Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
Ubiquitous computing and MIT's Project Oxygen

By Anonymous Zero in News
Wed Jun 21, 2000 at 04:51:15 PM EST
Tags: Technology (all tags)
Technology

This article from today's Boston Globe announces that some big players are about to join MIT's Project Oxygen. This 5-year project has the ambitious goal of creating a ubiquitous computational environment where computers are hidden in walls and all I/O interfaces are either voice or visual. The project started last fall with initial funding from the Department of Defense and now includes 250 MIT researchers, $50million in funds, and new players include Nokkia, Acer, and Phillips. And as you may know Microsoft Research is flailing in the same directions. But do people really want to talk to the walls like they do on Star Trek? I don't think so. Here's my little rant...


Computers without keyboard and mice is fine by me, are we ever going to stop using keyboards and mice? No, please. That's ridiculous. Keybaords and mice have survived for decades as our defacto input devices not only because they are cheap and simple but also because:

Privacy: Do you want to be sitting in your home office or cubicle at work dictating a deeply personal email into your PC microphone so everyone can hear you? No. Of course you're going to type it out instead. Do you want to be standing in your public library shouting at a periodicals search kiosk "I AM LOOKING FOR 18TH CENTURY NEWSPAPER AND MAGAZINE ARTICLES ABOUT GENITAL HERPES, THANK YOU". I think you and everyone around you would prefer that you type it out instead.

Efficiency and Control: Speech recognition is "not quite there yet" but even if it were I believe experienced writers and programmers will stick to the keyboard. Have you ever tried to talk someone through using say, PhotoShop? It's frustrating because you want to grab the mouse out of their hand and show them instead. How would talking a computer through the same types processes be any more effecient? "MAKE WINDOW BIGGER. STOP. TOO BIG. MAKE WINDOW SMALLER. NO I DON"T WANT A NEW WINDOW!!! ARGH!!!! GIVE THE MOUSE!!!" The keyboard and mouse are fine instruments and once you practice a little you realize subconciously that is the most direct route to translate your thoughts into actions.

Annoyance and Confusion: Can you imagine how your coworkers would feel about having to listen to you babble at your computer all day? You've probably seen that MPG video of the guy that beats the crap out of his workstation. Can you imagine an office full of people talking to the walls? It wouldn't work. "Are you talking to me?" "No, I'm talking to my email client." "What?" "Huh?" "Damnit delete that last line!" "Huh?" "SHUTUP!!!"

Overall I believe these clunkiness of certain aspects of computing persist because we want them to. Example: We prefer email instead of live voice streaming because we can answer email at our leisure and carefully edit our responses. Computer languages are complex because we are control freaks and want the ability to be as close to the hardware as possible.

I'm certianly not saying speach recognition and computer vision are pointless, I'm merely saying when tech pundits chime on about a future where we talk to computers instead of using a keyboard, I have to say "Slow down there sparky."

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Related Links
o article
o MIT's Project Oxygen
o Microsoft Research
o Also by Anonymous Zero


Display: Sort:
Ubiquitous computing and MIT's Project Oxygen | 53 comments (51 topical, 2 editorial, 0 hidden)
interesting project... (2.00 / 3) (#1)
by hurstdog on Wed Jun 21, 2000 at 03:25:40 PM EST

I think its an interesting project, and definitly has potential to be really neat. I do agree with you though in that it wouldn't work for many applications. One I can think of off the top of my head is coding. It would definitly be a lot slower saying your code than typing, and I almost never type a line correct the first time, speaking it would be even worse. (On a side note, what about people who stutter? Would they be forced to just not use the computers or would there have to be a specialized setup for each person?)

I do think it will be useful in some applications though. None that I can think of right away, but that doesn't mean it won't be useful. Look at some of the older technologies, like the GUI for example. Xerox didn't think it would be useful at all, now look where we are...

-Andrew

Re: interesting project... (4.00 / 1) (#4)
by Eimi on Wed Jun 21, 2000 at 04:28:29 PM EST

I've thought about that as well. I think there are two reasons coding works so much better with a keyboard. One is editting text in general. I have trouble seeing how a voice control system could be anywhere near as easy to use as vi, once you get used to it. The other reason, though, could change pretty easily. That is that programming languages are designed for our keyboards, and so use a lot of symbols. Sure it's easier to type "@_[0]" than to speak it, but if voice did take off, we'd probably start using languages that work better for speach.

The interface gizmo that I'd really like is a webcam-like device to follow my eyes, and move something akin to a mouse cursor based on that. Throw in sloppy focus, and a mouse button on the keyboard, and your hands would never need to leave the board. Mmmmmm.

