 17y ago
I've thought about this one too. I think I can write less than an essay, though: the solution is basically tagging in reverse. Comment threads aren't attached to individual links, but rather to tags (thus there is a 1:1 correspondence between "tag" and "discussion"; the words can be used interchangeably.) Discussions keep a list at the top of all the articles that go into them. The score of a link isn't an independent variable, but rather the sum of the "relevance" scores for each link-tag pair (that is, the weights of the graph edges.)
- joshu 17y ago
  so, an IRC channel?
  - derefr 17y ago
    Not quite—it's about half-way in between a Reddit-like model and an IRC server. Unlike an IRC server, everything someone "says" (links to) must be cross-posted to all the threads/channels they think it's applicable to, instead of hidden within the context of the conversation being had at the moment. Unlike an IRC server, people can vote on the articles' relevances to each conversation. And unlike an IRC server, the articles with the highest aggregate relevances show up on a "front page." It would take about the same amount of work to make an IRC service daemon that managed voting and gave a "front page" listing as it would to extend a social-news/bookmarking service with tag-link-pair voting and move discussions to tags. Plus, I think the IRC culture would be less willing to accept the resulting "conversation bleed" than the social-news/bookmarking culture.

by GavinB 17y ago

jimmybot 17y ago

I totally agree; it's a problem I've thought about as well, and it's not just HN either. People scan news headlines (whether on news media outlets or on news aggregators) always looking for something shocking or moving or hilarious or despicable. But after the initial reaction, often it fizzes out and there's no follow up. I would say though comments fairly quickly, the discussion can and does continue at times through a second or third link posted that gets voted up. But yeah, not enough times, and it's not even obvious often what it was posted in response to. Well, I'd be interested in looking at that long essay if you end up posting it somewhere.

philwelch 17y ago

Along the same lines--a wiki that contains all documented arguments and counterarguments on a given topic instead of rehashing them all the time over and over again on social discussion sites.

derefr 17y ago

I've thought about this one too. I think I can write less than an essay, though: the solution is basically tagging in reverse. Comment threads aren't attached to individual links, but rather to tags (thus there is a 1:1 correspondence between "tag" and "discussion"; the words can be used interchangeably.) Discussions keep a list at the top of all the articles that go into them. The score of a link isn't an independent variable, but rather the sum of the "relevance" scores for each link-tag pair (that is, the weights of the graph edges.)

joshu 17y ago
so, an IRC channel?
- derefr 17y ago
  Not quite—it's about half-way in between a Reddit-like model and an IRC server. Unlike an IRC server, everything someone "says" (links to) must be cross-posted to all the threads/channels they think it's applicable to, instead of hidden within the context of the conversation being had at the moment. Unlike an IRC server, people can vote on the articles' relevances to each conversation. And unlike an IRC server, the articles with the highest aggregate relevances show up on a "front page." It would take about the same amount of work to make an IRC service daemon that managed voting and gave a "front page" listing as it would to extend a social-news/bookmarking service with tag-link-pair voting and move discussions to tags. Plus, I think the IRC culture would be less willing to accept the resulting "conversation bleed" than the social-news/bookmarking culture.