Dragon/speech friendly Notepad replacement for voice coders

If you’ve tried to code by voice, you’ve likely realized that there are no compilers that support select and say. (If you found one, tell me, please.) Therefore, you dictate the code into Notepad, then copy and paste it into a compiler such as Visual C++, then try to compile and run it. The rub comes if Visual C++ turns up errors. It will helpfully give you the line numbers, and you may even be able to tell Dragon to go down the number of lines. The problem is that Visual C++ is MISERABLE when it comes to selecting and correcting text, and even dictating. Therefore, you have to go back to the original code in Notepad, find the errors there, fix them in Notepad, copy the document, go back into Visual C++, delete the code there, and paste the new code.

The problem with Notepad is that it doesn’t display line numbers; you have to count down the lines yourself. If you have a fair amount of code, your eyes may trick you and you might skip lines, making it rather tedious. But I found a program that might make it a little easier to dictate and vocally edit code. It’s the one Notepad replacement program out of I don’t know how many I tried that actually supports select and say: Win32Pad by Gennady Feldman. Even better, it has line numbering.

The line numbering is not turned on by default. It’s under View — Line Numbers. (The menus are vocally accessible as well. There are also keyboard shortcuts for most major functions.) Once you have turned on line numbering, it becomes easier to tell Dragon to go down or up some number of lines to find your error and correct it. Dragon’s own “go to line” command doesn’t work, but Win32Pad has its own shortcut, which you can access by saying “Edit,”then “go to line,” then saying the line number and OK. You could also tell Dragon to press Control G.

SDK: How to make a Dragon/speech recognition compatible program

In the recent brouhaha over Microsoft’s Word lawsuit, I couldn’t help wondering what would happen if Word really did disappear. That’s a longshot, but all the same… what would users of speech recognition be left with, besides WordPad? For once I’m not talking about tinkering or forcing things to work. I mean something that comes with select and say capability and vocally accessible menus — a program with a little more forethought. So…

Jarte Plus. One program. It’s a decent program, and it’s current and cheap — $19. But come on. Over the course of my experiments, I must have rejected at least 50 text editors and word processors because they just didn’t bother. Are you telling me that only one company is aware of speech recognition users enough that they KNEW that we’d be trying it out? Seriously, check the Jarte help site. There is a setting file in Jarte Plus specifically called Detect DNS. DNS does actually stand for Dragon NaturallySpeaking. (The issue mentioned in the help site doesn’t seem to apply to the Plus version, so no worries.) Carolina Road Software knew that basing something on Microsoft’s code would make it accessible to a broad variety of assistive technology, even if they weren’t users themselves. You can bet I sent a thank you note.

I really don’t understand the lack, though. Is it that people who go to school for programming aren’t taught about accessibility issues in their classes? Or is it that they think it’s too much work? Or do they not understand that it is NOT 100% the assistive technology programmer’s job, that they actually have to work together? Or is it just that they haven’t heard of SAPI or speech recognition soon enough? Or do they just figure that some other programmer will do it, and it’s not their problem? I don’t know.

There is help for the 2nd question though. On their development website, Microsoft has an overview of SAPI for speech recognition. Microsoft also has an SDK — a developer’s kit that actually has samples and instructions on how to make programs accessible to speech recognition. What a concept! There’s even a simple word processing program as an example, called Dictation Pad — the Windows equivalent of DragonPad, I guess. You use that as a springboard to your own applications, so that you know what you’re doing. Part of the work has already been done for you. So why aren’t more programmers making use of SDK? I don’t understand that either. This is why I wish that my study of C++ programming didn’t occur in such fits and starts, or I’d do it myself. But I’m nowhere near that point yet and don’t know if I ever will be. Dammit.

PS — once again, I’m not making light of the work involved in programming. I just don’t understand why programmers aren’t availing themselves of things that would help them to understand accessibility…if lack of understanding is indeed the problem.

Hey programmers! We’re down here!

It strikes me that half the e-mail I get in my inbox is useless to me. All the technological newsletters talk about is Windows versus Linux/UNIX, open source or “indie” software versus Microsoft, and so forth. If you depend to any great extent on speech recognition, you still don’t have a choice. It’s usually Windows and Microsoft all the way, because like it or not, Microsoft is the only one who’s explicitly going to make things accessible for us. (Not everything — see Visual Studio — but more than most.) Many programmers don’t seem to give a damn, or they’re quick to blame everything on Nuance or Windows Speech Recognition. (There are a few exceptions, e.g. Jarte.)

Now, it galls me to stick up for a company whose technical support is rude and/or clueless, and who’s taken to borderline false advertising, but I will say this much: at this point, Nuance seems to have done as much as they can. Compatibility with assistive technology goes 2 ways. Speech recognition can’t work fully with applications that don’t support SAPI. SAPI is the brain, or central nervous system, of speech recognition. Speech recognition can’t run on anything else but SAPI. Supporting SAPI-dependent technology is, like it or not, the other programmer’s job. Also, it looks even worse if the other programmer can’t even be bothered to use general menu controls such as Microsoft Active Accessibility (MSAA) or its successor. MSAA is not SAPI, but it helps a lot when we want to access menus vocally.

Compared to something like screen magnification or screen readers, speech recognition is a newer technology (though it’s been around for quite a while), so I don’t fault programmers for being possibly unaware of SAPI. Also, I am aware that integrating technologies is difficult and takes time. So I don’t fault programmers for saying that it’s difficult, or having to go through quite a lot of trial and error. But from some of the attitudes I see — “Oh, it’s faster for me to just reach over to the keyboard and mouse, why would you need speech recognition?” “I can type 100 words per minute, that’s lazy” — it looks on a bad day like they just can’t be bothered. If compatibility can’t be achieved because of a genuine program-language barrier, that’s understandable and acceptable. But what gets me is that, with so much apathy, there’s no way of knowing whether the problem is that nobody can do it, or just that nobody’s thinking about it. If you’re one of the programmers trying to do something about it, I’m not referring to you, and I thank you.

P.S. I don’t mention Macs because they do, at least, have MacSpeech Dictate — though I hear it’s still catching up to the Windows Dragon. Also, I am aware of the OpenOffice API Project, though I don’t know if they’d have anything to do with SAPI. I am also aware that Linux truly cannot physically support Dragon, as is explained here.