[ Parent ]

Re: interesting project... (none / 0) (#10)
by fluffy grue on Wed Jun 21, 2000 at 05:11:26 PM EST

Just add this to your .fvwm2rc:

Key k A CS CursorMove 0 -10
Key j A CS CursorMove 0 10
Key h A CS CursorMove -10 0
Key l A CS CursorMove 10 0
Key k A MCS CursorMove 0 -10
Key j A MCS CursorMove 0 10
Key h A MCS CursorMove -10 0
Key l A MCS CursorMove 10 0
Now you have ctrl-shift-vi movement keys to move the cursor, or ctrl-meta-shift-vi movement keys to have a bit more finegrained control. There's no way to get Netscape to respond to artificial MouseClick events, as far as I can tell, so you still need the mouse to do Netscape, and for things like GIMP my Wacom tablet is vastly superior to any conceivable mode of keyboard input (and the mouse, for that matter), but overall, I almost never need to touch the mouse unless I'm webbrowsing or using some other inherently-graphical application. (Everything else I just use the keyboard movement to move between windows and take advantage of SloppyFocus. :)

I've also put in similar stuff to go the other way around, accessing typically keyboard things using the mousewheel. For example, ctrl-alt-mousewheel flips desktops (very nice for when I'm browsing a BUNCH of webpages at once). I don't have the settings with me right now though (I only have that bit setup at home).

Don't have fvwm2 or something fvwm2-based? Sorry. :) Actually, my friend Ross has hacked up tvtwm to add in a LOT of customization (including keyboard-mouse stuff), and he just might get around to releasing his changes someday.
--
"Is not a quine" is not a quine.
I have a master's degree in science!

[ Hug Your Trikuare ]
[ Parent ]

Re: interesting project... (none / 0) (#38)
by drivers on Thu Jun 22, 2000 at 04:29:27 PM EST

Just curious: Did you do that keyboard stuff because of your wrists, or because it's more convenient?

[ Parent ]
Re: interesting project... (none / 0) (#51)
by fluffy grue on Sun Jun 25, 2000 at 01:06:59 AM EST

Both.
--
"Is not a quine" is not a quine.
I have a master's degree in science!

[ Hug Your Trikuare ]
[ Parent ]

Re: interesting project... (none / 0) (#44)
by Eimi on Thu Jun 22, 2000 at 08:10:31 PM EST

If you can't click, what's the point of moving the mouse with the keyboard? Only reason I can think of would be to move among windows (sloppy focus style), but for that you could do

Key h A CS Direction West (!Iconic !Sticky) FocusAndWarp
Key l A CS Direction East (!Iconic !Sticky) FocusAndWarp
Key k A CS Direction North (!Iconic !Sticky) FocusAndWarp
Key j A CS Direction South (!Iconic !Sticky) FocusAndWarp

and move pretty directly. Other thing I do is

Key KP_Home     A       N       GotoPage 0 0
Key KP_Up       A       N       GotoPage 1 0 
Key KP_Page_Up  A       N       GotoPage 2 0
Key KP_Left     A       N       GotoPage 0 1
Key KP_5        A       N       GotoPage 1 1
Key KP_Right    A       N       GotoPage 2 1
Key KP_End      A       N       GotoPage 0 2
Key KP_Down     A       N       GotoPage 1 2
Key KP_Page_Down        A       N       GotoPage 2 2

I keep 9 destops open, and then move among them using the number keypad with numlock off. Very convenient.

[ Parent ]
Re: interesting project... (none / 0) (#52)
by fluffy grue on Sun Jun 25, 2000 at 01:07:50 AM EST

It's useful because I use an XFree86 virtual screen on a much larger desktop (1024x768 display on a 1600x1200 root window). Warping the pointer doesn't work very well if the window happens to be down and right and offscreen.
--
"Is not a quine" is not a quine.
I have a master's degree in science!

[ Hug Your Trikuare ]
[ Parent ]

GUI (none / 0) (#8)
by Neuromancer on Wed Jun 21, 2000 at 04:57:22 PM EST

GUI is useful for some things, not for the things that were popular uses at the time that Xerox PARC developed the GUI. Most things, I would prefer a text prompt. Or perhaps voice activation "computer, emacs girlfriend(underscore)letter"

[ Parent ]
Password your shutdowns... (2.33 / 3) (#2)
by genehack on Wed Jun 21, 2000 at 03:30:21 PM EST

(Forgot to post this when I voted...)

In Head Crash, the lead character mentions how the shutdown process on his voice-operated computer is password-protected, because when he was in college, a frat sent pledges through the library during finals week, having them shout "SHUTDOWN! SHUTDOWN!".

