Indie-Resource.com

After hearing a story of a man who had both his legs and arms amputated after an injury serving in the military, I started thinking about what I would do if I was in some situation that caused me to lose the use of my arms. I would obviously still want to code, but how easy could that be? I did some looking around online, and there are few tools out there to help the severely disabled write code productively. One of the few things I found was this, which is a Stack Overflow question, and there was just a hodge-podge of various little projects that tried to make using existing languages bearable. This is sad to me since programming could be a great escape for many severely disabled people, and could even allow them to enter programming as a career. Sadly, many quadriplegic patients unfortunately wrestle with feeling "useless" and like they can't contribute to their family and are being cared for. A programming language focused on being 100% speech-to-text compatible could be a huge help!

What a language like that would need would be to largely remove symbols from the language, and to remove ambiguity from spoken code. For example, take a Java method:
This would be quite awkward to dictate. I am thinking something like this could be better for speech-to-text:
Obviously some of the math stuff could be converted to something a bit more eye-friendly, but I just wanted to note how I think it could sound. Also the IDE could know to run together any identifier between "function" and "passes", which could also be "passes nothing" for methods with no arguments. That would be an example of how the IDE would have to be intelligently setup with well-known rules to allow the dev to proceed quickly, without having to constantly go back and fix the problems current languages suffer when using speech-to-text. Having a good way to write large blocks of "practice code" to see how it looks would also be nice, like an expanded C-like block-comment feature, with additional features, such as the ability to easily choose parts of the practice code to merge into live code, and a way to dictate pseudo-code right in the IDE.

Some fully able-bodied programmers may like it if they are beginning to suffer from carpal tunnel syndrome, or simply don't like typing. A .NET-compatible language would be easier to create than an entire compiler from scratch, and it comes with the huge ecosystem of .NET. I don't know much IL or I would probably give it a whirl myself. I just wanted to get the idea out there to see what anyone else thinks. Sorry for a huge post lol!

That would be an example of how the IDE would have to be intelligently setup with well-known rules to allow the dev to proceed quickly, without having to constantly go back and fix the problems current languages suffer when using speech-to-text.

Not necessarily, the speech recognition software was the one that needs to be the smart and convert the speech to the IDE correct syntax.

Example: You say private static function name Init passes nothing
It would recognize it, and immediately type: But that was a nice thought and i never had it but now i really think that programming is a good hobby for people. it helps us to think and solve problems

It would be nice for a speech recognition program to do that, but I don't think any do, nor do I think they will in the near future. Maybe if one of the major players was extensible that could work, but I don't know of any that are. If so, that would obviously be much easier than creating a whole new language. But if not, then creating a new .NET language would be much easier than creating a totally new speech recognition program.

Jackolantern wrote:It would be nice for a speech recognition program to do that, but I don't think any do, nor do I think they will in the near future. Maybe if one of the major players was extensible that could work, but I don't know of any that are. If so, that would obviously be much easier than creating a whole new language. But if not, then creating a new .NET language would be much easier than creating a totally new speech recognition program.

I'm not an expert on speech recognition, but lets say your speech is recorded in a variable called ORDER

i assume that the program checks if ORDER == DEFAULT_SOMETHING

Then if it is, it will say something right? Then instead of say something, it would type pre-configured syntax.

I just don't know how easy it is to get into the mechanics of the major speech recognition players. I believe Dragon is locked-up tight as a binary. Perhaps there is a good open source speech recognition program that would allow "modes" so that a user could switch to a programming mode. Making a high-end speech recognition program from scratch can take years by a full team and requires many different disciplines and tens of thousands of hours of tuning and testing, so that really isn't an option.

If something like this was done on the speech recognition side, Python would probably make a good choice for a language to focus on since it is not very symbol-ish. It may still be hard to deal with a few ambiguities, though.

Jackolantern wrote:I just don't know how easy it is to get into the mechanics of the major speech recognition players. I believe Dragon is locked-up tight as a binary. Perhaps there is a good open source speech recognition program that would allow "modes" so that a user could switch to a programming mode. Making a high-end speech recognition program from scratch can take years by a full team and requires many different disciplines and tens of thousands of hours of tuning and testing, so that really isn't an option.

If something like this was done on the speech recognition side, Python would probably make a good choice for a language to focus on since it is not very symbol-ish. It may still be hard to deal with a few ambiguities, though.

Could make keywords to symbolize you want to create a symbol, for example

if(y == x){
echo "yes";
}

The user would say "If" then they would say the keyword, something they wouldn't normally say, then they would say parentheses, y, keyword, equals, key, equals, x, key, parentheses, key, curly, echo "yes", key, semi, key, curly

It appears as if it you would be saying the keyword alot, but in actually saying it and especially if it's short phrase, something like "kslash", it would work. Now as far as actually implementing this, I have no idea.

That is the way it works now, and from what I am reading, it doesn't work very well

You can always hire a guy to code while you talk :PPP

vitinho444 wrote:You can always hire a guy to code while you talk :PPP

There is always that

Or a monkey