Slashdot Mirror
Microsoft Voice Command Almost Here
PDA User writes "The new Microsoft Voice Command for Pocket PC isn't supposed to be out until the next Comdex, but someone inside the company posted details to Handango and Geekzone posted a preview.
The application notifies users of appointments, and answer simple English questions. It does not have "Do you want fries with that?" in the vocabulary though."
Besides the obvious excuse to use the word "paradigm" I have to ask if this is going to really be that big a deal. I mean, yes it's great to have this kind of thing for the disabled... but the majority of the use cases for this kind of thing seem hampered.
I can't think of a lot of reasons I'd want something like this on a home system. Besides the creepy "I'm talking to the walls" feel, I can't imagine it would be any faster or more engaging to interact in this manner.
Public use cases seem like a good idea, until you realize that background noise and assholes shouting commands over your sholder could end up causing more problems than the system solved.
When this kind of thing can be attached to a really powerfull datamining engine and equiped with a much larger vocabularly I'll be impressed. Right now it looks like more of a toy.
Killfile(TGK)
No trees were killed in the creation of this post. However, many electrons were inconvenienced.
Why can't you talk to your computer? Sounds like a good idea, right? Hell, it's a star trek fanatasy come true for 90% of the geeks out there. So why can't you?
The answer is not because the technology is not good enough. Speech command software has come a long way, and in most situations, with the right microphone, it can be very realiable...if you talk clearly.
The problem is that people don't naturally talk clearly. They repeat themselves, add in lots of "ummm"s and "errr"s and "like"s, and generally speak in ways that only another human could symantically understand. Because of this, using a keyboard or mouse to communicate with a computer is always going to me a more effecient mechanism.
Think about it...which is faster: Saying (in a clear, even tone) "Select the 3rd item in the list" or just clicking on it. Even in ideal environments using buttons on a PDA is going to faster and more reliable than voice command.
The only reason humans can use speech to effciently communicate with each other is that along with speech comes tone, body language, and symantic context that conveys as much (if not more) information than the vocal message itself. Computers cannot pick up on those kinds of things.
I seriously doubt that anyone will make extensive use of this feature.
The problem with every software that I have used that tries to decipher human language (like Zork or the game included with emacs for X) is that you have to know what words the software understands and in what context.
I have seen the same problems with automated phone systems that are supposed to recognize a generic voice and I can see the same thing happening here.
The main difference here though, is that when entering text, you know exactly what you input before pressing enter. With voice recognition software, how do you know that the software "hears" exactly what you say? If you say somethign like "What are my appointments for the thirteenth?" and it hears, "What are my appointments for the thirtieth?" you would be receiving the wrong information.
I hope this is a success but I don't have my hopes up.
Not everything is analogous to cars. Car analogies rarely work.
People in suits talking to machines on next years CeBit while trying to demonstrate that it really works, "yeah, it does!", while all I can see is that dialing on my Nokia 6210 outperforms any voice recognition attempt by minutes. Sorry for beeing cynical, but I was promised these kind of things since back from 0S2-Warp days and everything so far was crap.
If Microsofts attempt on handwriting recognition is any indication, this thing will fail terribly. Have you ever tried to use a Tablet? You must be a real bad typer if this makes you more productive.
Just my guess, please proof me wrong since these things would be really cool if it worked.
cu,
Lispy
Yes there's a need among the handicapped, but the real demand for this has been among the old farts who still see typing as "secretarial" work.
When I worked for a large law firm, that was one of the most aggravating and consistent comments we would get was "Why do I have to do all this typing? Can't you just load some software that'll listen to what I want?"
I've never seen a real practical use for this in an office environment.
*
The answer is not because the technology is not good enough. Speech command software has come a long way, and in most situations, with the right microphone, it can be very realiable...if you talk clearly. The problem is that people don't naturally talk clearly. They repeat themselves, add in lots of "ummm"s and "errr"s and "like"s, and generally speak in ways that only another human could symantically understand. Because of this, using a keyboard or mouse to communicate with a computer is always going to me a more effecient mechanism.