Voice operation is potentially a Very Good Thing for people who can't use keyboards or mice, but it's only a Very Cool But Not Very Practical Thing for those of us who can.

Re: Password your shutdowns... (none / 0) (#27)
by jabber on Thu Jun 22, 2000 at 10:04:25 AM EST

So you have to speak the password to then say SHUTDOWN?

Or do you have to type the password? If so, what's the point of having to then say SHUTDOWN.

Makes most sense to just make the sensitive tasks hands-on, and the safe tasks vocal.

[TINK5C] |"Is K5 my kapusta intellectual teddy bear?"| "Yes"
[ Parent ]

Re: Password your shutdowns... (none / 0) (#30)
by genehack on Thu Jun 22, 2000 at 10:55:16 AM EST

IIRC, in the book, there were a couple of requirements (for the passworded shutdown, that is)

  • It had to be above a certain dB level
  • There was some other key phrase that had to be said

As far as "just make the sensitive tasks hands-on, and the safe tasks vocal", I don't think these boxen had keyboards.

If you haven't read the book, it's worth a go -- Head Crash, by Bruce Betkhe(sp?) -- it's sort of a send-up of the cyberpunk genre.

[ Parent ]

I respectfully disagree (2.50 / 2) (#6)
by Neuromancer on Wed Jun 21, 2000 at 04:53:42 PM EST

I have worked on a few projects out of an interest for ubiqitous computing, and I think that it will be there eventually. It will not replace the desktop, but it WILL be alongside it. In some cases, it will replace it. I'm never going to hammer out code by yacking at a computer, but I will make phone calls without touching my phone, browse the net, make queries, things that most people who "don't really know anything about computers" do.

Re: I respectfully disagree (none / 0) (#13)
by w3woody on Wed Jun 21, 2000 at 06:15:13 PM EST

I have worked on a few projects out of an interest for ubiqitous computing, and I think that it will be there eventually. It will not replace the desktop, but it WILL be alongside it.

Well, in one sense, we already have ubiqitous computing. In my living room, for example, there are 8 different computers, all tasked to a specific task: one in my television set, one in my DVD player, one in my laserdisk player, one in my VCR, and three remote controls.

I think the problem here is a misunderstanding of how "ubiquitous" computing may work in the real world. To me, "ubiquitous" means that a computing system is used because it's cheaper, easier, and better to use a computer system in place of an analog or mechanical device. And to me, that means embedded, not some wall unit that looks like it belongs on the set of Star Trek.



[ Parent ]
Check out Hive (1.50 / 2) (#7)
by mebreathing on Wed Jun 21, 2000 at 04:57:18 PM EST

Hive is a framework for ubiquitous computing, also out of MIT. Check out http://www.hivecell.net/

I'm not sure I like this (4.00 / 3) (#9)
by feline on Wed Jun 21, 2000 at 05:02:35 PM EST

from MIT Oxygen Project home:

'Instead, "anonymous" devices, either handheld or embedded in the environment, will bring computation to us, no matter where we are or in what circumstances.'

Aren't they the least bit concerned with abuse of these 'anonymous' devices floating around in the environment that can have quite a bit of control over other people, esspeccially since these devices can be used by anyone?

'We will not need to type or click, nor to learn computer jargon.'

I don't see what would be any fun about this, what makes _us_ special? :(

'The Oxygen system must be:... * embedded--it must live in our world, sensing and affecting it...

Again, what about the abuse issue, if it's actually embedded with our environment, and it affects it directly, I mean, what could someone be capable of doing?

It seems that their whole idea is based on trust of everyone, which is just not logical with humans. Some people are going to get kicks out of hurting others, some are going to get desperate and see if they can snake a bit of dough out of the system, I'm just saying that this can't work unless they come up with some kind of security measure to prevent this, which would, in my oppinion require manipulating every human brain to be good, which is just impossible.

Brain wave generating chips in every box of corn flakes?
------------------------------------------

'Hello sir, you don't look like someone who satisfies his wife.'

Re: I'm not sure I like this (5.00 / 3) (#26)
by jabber on Thu Jun 22, 2000 at 10:01:24 AM EST

/*Aren't they the least bit concerned with abuse of these 'anonymous' devices floating around in the environment that can have quite a bit of control over other people, esspeccially since these devices can be used by anyone?*/

Anonymous devices are the key to nomadic computing, and that is, right or wrong, the future. Rather than thinking of these things as PC's that "are not yours", consider them publically accessible on the user level only. They are part of the information processing infrastructure - just like ATMachines, public phones, those dinky little photo booths, automated toll gates on the highway, pay-at-the-pump gas stations... What are your feelings about those?

