@alexboche sorry for the major delay in responding to your question. My academic life was kind of thrown into chaos around the beginning of May (having major problems getting along with my PhD advisor), and I sort of withdrew from a lot of outside projects and communities. I think I'm back now. Anyway, yes, when I say that I do programming with my eyes, I am using Optikey. If you're curious, it is another really cool project to check out. I actually designed some custom keyboard layouts specifically for programming that are now included in the project.
@mrob95 I just pulled down the latest version of master, and the latex and latex_maths modules are not loading correctly. I'm seeing error messages
Ignoring rule 'latex'. Failed to load with:
Duplicate keys! (line 169 column 1 char 5622)
Ignoring rule 'latex_maths'. Failed to load with:
Duplicate keys! (line 169 column 1 char 5622)I'm going to go take a look at the file and see if I can identify/fix the problem.
@mrob95 Please ignore my previous message. This was a problem with my use of git. Basically, I had added the in-line command, and then I merged in the master branch which also had the in-line command. And instead of a merge conflict I just ended up with two copies. Oops!
@mrob95 Would you be willing to include the tex source for the documentation in the repository (or in a parallel documentation repository)? I am working with the latex commands and have already added several common latex commands that were missing. It would be great if I could modify the documentation at the same time, and then send you a pull request with both updated code and documentation.
@annakirkpatrick Sorry to hear about those issues with your advisor, that sounds difficult. No worries, there is no rush. I haven't been doing mathematical writing recently, but I'll be starting a PhD program in mathematical economics in a couple months so I'll be writing math by voice regularly for many years to come. Would be interesting to see a video of you coding with your eyes sometime if you have time.
At some point,it might be worth compiling a set of mathematical vocabulary words.Someone mentioned there is a way to add words en masse using Natlink. not a priority right now just mentioning it to get it into the general consciouusness. Rüdiger's dragonphrases is another product that mightt be useful for this . other engines will bring new possibilities too
@mrob95 I wanted to check with you on what I think is an omission from the LaTeX module. I don't see any way to dictate the double backslash ('\\') , beyond the brute force way anyway. I see that you do include the \linebreak command, but it is really more of an advanced command for special purposes or power users. I noticed that one of the templates does include the double backslash in a table so I just wanted to check that I'm not missing something here. Thanks!
And I'll add a command for double backslash; it gets used a lot!
in might be good to have the
add_words vocabulary list be a dictionary that has spoken and written forms. I don't know how that works,I think Natlink has a way. Caspark mentioned this (pseudo-?)code on the Talents slack. Not sure does the spoken forms:import natlink
def add_words(words):
for word in words:
known_word = natlink.getWordInfo(word)
if known_word is None:
natlink.addWord(word)
I'll do some research
Dragon is usually reasonably good though at getting the pronunciation right
The Dragon gui lets you import a list (plain text file) of words and phrases with written and spoken form specified. They are separated with a backslash. (And hence you can't add anything that itself contains a backslash. Nope, there is not an escape mechanism. I did check.) Just sharing this in case it gives some insight into what natlink is looking for.
@mrob95 I have implemented the line and command as you suggested, specifically I chose the version with the carriage return as it is common style to include a carriage return in source in this setting. I'm getting the double backslash and carriage return output correctly, but I am also getting a space after the carriage return. Which isn't a huge problem, because LaTeX ignores a lot of the white space details from the source file, but it is annoying if you are trying to keep clean human-readable source . I'm guessing this behavior has to do with inclusion in the section [commandnoarg], as generally putting a space after a command is the desired behavior. Does this may be indicate that the command belongs in another section? Maybe the misc_symbols section? Or is there a better way to handle this?
@mrob95 in the documentation, did you intend to use a long dash for "or" when listing spoken forms? I'm guessing you actually meant |, as that is the character you use in the source. But | needs to be escaped in LaTeX outside of math mode. Since I'm already working on the documentation, I'm happy to fix that issue while I'm at it. Just wanted to be sure I correctly understood your intent before I started making changes.
@alexboche congratulations on getting accepted your PhD program! At the moment, I'm probably not the right person to tell you how great graduate school can be. (I'm still super frustrated with my advisor, but I am slowly working through it.) I do believe it is a worthwhile endeavor though. If I didn't, I would not still be doing it myself!
If I can remember, I will turn on a screen recorder next time I'm using eye gaze for code on my personal computer. It is pretty cool, and totally different than speech to text.
If I can remember, I will turn on a screen recorder next time I'm using eye gaze for code on my personal computer. It is pretty cool, and totally different than speech to text.
Hope pulling changes isn't causing too much bother
And pulling changes is no biggie. I'm pretty comfortable with git and certainly couldn't suggest an easier way.
np, if you need more flexibility for commands then dragonfly is not at all hard to learn see e.g.
https://github.com/mrob95/MR-caster/blob/master/caster/apps/chrome.py
@mrob95 thanks for the link! Once I get the basics of LaTeX dictation down (both adding commands and getting over most of the learning curve), I am thinking about setting up better integration with vim for easier editing. That looks like a really useful example to work from.
