Re: speech recognition support

Angel Ortega wrote:
> Eric S. Johansson (esj_at_harvee.org) wrote:
>
>> a note from the eclipse bug on speech recognition support.
>>
>> ----------
>>
>> http://ct.scansoft.com/customerfiles/kbasefiles/5176/wp_WinAppsCompatibleWithDNS.pdf
>>
>> You need to either use standard Windows edit controls or respond to these
>> Windows messages:
>> EN_SELCHANGE
>> EN_REPLACESEL
>> EN_SETSEL
>> EN_GETSEL
>> WM_GETTEXT
>> WM_SETTEXT.
>>
>> ----------
>>
>> since I don't program in windows, I have no knowledge of how hard it is to make
>> these changes.
>
> Hi, Eric. I'll take a look at that document. As I told you, I'm a total
> ignoramus about speech recognition, and that can prove helpful.

it is great that you are willing to take a look. here is vr-mode for emacs.
http://sourceforge.net/projects/emacs-vr-mode

the main reason I'm showing at you is because you might be able to lift some
code or architecture for what ever you might be able to do and, (if I'm not
asking for too much), make it work across machine boundaries albeit more
successfully than vt-code does. The primarily way VR-mode fails is that it
uses the title bar of the window to communicate both state information back to
the speech recognition engine. Unfortunately, in systems like Emacs, the
taskbar is near constant nor controllable and we need to uses different
techniques for binding what you say to which target.

  for what it's worth, the reason for cross machine access is that, like many
people, I use multiple machines to get my job done. Sometimes I'm hitting as
much as 10 machines at once. It would be really swell if I could dictate into
each of them with the same level of comfort as I could with a native instance.
The other option would be some sort of a "on the spur of the moment" filesystem
like interface to the other machines. ssh works most of the time and would
solve the "install everywhere" problem. Trouble being that most of the time
when I used such an interface (i.e. tramp, Web Drive), performance has been
livable and caching has made it unacceptable. For example, if I change a file,
save it, and then go to the remote machine to install a Python module containing
a file, I don't like having to wait 30 seconds for the cache to flush the data.

another reason for cross machine boundary mechanisms is that naturally speaking
works in wine and it would be really cool to be able to access mp in the linux
context.

Sorry about these things being so complicated but that's the world I live in
when trying to build an accessible framework out of pieces that were not
designed to be accessible

---eric

-- 
To unsubscribe, send mail to mp-unsubscribe_at_lists.triptico.com.