LexiconCode on master
Updated Dragonfly requirements … (compare)
@beyondaxis while I personally prefer sublime, I agree with @LexiconCode that you should first give it a try with the editor you are already most comfortable with, as trying to adapt to a new editor while also trying to adopt to a voice only style would most likely result in even lower productivity ,until you have customized enough.Either way, if you want things like vim shortcuts there are extensions that should be able to provide you with functionality you need at a good enough level.
now regarding the excessive mouse clicks, do you thing the big contributing factor is
Clicking/Interfacing with UI element/buttons perhaps you can find keyboard shortcuts for?
Or simply for navigating the source code?
If it is the former, then select-and-say functionality ,to click buttons simply by saying their name,can be your friend. Dragon has its own implementation but there is quite a lot of community work going on at #814 to improve it using accessibility APIs. If that's not enough, then eye-tracking is by far the way to go and there is independent work by @wolfmanstout and @ileben that integrates it with OCR to grab screen text around where you're looking, which should make interfacing with UI elements even easier.
If it is the latter, then your first stop should probably be the excellent TextManipulation commands if it hasn't already. Furthermore quite a few folks in these channel use an Acejump extension for their editor of choice. Finally, you can take a look at any us three projects PythonVoiceCodingPlugin,serenade,VoiceCodeIdea providing syntactical navigation capabilities. Unfortunately, only the second option,which is a whole different stack,supports currently vscode , but by the beginning of the next school year I hope that should no longer be the case
do you know how it works with Dragon
I partly miss read this quote. Grammars are compiled and injected ultimately through nalink. So any grammar we create or that in Caster runs alongside any existing commands in DNS. It would be nice if we could directly override DNS commands. There is a grammar in Caster that allows you to customize control over DNS. DNS Quick reference Documentation. If those commands don't work make sure you have the rule enabled by saying
enable dragon and that's true of all applications that the need to be enabled at least once.
Numbers: numb <wnKK>, 906
for me the numb command doesn't work correctly if I say "zero" two times. like I can say numb 90066, and it will come out as 906. can anyone else confirm?
Interesting so I'll have to investigate the further. I can't say
ninety thousand as it prints out
90 and it is only recognized as
ninety. Testing with Dragonfly
numb ninety thousand, I get 90. that's probably what you meant
"On a random notice, a small update in case anyone is interested in "sounds" with Dragon. after some experimenting, it seems that Dragon can support quiet a few "sounds" without needing to launch accuracy tuning. At the moment I think I have tried around 30 different ones for which we can make dragonfly rules including stuff like shhp,pppf,pppt,tttk,ssss,xxxx. It seems that that the key to get Dragon to recognize those sounds, is to drag the dragonfly specs to them 4 letters or longer. for example, a rule ffp will not work in my end, whereas fffp will!
Unfortunately accuracy can drop if you try to use too many and some combinations do not work well due to phonetic distinctiveness issues. Furthermore, wannabe sounds might be a better description as:
I have no idea how to get things to work with sounds that can not be mapped to letters( so for instance no pops), if anyone has experimented with that and would like to share that would be very interesting!
interestingly, the dragonflies spec will be recognized both if we make a sound or if we spell it. For example , for fffp you can make a sound like here (or adapt a little bit), but it will also recognize if you say FP ( pronounced the same way you would speak fps)
so this is clearly inferior to what dedicated noise recognition systems can do, but I would say it is good enough to get some basic things done. I have been using these lately for things like going up or down, switching between tabs, triggering vimmium , going back and forth in history when browsing or for switching between commonly used applications and so far so good! on my end with DNS 15 home and FlexyMike DEC, unless overloaded Dragon doesn't usually have much of a problem differentiating for instance between shhp,sssp,fffp,tttp,pppp,xxxp,kkkp or fffp,ffft,fffk,fffx,fffs and so on,so you have quite a few options!"
@esc123 thanks for sharing my old post and giving it a try! It is good to know that others reproduce this:)
@beyondaxis While general-purpose noise recognition with arbitrary sounds does not seem possible with DNS, it does end up ,whether intentionally or by accident, supporting a whole bunch of "letters like" sounds such as
fffp,shhh,shhp,pppx,pppt, as long as you pay attention to the four letter trick. of course this is kinda limiting( so you cann't do that tongue click) but it is more than good enough to replace a lot of monosyllables but you might be using for various common tasks like page_up/page_down,next/previous tab,triggering vimium and so on. At the moment, I think I am using about a dozen of such sounds and it is making a world of difference for me,I can highly recommend giving it a try.
For those interested,at #825 there is an example noise grammar for Chrome to get you quickly started
Editing the transformer file words.txt doesn't seem to be working for me I tried to change the directions to their regular counterparts and it seems to take no effect on reboot. Any ideas? All I did was edit that text file as shown in the docs save and then reboot caster.
TextReplacerTransformer = false is
true found in
transformers.toml your user directory in the settings.