possible help for programming by speech recognition

[esj]

my goal is to try and make as many tools accessible by speech recognition is possible. I really like reinteract but it needs some help to work with people like me.

1) I would like to add "Toggle words", a model that enables someone using unenhanced speech recognition to create typical mangled names from an English phrase and more importantly, reverse the process so one can edit the result

2) Enhanced navigation by language features

3) mapping all keystroke commands to something that makes it more resistant to misrecognition events. Try dictating into Thunderbird and see what happens for e-mail.

I can give you a URL for a white paper I've written on the topic of toggle words. In a nutshell, you identify a string of Words or a symbol around the current cursor location in that region is manipulated based on the type of field in that region.

If you have other questions, don't hesitate to ask. I'll explain far more than you want to hear