The thing is that this new idea isn't any different, really, it's just more of the same, and interconnected.

Imagine being able to make a withdrawal from your bank at any pay-phone. Imagine being able to talk to a live teller (or AI, in any case something complex) at an ATM. Imagine getting into your car after a hard day at work, and having the house know that it's time to turn on the AC in the summer and heaters in the winter. Imagine a self-diagnosing car that, in the event of a brakedown, will inform your mechanic with the nature of the problem, request service, inform your destination that you will be late, all before coming to a slow and safe automatic stop at a safe area of highway.

/*I don't see what would be any fun about this, what makes _us_ special?*/

"US" being those trained in computer usage, right? Computers have a steep learing curve for most people. "We" happen to understand them, and enjoy the benefits of being special because of it. This is the same issue that keeps rising up between the PC and the Mac crowds. The PC let's "Us" tinker and experiment and learn obscure tricks and magical commands - Unix is even more like this. Macintosh on the other hand abstracts you from the workings of the computer as much as possible. It makes the computer into a tool, and the result is that people tend to focus on the product they are trying to produce rather than on the means of producing it.

This view is very valid for all computer users (not professional computer people, but end users). They have no interest in HOW the computer does it's job; only THAT is do it, and in the least complicated way possible. Natural language communication is the Holy Grail here - with the more realistic goal being an intuitive set of metaphors.

CLI sucks as a metaphor. GUI is closer to reality, since we're very visually oriented as a species. MultiMediaUI is even better because little auditory clues are a good thing. The point is that we're looking at this Ubiquitous/Embedded/Pervasive computing approach with bias - we're professionals after all, and we WILL take apart the black box to see how it really works. But the general public couldn't give a toss. Just as I do not much care about things unrelated to figuring out how things work. :)

/*their whole idea is based on trust of everyone, which is just not logical with humans. Some people are going to get kicks out of hurting others, some are going to get desperate...*/

The ATM again. Devices can be made such that any tampering results in self-destruction of the usefulness of the device. Ever try to crack open a SecureID card to get at the ICs? Once you crack the seal, it's dead hardware.

There are valid arguments agains pureply "human communication" I/O to machines. But most only apply to those of "us" who deal with computer related abstractions and specialized applications. You can't SPEAK CODE into JBuilder effectively. You can not express verbally the intention to add a little more fuschia to the background of your image.

For graphic artists, using real world tools with I/O to a computer is the answer. Programs such as Fractal Designer are a nice step in the right direction, as are intelligent drawing pads (maybe with an LCD underlay of the image) which are sensitive to a pen vs charcoal vs brushes... Interesting subject in itself.

Coders are still searching for a 'better' metaphor than the keyboard and mouse. We may not find it, but then again we may. Some sort of a VR representation of basic constructs and functionality perhaps... I very much liked the Matrix imagery of Neuromancer. I think that being able to come up with personalized abstractions might be the key... Remember the sequence where the (Kuang?) virus is penetrating the AI defense, and delivering Case into the belly of the system? The virus is presented as a shark, which on contact seems to freeze and very slowly creep past the AI's defenses. While I read that, I thought... Hmmm, a slow, unpatterned portscan and exploit search.. Slow enough, and random enough to not show up as anything unusual. See where I'm going? It's imagery that worked for the idea being handled; just as the scene in The Matrix, where Neo is quite literally 'bugged'.

I'm sure there are real world metaphors for what we do at the CLI and in code. Hell, sometimes the concept is named just so, to suggest the metaphor. Unix sockets for example.

Ok, enough rambling. I should go back to work now. :)

[TINK5C] |"Is K5 my kapusta intellectual teddy bear?"| "Yes"
[ Parent ]

Re: I'm not sure I like this (none / 0) (#49)
by feline on Sat Jun 24, 2000 at 12:29:17 AM EST

'Rather than thinking of these things as PC's that "are not yours", consider them publically accessible on the user level only. '

You've never heard of end-user level access of things leading to abuse? Phone phreaking, 'Cracked!'. The former is just a harmless fun game that we used to play (damn these digital phone systems!), but the latter, look at the devistation that this caused, and this was just bits flying around different transistors and wires, but, with oxygen, immagine the power someone could have, and and and, a viral effect might possibly take effect.

