Enter a search term and you get back a report on how many positive, netural, negative and the total tweets for the subject. The app ranks the tweets based on searches for words like “happy” or “the best” and says these are positive. The words “worried,” “confused,” and “hypocrisy” make a tweet negative.

The problem? Irony. Take this tweet, for example:

Sen. mcain is that the best you can do? every day it more clear why barack obama is the best candidate, please vote dont let mcain distract (view)

It gets ranked as positive, though a human rater would probably put it in the negative category.