[Accessibility] evil thoughts

[Eric S. Johansson]

 just a few ideas I want feedback on when I wake up in the morning.

How do we specify a speech driven interface for editing text. NaturallySpeaking 
uses "Select-and-Say". I think we can do something based on Mark and point 
generalized to any textbased window.

"Mark after speech driven interface
point after to any
region bold
region cut
region copy
point before textbased window
paste it
drop it
yank it back


Now, for extra points make it work across the network. Text window on one side, 
recognizer on the other

For extra extra points, embed the client-side into Emacs. Create a recognizer 
side for natXXXXXXXXXXXXXX so that people can do real work with it we can learn 
about how well it suits their needs

-- -- --
I'm assuming that NaturallySpeaking in Windows in a virtual machine will never 
be politically correct even though it's extremely practical. I know the audio 
channel is a stinking pile of crap virtual box. Is there any other free virtual 
machine that can run XP and NaturallySpeaking with a decent audio channel or 
decent USB traversal? If not, do we have the expertise to rework USB handling in 
a virtual machine or do a host to guest transfer of audio without any skips, 
gaps, pops, static, resends etc.

More later, I'm an hour late getting to bed because you guys are just so much 
fun