Enslaving Twitter

By on 4/1/2010

I've noticed that in recent years, twitter has become the near defacto "engine" for a number of things:
  • Random list data for demos Scott Guthrie's recent MIX keynote demo is a perfect example of this. He uses twitter to populate a simple listbox using his own feed. I do something similar in the sample project for nBayes where I let the user choose two different search terms to populate the two indices from twitter.

    It's convenient because you don't have to manage the data set. It's a bottomless well of perfectly randomized sample data which makes it perfect for this purpose.
  • Using it as a data communication channel One of the more interesting usage patterns, the use of Twitter to enable large scale communications and/or data transfer. Jeremy Hilton shows how to use it to manage a voting campaign. Other more "ambiguously practical" projects include TwitterDrive, and even a WCF Communication Channel!
The only unfortunate thing is that whenever I see this stuff used in a demo, they almost always use some xml parsing api to parse out results from the xml API. This can tend to be ugly and error prone. Linq2Twitter seems like an interesting solution because it lets you write  nice strongly typed linq queries and get back the results.

Depending on what you're doing, having organic sample data can be very useful, there have been many projects attempting to address this in the past. And the use of twitter as a communication channel could be nice because it's infrastructure you wouldn't have to manage and maintain (just be prepared to handle fail whales).

See more in the archives