My thoughts on these devices:

ATMachines-these are heavily secured, and users of these are tracked and even video taped, are we going to put a system like oxygen in a police state, remember the power that these itsy bitsy machines will have, the amount of people that will be influenced?

public phones-these are vandelized, I even read a story about how someone set up a whole system of other people's phone lines (grey boxing) and hooked those into a home-made switch board, and then went so far as to hook a cordless phone into a payphone across the street and listn to other peoples calls. The story was false (it said so at the end), but it was all very plausable (I was thinking seriously of trying it myself ;) ).

automated toll gates-in my city, people routinly go through the man-manned without paying (they have a weird system of leaving the gate up and putting a cone there when no one is there, I guess they'd rather have a cone moved, and loose a few bucks than to have someone trash their facilities).

'...having the house know that it's time to turn on the AC in the summer and heaters in the winter...'

They already have systems like this, but I see your point, having your car start up in the morning right before you leave, but only if it sees that you got up, perhaps by seeing if the shower was run for ten minutes that morning.

'Ever try to crack open a SecureID card to get at the ICs? Once you crack the seal, it's dead hardware. '

I've never heard of the SecureID, got a url for that?
------------------------------------------

'Hello sir, you don't look like someone who satisfies his wife.'
[ Parent ]

Think STAR TREK(tm) (4.20 / 4) (#11)
by Tin-Man on Wed Jun 21, 2000 at 05:23:32 PM EST

You have to think Star Trek in order to understand pervasive computing.

By the time we have computers capable of 100% accurate speech recognition, we should also have been able to work out other things, such as better AI, improved security based on biometrics, increased privacy controls embedded into the OS, etc.

And remember, for all the voice-recognition capabilities of the Enterprise, there were still little consoles where the crew members actually worked. Pervasive computing isn't about making computers disappearing; it's about making computers more integrated into the environment, rather than a beige box we clunk onto our desktop.
--
The future sure isn't what it used to be!

Re: Think STAR TREK(tm) (3.00 / 1) (#14)
by nuntius on Wed Jun 21, 2000 at 06:18:16 PM EST

Yeah, no. I've always kept quiet on this one, but the integrated systems on the Enterprise would have been the most beautiful security exploit.

Kirk: Delay the enemy, Spock's going to try this new missile thingie. (thingie is a technical term)
Klingon: Captain, did you hear that? Stop playing with them!

Ok, so it didn't happen in any of the movies. However, take the relative difficulty of planting good bugs in today's home. Small size means poor microphone reception, etc... Now think in terms of an already wired house. Big, well placed microphones, which a simple security exploit allows you to access across the Internet.

Legal precedent would quickly give law enforcement the authority to tap these active grids, and insecure computers would grant access to everyone else.

--Its not paranoia if they actually are after you. (I forget who that's from)

[ Parent ]
Re: Klingon bugs (4.00 / 1) (#16)
by Tin-Man on Wed Jun 21, 2000 at 06:55:26 PM EST

You're assuming that the Klingons were able to get on board the ship to plant a bug, and that it could transmit a signal without being detected. Surely the Enterprise wouldn't transmit that to the enemy of its own accord.

Taking this back into the not quite so distant future, given the proper advances with security and privacy controls embedded into the OS, much like current security controls for file permissions, your software would not forward information without proper permission. So even though a house is completely wired and connected to an advanced network, none of the owner's devices would transmit that information. You'd still have to get inside and plant your own bugs. And they can't simply tap into an already (properly) configured network without being detected.

Obviously, the technology that we have today isn't ready for either the technical aspects, nor the legal/moral aspects of pervasive computing. But it is coming. There will be growing pains as companies rush to market and forget about or ignore the security and privacy issues. But much like security problems today are fixed in robust operating systems, we will learn from our mistakes and make corrections.

We already have technology that can detect if someone is tapping into the middle of a communications channel. Other innovations will follow. It's only a matter of time. Just as mainframe geeks couldn't stop the microcomputers from coming, and microcomputer devotees couldn't stop personal computers from taking over, and now as IT managers are resigned to figuring out how to deal with all the Palm computers, pervasive computing is coming, and, short of global holocaust, there's nothing we can do to stop it. All we can do is try to work out the potential problems.
--
The future sure isn't what it used to be!
[ Parent ]

Re: Klingon bugs (none / 0) (#17)
by Paradox on Wed Jun 21, 2000 at 07:26:30 PM EST

