But we have had voice controlled programs and they just never seem to really take off. Dragon sounds great for just talking into my iPhone to text but it isn't that popular. Why are people thinking Siri will be the difference?
Because of the architecture. Voice Control has been done well for years. I've used Nuance's Dragon NaturallySpeaking and after brief training and correct enunciation it is more accurate than 90% of typists.
What makes Siri different is that it mashes together the necessary ingredients.
Natural Language Input with Speech to Text
Latent Semantic Mapping to actually analyze and understand the words
Context sensitivity (Today, Tomorrow,)
Conversation flow and Delegation (reschedule to 2pm, reply, etc)
What people know now is Speech to Text with the results thrown into a search engine. That was impressive a few years ago but these systems are easy to fool and thus become novelty features much sooner.
Siri is broken up into a module architecture. The front end handles the speech input which should support far more languages. The middle handles the analysis and the rear is the their domain systems that link to external API like Yelp, Wolfram and much more.
For more potential API simply look at
http://www.programmableweb.com/
Which details startups and more with API based services. They are all natural candidates for Siri.
Also I'm betting that WWDC 2012 announces developer support for Siri via an API. I'd love to be able to say "Create a task" and have Siri understand that I want OmniFocus to get that task and not another app.
Classic!
