The Twitter Trend lecture will leave you with a number of ideas about tracking tweets.
It occurred to me watching the video that a Twitter stream could be used as a feed into a topic map.
Not the same as converting a tweet feed into a topic map, where you accept all tweets on some specified condition.
No, more along the lines that the topic map application watches for tweets from particular users or from particular users with specified hash tags, and when observed, adds information to a topic map.
Thinking such a feed mechanism could have templates that are invoked based upon hash tags for the treatment of tweet content or to marshal other information to be included in the map.
For example, I tweet: doi:10.3789/isqv24n2-3.2012 #tmbib .
A TM application recognizes the #tmbib, invokes a topic map bibliography template, uses the DOI to harvests the title, author, abstract, creates appropriate topics. (Or whatever your template is designed to do.)
Advantage: I don’t have to create and evangelize a new protocol for communication with my topic maps.
Advantage: Someone else is maintaining the pipe. (Not to be underestimated.)
Advantage: Tweet software is nearly ubiquitous.
Do you see a downside to this approach?