Well, I agree pervasive computing is the Next Big Thing(tm).. however, I don't like the idea of microphones distributed about the house like that. I have a speech pickup on an old macintosh, and it listens to me sometimes. I think that's cool. But if the government decides that it's legal to TAP these over the internet (and with DSL, cable and fiber running to more and more homes it is not only feasible, it's likely!) then you can't do a damn thing about it. Sure, maybe you can figure out they ARE listening, but whatcha gonna do about it? Tell them to piss off? Basically, we need to be really careful and realize that speech would be good for small, common commands ("Turn on. Check my mail. Make coffee." (a la serial experiments lain or something similar) but that's about as far as I think I want it to go, unless I can get garuntees the next "I love you" wont be a voice pickup hack. :)
Dave "Paradox" Fayram

print print join q( ), split(q,q,,reverse qq;#qsti
qq)\;qlre;.q.pqevolqiqdog.);#1 reason to grin at Perl
print "\n";
[ Parent ]
microphones vs. speakers (none / 0) (#28)
by Anonymous 242 on Thu Jun 22, 2000 at 10:21:16 AM EST

Just think. The basic hardware design of speaker allows it to also be used for a microphone. With current hardware this is not such a big deal, but as technology progresses and more and more generic processors that can evolve or adapt to the task at hand are designed and produced, the ability of anyone with the proper knowledge to record your every sound should be taken for granted.

Read Graham Watkins' thriller Virus. While I felt like the end was sort of a cop out, the ideas presented have the ability to scare the living daylights out of your average computer programmer. If Penny can do this, so can Big Brother.



[ Parent ]
Perl (none / 0) (#34)
by Anonymous Hero on Thu Jun 22, 2000 at 01:33:00 PM EST

An optimisation:

perl -e 'print "god i love perl \;)\nits #1\n";'

[ Parent ]

Re: Perl (none / 0) (#36)
by Paradox on Thu Jun 22, 2000 at 02:42:08 PM EST

Haha.
Sorry, but I have to leave it obfuscated. It's just too much fun in that state.

That's not even my newest sig. This one looks pretty tame by comparison.
Dave "Paradox" Fayram

print print join q( ), split(q,q,,reverse qq;#qsti
qq)\;qlre;.q.pqevolqiqdog.);#1 reason to grin at Perl
print "\n";
[ Parent ]
Re: Wire taps (none / 0) (#35)
by Tin-Man on Thu Jun 22, 2000 at 02:35:33 PM EST

if the government decides that it's legal to TAP these over the internet (and with DSL, cable and fiber running to more and more homes it is not only feasible, it's likely!) then you can't do a damn thing about it. Sure, maybe you can figure out they ARE listening, but whatcha gonna do about it? Tell them to piss off?

That's exactly what you'd have to do about it. Currently, our government has laws against spying on its own people. In order to spy, there has to be an order from a (hopefully) objective judge who decides that there is enough evidence that an individual is engaging in wrongful behavior that it warrants such spying. So if you are a bad guy, you make sure your lines are clean before transmitting, and if you're not a bad guy, you get alerted when you are being spied on, and you do something about it. Even if it only amounts to "telling them to piss off." Under current law, you can probably get a bunch of money from a lawsuit like that.

So if government passes a law that gives them the right to tap your private communications, you do the same thing, and encourage others to join you. I think (or, I hope?) the current tide of public opinion is gathering strength and soon will be able to crush those who oppose allowing private citizens their privacy.

I also agree with your statements about speech recognition being mostly good for simple commands. At least for the next few years, our AI won't be good for anything more. But, as always, we will continue to build upon our current technology, ever in pursuit of that next Big Thing.
--
The future sure isn't what it used to be!
[ Parent ]

this happened in multiple episodes and books (none / 0) (#33)
by Anonymous 242 on Thu Jun 22, 2000 at 11:50:57 AM EST

The ones that come immediately to mind is the original start trek episode where the enterprise tapped into the system of a romulan vessel to see what was going on and the book where the enterprise computer fell prey to a practical joke trojan of some sort that was fixed by transmitting the predatory routine to a klingon vessel.

I believe that there have been other episodes where (1) either the computer itself was the enemy through something bad happened (especially in the books) or (2) the enemy some how gets control of at least portions of the computer. Did anyone else think that the second next-generation episode with the Moriarity AI copped out at the end when transporting Moriarity into a virtual universe?



[ Parent ]
Re: Think STAR TREK(tm) (3.00 / 1) (#25)
by Anonymous 242 on Thu Jun 22, 2000 at 09:56:36 AM EST

By the time we have computers capable of 100% accurate speech recognition, we should also have been able to work out other things, such as better AI, improved security based on biometrics, increased privacy controls embedded into the OS, etc.