@mrob95 I packaged up my changes so far into a pull request. It's only a handful of commands, but they are significant in terms of usability. I'm sure there will be more as I continue working, but you might as well have these for now.
I get an absurd amount of value out of dragonfly and use it for pretty much everything - text editor commands, programming language commands, math commands, git commands, web browser commands, etc etc
@mrob95 I'm hoping maybe you've seen this error message before and can point me in the right direction. Here's the sequence of events:
- I started up Dragon but natlink didn't start. I did not get the "messages from Natlink" window, and commands like "enable X" were not recognized as commands.
- Pulled latest version of mathfly from master, just to be sure. (Couldn't run rebuild command.)
- I ran the natlink configuration and reregistered NatLink. Rebooted computer and restarted Dragon.
Same problem as 1, so I checked my path. Somehow Python 3 had jumped over Python 2 in my path. I fixed that. Rebooted computer. Ran NatLink again being sure I was using Python 2.7. Reboot again. - Now Dragon starts with NatLink, but I am getting a bunch of error messages.
```
Vocola not active
Ignoring rule 'sublime'. Failed to load with:
cannot import name ContextAction
Ignoring rule 'sumatrapdf'. Failed to load with:
cannot import name ContextAction
Ignoring rule 'wordpad'. Failed to load with:
cannot import name ContextAction
Ignoring rule 'core'. Failed to load with:
cannot import name ContextAction
Ignoring rule 'ScientificNotebook55'. Failed to load with:
cannot import name ContextAction
Ignoring rule 'LyX'. Failed to load with:
cannot import name ContextAction
Ignoring rule 'latex'. Failed to load with:
cannot import name ContextAction
Ignoring rule 'latex_maths'. Failed to load with:
cannot import name ContextAction
Ignoring rule 'alias'. Failed to load with:
cannot import name ContextAction
- Starting Mathfly -
Say "enable <module name>" to begin, or
"configure <module name>" to make changes.
Modules available:
natlinkmain started from C:\NatLink\NatLink\MacroSystem\core:
NatLink version: 4.1whiskey3
DNS version: 13
Python version: 27
Windows Version: 10
User defined macro's (UserDirectory) enabled: C:\Users\Anna\Documents\NatLink\mathfly
```
Any advice?
ContextAction is something new which was only added recently to dragonfly
Jeez it's been so long since I used Lyx now that I can hardly remember why I switched! I recall SN being more convenient for voice rec because it was more streamlined to write multiple lines of Math continuously. Just "math mode" and away you go. I believe in Lyx when you begin a new line you must then specify that you will be writing math in that line. Probably an easy issue to overcome through creating specific commands for such but yeah.
Also I remember there being significantly less requirement in SN for moving the insertion point to the correct place before dictating the actual math. With Lyx I remember having to constantly move the insertion point using ross, dunce etc whereas again SN seemed much more streamlined in that regard.
Furthermore for me specifically , I have a touch screen (Surface Pro) and I am now using the pen to highlight math in SN I've already written so I can copy and paste repetitious math content from one line to the next and then just change the values you are calculating, very convenient for Linear Algebra say. This is quicker and puts less strain on the voice. This way I am using voice commands as an adjunct to this copy/paste with pen along with AutoHotKey scripts to further streamline repetitive tasks and common commands. Anyway, using the pen to highlight text in this way thankfully doesnt really do to much to RSI symptoms and SN + touch screen makes it very easy to do so. This wouldn't be the case with Lyx
@alexboche pretty much what @esc123 said. In scientific notebook you can go into a permanent math mode where everything you type is assumed to be math, regardless of new lines et cetera, whereas LyX requires you to constantly reset the mode and it is easy to accidentally jump out by going to the end of the line or w/e. Also in SN you can select and manipulate blocks of math easily. Since I mostly used it for bashing out homework assignments or completing exams the ease of getting stuff down in SN easily outweighed its lack of complex formatting options.
I think the main use case for LyX is as an alternative to raw LaTeX for technical writing, but since I know LaTeX pretty well and am comfortable in a text editor I don't really have a need for it.
btw if you are considering trying SN, then I would definitely recommend version 5.5 over version 6 if you can get it. Version 6 is much closer to the behaviour of LyX and I really think they have made it worse.
@mrob95 I've noticed in dictating LaTeX with mathfly that Greek letters always have a space after them by default while latin letters do not. Was this an intentional choice? While it doesn't cause any real problems, expressions like
$\Omega _{k}$ in source are nonstandard from a style standpoint. In case it's not clear, $\Omega _{k}$ and $\Omega_{k}$ compile to the same typeset result, since LaTeX ignores most whitespace characters in math mode.