One of the reasons YouTube videos are such powerful brand marketing aids in a semantic web is because the visual element of a video makes it easy to absorb content on the fly and engage in a way that allows brand sentiment to surface.
Comments are judged according to:
Directionality - where did the connection come from? Was the vicinity domain relevant (i.e did the comment come from a group of people who regularly comment on this subject?) Was it from a bad neighborhood (which might have been trollish) or from a domain not usually associated with the current interaction?
Temporary or Persistent - was the connection haphazard, or by design? Did it lead to a real connection because it contributed something significant or was it a miss?
Transivity - does the current connection stand alone or are there shared friends and interests linking the commenter and the person who posted the video? If yes, what level are they at?
Priority - Was the connection one that resulted in a response and further engagement? Or was it ignored as irrelevant?
What all this reveals is that comments and engagement, interaction and connections in the semantic web matter as much as they do in real life.
Excerpt from: http://goo.gl/WvlYiy