Exactly

Look at this portion of BeOpen's interview with Andy Hertzfeld at http://www.gnulinux.com/interviews/hertzfeld_part2.shtml. Andy Hertzfeld was one of the designers of the original Mac UI. Basically what he says in the interview is that vr (voice recognition) is basically a solved problem. The problem with vr's usability is with language recognition. Great, the computer can turn your words into bits. The challenge of turning vr into an interface more usable than what we currently have is not simply turning the sound waves into bits in an accurate and consistant manner. The challenge is to transcend the keyboard/mouse paradigm that all current vr systems use and develop a ui that is intuitive and easy to use based around natural language. I have my doubts about the solvability of that problem within the current and forseeable generations of hard ware.



[ Parent ]
Re: Think STAR TREK(tm) (none / 0) (#29)
by YellowBook on Thu Jun 22, 2000 at 10:28:55 AM EST

That's very interesting. I think it would be useful to start off by making a more linguistically intelligent CLI and then moving to voice (which is an easier step). I envision the first iteration of this working something like an Infocom parser...

lingsh$  for each file whose name ends in ".mp3"
>> go on... >> apply oldstylecommand 'mpg123 -b 1024' to it
>> go on... >>

Okay, I'm applying an old-style unix command (mpg123)
to the following files in sequence:
rms_free_software_song.mp3
free_software_song_techno_remix.mp3
...

Okay.
lingsh$ look at files

There are very many files here, mostly with names ending in .mp3.  There is a list of files here.

linghs$ read the list of files

<unix-style ls -l listing>

lingsh$

Or is that silly...



[ Parent ]
Re: Think STAR TREK(tm) (none / 0) (#40)
by naasking on Thu Jun 22, 2000 at 04:45:49 PM EST

This is the problem that most people fall into when considering doing new interface paradigms. They're still stuck with old ways of thinking. They try and extend the old paradigm to do something for which it wasn't intended or they try and shape the new idea to look like the old. This is a serious problem.

Why would you want to manipulate your computer on a file by file basis through voice control? It's obvious that doing this by dictation is very cumbersome. Handling files is much quicker and easier with a GUI or CLI.

The Voice interface would have to be stuctured much differently IMO. Users of such an interface want to be able to manipulate and access information in the computer, not files. They don't want to have to worry about file organization like in your example.

The intuitiveness of the Voice interface is that you should be able to just tell the computer to "play the rms_free_software_song" and it would take care of the rest. It would go to where it stores its music files, find the right song and play it according to the format it's stored in.

If it doesn't find it locally, you should be able to tell the computer to go and get the song, or at the least say contact this server and search for rms_free_software_song.mp3, then get the appropriate one.

Users of speech based interfaces want to be able to say, "open the 2000 Q2 financial results document" or at least "open document 'Q2_financial_2000'," and not, "go to folder 'this/blah/another/thing/finally/here' and open document 'Q4_financial_2000.doc' in Word" or anything like that. It's the big difference between handling files and information.

To be honest, I'm probably getting it wrong too. There are guaranteed to be a better way to do things than what I've suggested, but I think it's at least more intuitive than your suggestion.

The main article itself is flawed in this respect as well. Why in god's name would you use one interface to control another? Do you see people typing commands into a CLI by clicking with a mouse on the keys on an on-screen GUI keyboard? No. Why? It's prone to error, slow and makes absolutely no sense. Just like typing is for a keyboard not for the mouse, so you would never use a speech interface to control a GUI. It's just dumb.

Voice control is it's own unique interface and requires it's own ideas, abstractions and approach. Don't try and apply other abstractions and ideas to it; don't compare apples and Ferrarri's, please. Sure they're both red, but isn't obvious that they're each in a class of their own?



[ Parent ]
Re: Think STAR TREK(tm) (5.00 / 1) (#41)
by naasking on Thu Jun 22, 2000 at 04:45:59 PM EST

This is the problem that most people fall into when considering doing new interface paradigms. They're still stuck with old ways of thinking. They try and extend the old paradigm to do something for which it wasn't intended or they try and shape the new idea to look like the old. This is a serious problem.

Why would you want to manipulate your computer on a file by file basis through voice control? It's obvious that doing this by dictation is very cumbersome. Handling files is much quicker and easier with a GUI or CLI.

The Voice interface would have to be stuctured much differently IMO. Users of such an interface want to be able to manipulate and access information in the computer, not files. They don't want to have to worry about file organization like in your example.

The intuitiveness of the Voice interface is that you should be able to just tell the computer to "play the rms_free_software_song" and it would take care of the rest. It would go to where it stores its music files, find the right song and play it according to the format it's stored in.

If it doesn't find it locally, you should be able to tell the computer to go and get the song, or at the least say contact this server and search for rms_free_software_song.mp3, then get the appropriate one.

