I think 5 years from now, we'll be communicating with the digital world through... voice & gestures. Sorry if that was a let-down. Yes, we already have pretty good speech recognizers as well as text-to-speech. With Siri on iPhone4, we even have a voice activated 'personal assistant'. But I still think that the primary form of interaction would be voice in five years. With Siri, and other similar software, we're just starting to explore commercial voice activated software on a large scale. Sure, there's been research in this area for many years now, but latest smart phones are the first commercial products to make these features available to everyone.
In spite of that, there's lots of scope for improvement and making these 'agents' even more intelligent and addressing accessibility issues. And I suspect, that that's going to take the next 5 years. So we'll probably have very sophisticated agents that function very well with voice and maybe even gestures (like the Kinect)? Maybe we'll see an integrated software that can understand voice, gestures, touch etc. I also hope we see some usage of 'smell', to give us the 4D effect!
Of course if you asked me how I'd communicate with the digital world in about 35 years, I'd say with my 'mind'! We probably won't even have to say or use gestures then, just think of it and your device would know it :) While we're at it, I hope that we have a common database of human knowledge then - just like Issac Asimov's people from Gaya. That would eliminate the need for classes - simply access the information from the comprehensive database!