My take on most technology is - just think of something that is just plain silly and then come up with a great marketing campaign. You'll make a mint.
I still can't believe that people spend so much time using SMS - what a waste. Just call the person. Why would I want to spend 10 minutes attempting to "type" into an interface that is NOT designed for text input to get 2 lines?
dpyers' idea on testing is good. There is a serious lack out there. Most of the "learning" solutions I've seen are complete crap. They are designed by programmers with no regard for the learning process. The specifications are given by educators who have zero understanding of technology. MASSIVE gap there to be filled.
For a project though, you are best to focus on 1 simple thing. If you get too broad, you'll fail to produce anything other than the junk already out there. The key today is either mass appeal with little added value or functionality, OR niche with real functionality. You've got to choose.