Why Hal Will Never Exist

aengblom writes "Researchers at the University of Maryland's Human-Computer Interaction Lab are suggesting what many of us have already guessed. The future of human-computer interaction won't be through speech--it will remain visual (they explain why). The Washington Post is running a story about the researchers and how they think we will get computers to do what we want. The article is a fascinating read and is joined by a great video clip (real or quicktime) of the researchers and their methods. The Post is holding an online discussion with the researchers tomorrow. Also check-out Photomesa the lab's software program that helps track images on a computer. (Throw a directory with a 1,000 high-res files at this thing and you can justify that pricey new computer you bought)."

Comments

Thinking out loud?

[galaga79]

"It turns out speaking uses auditory memory, which is in the same space as your short-term and working memory," he adds.

What that means, basically, is that it's hard to speak and think at the same time.

I don't know about this statement, I always find it easier to write and/or think when I am expressing my thoughts out loud. Wasn't this something we were tought in school, like it's easier to read out loud than silently? Mind you having done two years of psychology I realise there is a lot differing opinions about how the brain works, so can any psychology graduates tell me if his statement is true?

Re:Single Modality?

[_Quinn]

(Mod the parent up.)

Aside from this, making a speech interface anyone wants to use isn't about the speech; it's about the natural-language comprehension that most people (naively?) associate with speech recognition; e.g., the Enterprise's computer. Which, you note, the crew interact with on a technical level visually.

As for the specific example of italicizing text, natural language understanding should give rise to accurate _dictation_ systems, where the computer will insert the appropriate puncuation and emphases as you speak. If you're typing, instead, CTRL+I is your friend. :)

-_Quinn

The real issue

[00_NOP]

Is surely whether, in the future, computers will be bothered to talk to us.

There is no doubt that computers with greater intelligence - ie an ability to learn and adapt - than ourselves will be here, probably in the next 20 - 25 years.

When these machines get here they may well decide that speaking is a waste of their time.

Re:I agree

[CyberDruid]

Voice interface is excellent for communication from a distance. When I'm sitting in my couch, I don't want to go all the way over to my computer to check trivial things like if I have mail, when the Simpsons is on, what I have scheduled for today, playing an mp3-album, etc, etc. I just want to tell my computer to do it from wherever I happen to be. If I ask for information, the computer can use text-to-speech to give it to me.
I'm actually looking in to the possibility of setting up such a system for myself (mostly for hack-value, of course ;). Just need decent open source voice recognition for a few pre-defined commands. I'll probably need a way to place a few (2-3) cheap microphones in my apartment and connect them (in series?) to my computer, as well.

I don't know what to think about this article...

[heideggier]

I think that the bloke is right that speech is a really bad way of communicating with computers, as they are designed today. But think that it's a bit of a leap of logic to conclude that this will always be the case.

Case inpoint, today computers are normally designed around some kind of windows environment, a Wimp interface, where information in displayed as a metaphore, ie scoll bars, ok buttions etc etc. This is an environment that was never designed for interact beyound a mouse and a keyboard. DVD however do not follow this standard, normally being based on some kind of menu system. Clearly, the way you make something determines the way it is used.

If speech is to be a sucess on computers then the way that people interact with the computer needs to be changed. I think a system like the console where programs arn't very powerfull on their own but due to the way that they have been linked together would work very very well.

I long for the day when I can say, "dump down everything on slashdot and tell me if any of my post have been modded up" to read wget somesite | grep index.html | echo $whatever (please excluse this example), all you would need is somekind of AL which is able to manage the interpreation correctlly (at least most of the time).

I think, fundamentally, computers should be designed to so what you tell them to do (how I think such a system would work) and not force you to do things in a certain way, which is what current systems do today, One should never have to learn a interface.

I also think that this guy has limited his imagination somewhat, the main thing about hal was that he was everywhere, and that in the future, computers are everywhere. For example if you were on the loo, and just thought up a really good chess move, then you would just say, Hal queen to bishop 4, not get up, sit at a console, login a realise you've forgotten what it was you where about to do. Saying that in such a case it's easier to point to some graphic, cause you don't have to think to much, Seems kinda lame

Re:Single Modality?

[MoneyT]

You do of course relize that the comprehension part is the least of our worries. Try telling your computer to open a temporary file on your computer. Have you seen some of those file names? If we do go to speech commands, we're going to need to get a much better system of naming things (can't name your documents dsfk.txt anymore). As for just getting files or programs to open, Apple's speach recognition does this fairly well. Just place and alias (or the actual file) into the speakable items folder and then tell the computer to open [item name]. They even have a command to make a currently selected item speakable (places an alias in there for you). Admittedly, it isn't the best interface yet, but it's a start. And the voice passwords are just so friggen cool (OS 9 only, when do I get it for X?)