You miss the point. Just like with Pen input, everyone wants to think of the tech as a replacement for keyboards when in reality it is simply an alternative form of input that will be extremely useful given the right situation/environment. One obvious one is allowing input/control for those who are disabled (or close to it). A friend of mine's mother had horrible arthritis in both hands, it really kept her from being able to use her computer much. Using voice input allowed her to utilize it much more than what she would have been capable of otherwise. Another general scenerio is someone who needs to use their hands, but would be convenient to "use" a computer while their doing their activity. Heck, this could be useful even driving. One of the biggest complaints about the fancy computer systems in higher end cars today (e.g. BMW) is their complexity. Well this is perfect since BMW uses WinCE (for now at least) and telling the car to "adjust the temperature to 72 degrees" is simpler than wading through the menus (or to "give directions to Hudson St").
So you are correct in that it's not likely to walk into a cubicle farm and hear "File|Save As|myresume.doc" and "10 of clubs under jack of diamonds" coming from the mouths of four dozen workers. But there are PLENTY of other very useful applications for this technology.
This is something that the linux based PDAs should wait for a long time to have. While Microsoft maybe a bully in the marketing, it's bringing such technologies like subject-independant hand-written recognition and speach recognition to PDAs is a very strong move. How long would it take that the open-source community can bring such technologies to the scene? For anyone involved in signal-processing, it is clear that such things, though yet far from being complete, are only possible with huge R&D budgets. Maybe we yet need big companies and big moneys....
The big deal is that this is in a *Pocket* operating system.
Number one it's much more difficult because of limited resources. I don't know of many who've actually accomplished useable voice recognition on a pocket platform.
Number two, think about the uses of this. The two major bottlenecks of handheld systems are input and output. If the speech recognition gets good enough (which I doubt at this point but which will probably happen sooner rather than later) you might not have to use the little pin-pusher thumb keyboard or handwriting recognition. Also, think of this being used on a combination pda/cell phone. You can use your hands free set for the phone to control the pda and also get information from it. You could use it driving much easier then.
I'm not getting all lathered up over this, but it is kind of cool.
"He's more machine now than man, twisted and evil."
What appointments do I have today?
...
>> Acknowleged, changing permissions to Administrator.
No. List my appointments.
>> Delete all records in database: Are you sure you want to do that?
NOOO!!!!!
>> Yes. Ok. Database deleted. Continue with disk format?
Slashdot Syndrome: the sudden, extreme urge to correct someone in order to validate one's self.
While it's true that we're a long way from voice input becoming the normal mode of input for typical computing tasks, there are certain situations where it can be used to good effect. One example is cell phones. My most recent cell phone has voice dialing, and it has turned out to be a very convenient feature. The commands are simple enough that the phone gets it right virtually 100% of the time, and it eliminates me having to take my eyes off the road and push buttons to scroll through the phone list (which is a big safety improvement in a car).
Perhaps on a PDA the advantages are more marginal, especially since most common apps can be bound to PDA buttons, but there may be circumstances where it is convenient, say if you have to pull out your PDA in the car to get the address of a business contact you are on your way to see.
I do agree, though, that there are cases where we seem to be developing technology for technology's sake. A lot of today's new technologies seem to represent only minor improvements in standard of living compared to the landmark ones that came earlier like the automobile or the refrigerator.
As I recall, voice recognition still ain't quite 100% yet...
Dude, human voice recognition isn't 100% either. How many times a day do you say "excuse me," or "what?" because your ears missed something, or somebody mubled?
Voice recognition will never be "100%" because speech isn't perfect. At least not until voice recognition software designers realize how flawed speech is, and program software to say "Can you repeat the part about the...stuff?"
Hey freaks: now you're ju