Voice recognition

[Seraphimsan]

What are your thoughts on voice recognition as a primary means of interfacing with a home desktop computer? I've always been for it, up until last night when I was struck with inspiration for an invention ( that I later found has already been implemented but costs several thousands of dollars for anything that's worth buying ). But I digress.

Here are the pros in my opinion:

*Hardware is cheaper than touch capable hardware. (all you need is a mic)
*There are literally hundreds of free/opensource API out there for embedding voice recognition into your application.
*Voice Recognition software comes standard in at least one of the four big OS out there (Windows, Linux, BSD, OSX) It is in windows and works rather well in 7, I don't know if most linux distros have a native VR Software. Nor do I know if BSD or Mac OS X has one.


The cons:

*Software is still a bit clunky at times
*Requires training your computer to your voice (at least it does in windows 7's Speech recognition)
*Not quite as quick or intuitive as touch input.

What are your thoughts on this subject? Obviously it's not going to be the "one future of Human/Computer interaction. But outside of accessability is it something worth pursuing?

As a note I am currently creating a shell replacement for windows that is voice driven (primarily)

[Bazzy]

I'd hate to have to talk to my computer.
IMO,
Mouse+Keyboard >>> Touchscreen || Voice Recognition
( read >>> as 'much better than' )

I've seen that some people are developing ways to communicate with computer using only your brain:
with receptors of electrical charges on your head you can send your thoughts to a computer which will elaborate that input into some really basic commands

That is much more convenient than talking.

[filipe]

I agree with Bazzy. Having to talk to the computer or touch the screen instead of using a keyboard and mouse is clumsy and very tiresome.

[king214]

I agree with Bazzy. Having to talk to the computer or touch the screen instead of using a keyboard and mouse is clumsy and very tiresome.

Thirded!

Although I think that It would be cool I believe that after awhile it would just be a chore. There is one RTS game that I heard of that you use your voice as the primary means of control, but I just can't think of it right now...

[Kyon]

What are your thoughts on voice recognition as a primary means of interfacing with a home desktop computer?

As a primary means of interfacing, I'd say no. Not only is it a lot harder to process, it brings up a lot of inconveniences, since there are different languages, different accents and what not. Finally, the commands will probably be very vague and just not feel natural.

That's for the primary means of interfacing.. as an add-on, it would be interesting to say the least. COMPILE. -compiles current project-

[filipe]

Still, you'd have to have a special word that invokes a "command prompt" of sorts, otherwise you won't be able to talk about compilation to someone else while the computer is listening.

[Seraphimsan]

Well I've been using Windows Speech recognition with macros on 7 for almost 2 years now and I have to say I prefer it mouse and keyboard for most things. I simply say "load c plus plus [subforum]" and it goes straight to the sub forum i say. Or "play music by artist [artist]" and it loads a playlist with that artist. Stuff like that, It's really cool at times. Now for things like CLI programs, or image editing, or software development, thats where i prefer keyboard and mouse, but other than that, voice driven is a great interface....if you give the AI time to learn your speech patters.

As for touch being clunky and tiresome. I disagree entirely. (for me) it's quicker more intuitive and a hell of a lot more fun the point and click, and to be perfectly honest, all interface will be touch based (or purely digital if not that) in the near future anyways, That's my predicament. I could very well be wrong, but watching the trends of technology hasn't failed me yet (or my father, an ex computer scientist from the 80s, for that matter)

People said using a mouse wasn't as quick or fluid as the keyboard, but look at where that lead.

now, all things aside, (disclaimer) I respect all of your opinions, and do NOT plan on releasing this current project of mine up for sale, as it is a bad business decision, Speech driven environments isn't everyones cup of tea. But I will continue developing it as it's of my liking :)

[filipe]

As for touch being clunky and tiresome. I disagree entirely. (for me) it's quicker more intuitive and a hell of a lot more fun the point and click

Just compare how much you have to move your arm to touch the screen to how much you need to move your hand to move a mouse pointer across the entire screen. Consider having to use a GUI where there's lot of clicking to be done. Using a mouse is orders of magnitude less tiring.

