Funny as there wasn't really anything re-imagined about the online discussions -- they're still lists of effectively anonymous text.<p>I might be quite opinionated on this as much of my PhD thesis [1] was focused on what could be new ways to interact with people online.<p>There are several core problems with the way discussions work right now.<p>1. Too much content means you'll miss a lot. Sorting is usually effectively random, especially in relation to what you'd find interesting. Summary is one approach, but nothing has been executed spectacularly well compared to what could be.<p>2. The current design is purely a text entry (of only one comment) plus some tiny representation of the poster. In real life, we focus more on the person and what we can ascertain in order to determine credibility, if they're a jerk or someone we want to talk to, etc.<p>3. We're also making the assumption that keyboard-based text is the best & only way to express ourselves. Video/images in abstract are just as vague -- I'm talking more specialized interactions to build up a "speech act" in some new form/medium. Jeff Heer's sense.us [2] is just one of a million ways that could be done in a more fine-grained goal-oriented fashion.<p>Plus text is much less interesting to look at than something pretty and graphical.<p>4. There are varying arguments about the goal of these discussions, but they can at least be seen as<p><pre><code> a) correcting some non-participant's information (blog post, paper, link, etc)
b) attempting to determine main/alternative arguments for/against something
c) attempting to reach group consensus
d) normal social interaction for social interaction purposes (which itself is wide ranging)
e) establishing a sense of community, intrinsically linked to (d) but still different
f) sharing general knowledge
g) expressivity to react to something
</code></pre>
Each of these sub-goals can be supported to a certain degree by existing paradigms but clearly if one goal is much more important than the next then it should be clear that new paradigms or designs are needed to better address one extremely well.<p>It's also the case that many are unicorns: attempting to reach group consensus is possible when thinking about software for small sets of people who need it, but fails when you're thinking about something large-scale like say 'answering' political questions in the US. Trying to dissuade people of their existing biases is a loosing battle. And there have been many tries at something like this, mostly in the research community at places like CHI and CSCW.<p>I personally think #2 is the most exciting, as that's what my PhD was about :) The primary interface doesn't have to be just a huge amount of bottom-up text... if we think more top-down we can gain a sense of a community, discussion, sets of people in a more direct manner. Computational ability is on our side to be able to compress large amounts of data by recognizing varying dimensions in which to navigate. We don't have to just look at 'votes,' but can think about the larger arc of an individual across all of their participation. That as a basis allows us to do what we can't do in real life: segment, shift, and synthesize individuals into groups, style, credibility, cultural position, social position, viewpoints and more.<p>[1] <a href="http://azinman.com/pdfs/aaron_zinman_phd_disstertation.pdf" rel="nofollow">http://azinman.com/pdfs/aaron_zinman_phd_disstertation.pdf</a><p>[2] <a href="http://hci.stanford.edu/jheer/files/jheer-thesis.pdf" rel="nofollow">http://hci.stanford.edu/jheer/files/jheer-thesis.pdf</a>