Speech driven apps have been the single biggest craze since Apple introduced Siri in 2010, giving voice recognition much-needed spotlight. For the last decade speech recognition has been locked down to proprietary OS-based devices and largely confined to voice dictation. The flexibility of controlling a device through voice commands has been paraded in Hollywood blockbusters for years, but with speech recognition becoming mobile, we are finally questioning its limits. With most speech recognition infrastructure being built on language intended to communicate ideas, is it possible to grow speech recognition beyond anything more than simple dictation?
Eleuthera: Speech Driven Programming Application.
In 2012, I was afforded the opportunity to participate in
Google Summer of Code with an amazing group of people at the Inclusive Design
Institute (IDI). Still in school, a
disability beginning to affect the joints in my hands, and with no real
programming experience outside of the classroom, I opted to take on the most
challenging concept of all, to try and develop an application that would allow
individuals to program using the voice on any mobile device purely using
JavaScript.
No one else had even applied for the position at that time,
nor has there been much research (that I'm aware of) been done on it since. I jumped straight in with what I called Eleuthera
(freedom from chains), and quickly found
myself drowning in all of the complexities involved in the project. Unfortunately, the project never really got
off the ground because Google's open source speech API was still within its
infancy, and microphone access still relied on Flash. Two years later, however, the landscape has
changed and has breathed new life into Eleuthera. Over the next couple of months, I hope to
continue to release posts about my encounters as a disabled programmer
designing an application with the potential to aid other disabled programmers.
No comments:
Post a Comment