[Seraphimsan]

alright I'll give you the tiresome. but still it's more natural.

[NGen]

std::cout << "Hello, world!";

That would probably take about 1 or 2 minutes to write, given that I've had many occasions where it put the wrong thing and I'd have to say "Backspace" repeatedly, hoping it doesn't actually type "backspace". I'll stick with my keyboard and mouse, thanks.

[Seraphimsan]

actually with voice macros that could be printed with the utterance of any arbitrary phrase. last year I experimented with developing software in c++ using dictation. I wrote a couple macros that made writing large blocks of tedious code faster.

I would say "allocate RAM for 500 integers and initialize all to 30"

would output

int *i = new int[500];
for(int j = 0; j<500;++j)
{
    i[j] = 30;
}

or for creating the 'shell' of a class i would say "define class child that inherits parent1 and parent2"

and it would write

class child : parent1 , parent2
{

   child();
   child(child &cpy);
   ~child();

}

edit: accidentally hit submit before I finished the post

[z05DSL3A]

I have never got on with voice recognition (it has been on Macs for a very long time), mainly due to the embarrassment factor of talking to your computer but also environmental issues with background noise and 'other' conversations. I use voice dialing on my mobile phone and end up ranting at it for not being able to 'phone home'.

Touch screen is good on small form factor devices, but I would never use it on a laptop or desktop. I do, however, like Apples Magic Trackpad and Magic Mouse as touch devices.

[Seraphimsan]

To be perfectly honest I don't like any of the above nearly as much as the concept of virtual interfacing:

imagine you had a pair of glasses that hooked up to your multi core smart phone (possibly through wireless) that would display your phones screen in front of you, and on top of that displays a keyboard that you can type on without actually needing a physical keyboard.

edit: currently science fiction, but we have all the components, we just need to put them together to make this happen.

[Albatross]

without actually needing a physical keyboard.

During the first few months of release, it would look like the real world is now owned, maintained, and managed by Linden Labs, no?

-Albatross

[chrisname]

I absolutely can't stand touch. My phone has a touch screen and I hate it.

[Seraphimsan]

@albatross that would be funny, and second life is the single most disturbing collection of code I've ever encountered. People take it as a literal second reality, and that's just sick.

@Chrisname What kind of phone is it? If touch isn't done right it can be horrible, but with the iPhone, Android and Windows Phone 7 it's really done right in my opinion.

[chrisname]

@Seraphimsan,
It's an LG Viewty; it's horrible - excessively difficult to use, software is generally bad, poor battery life. To its defence it has a pretty good camera but that's only since I broke the little piece of plastic covering the lens.

This is in contrast to my Sansa Clip+ which is possibly the most satisfactory thing I ever bought. It's priced reasonably, works flawlessly, has a very impressive battery life (15+ hours, also bear in mind that it's about the same size as two of my thumbs) and, since I installed RockBox ( http://www.rockbox.org/ ), plays Doom. Rockbox also is amazing; it has everything. I can't see when I'm ever going to need to calculate MD5 hashes in my travels, but if something needs to be hashed, I, with my Clip, will be the man to ask.

[wtf]

I've seen that some people are developing ways to communicate with computer using only your brain:
with receptors of electrical charges on your head you can send your thoughts to a computer which will elaborate that input into some really basic commands

I've read about that too. But I think its freightening, really. What would happen if I were all connected and everything, and i thought about, just for a split second, "What would happen if I formatted my computer?".... then the computer would interpret that as... "Format Hard Drive. Ignore further user input until task complete. "

[chrisname]

I more interested in the possibility of "trusted" computing and brainwave-controlled computers... a whole new way for companies to enslave us... I guarantee rms will have something to say about this.

[filipe]

What would happen if I were all connected and everything, and i thought about, just for a split second, "What would happen if I formatted my computer?".... then the computer would interpret that as... "Format Hard Drive. Ignore further user input until task complete. "

It's the same difference between thinking "I'd like to move my arm towards x" and actually doing it, isn't it? Surely it can be solved. And of course, computers will still ask "Are you sure?" for every potentially damaging operation.