Users of speech based interfaces want to be able to say, "open the 2000 Q2 financial results document" or at least "open document 'Q2_financial_2000'," and not, "go to folder 'this/blah/another/thing/finally/here' and open document 'Q4_financial_2000.doc' in Word" or anything like that. It's the big difference between handling files and information.

To be honest, I'm probably getting it wrong too. There are guaranteed to be a better way to do things than what I've suggested, but I think it's at least more intuitive than your suggestion.

The main article itself is flawed in this respect as well. Why in god's name would you use one interface to control another? Do you see people typing commands into a CLI by clicking with a mouse on the keys on an on-screen GUI keyboard? No. Why? It's prone to error, slow and makes absolutely no sense. Just like typing is for a keyboard not for the mouse, so you would never use a speech interface to control a GUI. It's just dumb.

Voice control is it's own unique interface and requires it's own ideas, abstractions and approach. Don't try and apply other abstractions and ideas to it; don't compare apples and Ferrarri's, please. Sure they're both red, but isn't obvious that they're each in a class of their own?



[ Parent ]
talking to your computer (2.00 / 1) (#12)
by jetpack on Wed Jun 21, 2000 at 06:10:13 PM EST

Can you imagine how your coworkers would feel about having to listen to you babble at your computer all day?

Irrelevant. I already do that. I spent about15 minutes swearing at an HP series 700 while it booted just a few days ago. And I often verbally abuse my machine when my code is doing something particularly strange that it shouldn't. My co-workers think I'm nuts, but at least they are used to it :)


--
/* The beatings will continue until morale improves */

oh come on! (none / 0) (#15)
by feline on Wed Jun 21, 2000 at 06:36:30 PM EST

that's nothing, have you ever been at a party where someone says something funny and you blurt out 'L-O-L!' ? ;)
------------------------------------------

'Hello sir, you don't look like someone who satisfies his wife.'
[ Parent ]

Re: talking to your computer (none / 0) (#45)
by Eimi on Thu Jun 22, 2000 at 08:14:45 PM EST

That's exactly why I don't want voice recognition. I'd much rather my systems not know how much I'm swearing at them and complaining about them behind their backs.

[ Parent ]
Talking to my Macintosh (2.00 / 1) (#18)
by jdigital on Wed Jun 21, 2000 at 07:57:41 PM EST

I while back I fell for the marketing hype over Macs, and left my comfort in *nix to buy a G3. The machine was terribly overpriced, underspeced, and after a few weeks I sold it for under half price.
But, there was some joy I got from it - namely the joke interface. :)
It is possible to set up on a base install of MacOS 8/9 a system whereby you can talk to your computer, and it will respond (some of the time) to spoken commands such as "Open Netscape", or "Minimize Window".
As mentioned in the main article, these commands were absolutely useless, with half of the time the Mac doing the exact opposite of what I wanted.
The only thing I could get working was asking the computer to "Tell me a Joke", to which it would respond "Knock Knock", after which is would reply with a various range of stupid jokes.
eg:

-Knock Knock
+Who's There
-Orange
+Orange Who
-Orange you glad you bought a Macintosh




The RAW chip (2.50 / 2) (#19)
by royh on Thu Jun 22, 2000 at 02:28:24 AM EST

I agree with you; the Oxygen project seems to me like a lot of fluff, but there is one thing that I find very very interesting: their raw architecture.

The raw chip is kinda like taking the simplest possible processor you could make and parallelizing it. The ideas (IIRC) are:

1: Hardware should be minimized. Compilers will be able to handle the optimization, and will do a better job of it.
2: You can just buy grids of the stuff and tack it onto your existing grid to upgrade.

(This is the idea - I do know that the architecture/instruction set is not as excrutiatingly simple as it could be - practical problems I think).

Anyways, I really think, once software really becomes a mature science (encapsulation up the wazoo; bug free code (no I am not joking)), we'll see a lot more generalized hardware like this.

easily the coolest part of Oxygen.

ridiculous? (2.00 / 1) (#20)
by Dacta on Thu Jun 22, 2000 at 02:47:08 AM EST

Computers without keyboard and mice is fine by me, are we ever going to stop using keyboards and mice? No, please. That's ridiculous.

Well, I'm not sure about voice interfaces, but I don't think it is at all ridiculous that someone will come up with a better idea than a keybord & mouse.

(We've only been using mice for 15 years, really, anyway.)

Think of Palms and the like - I think there is potential there for a scaled up & more powerful version of graffiti sutible for desktops.

palm like graffitte imput interface (none / 0) (#21)
by feline on Thu Jun 22, 2000 at 04:53:13 AM EST

'...I think there is potential there for a scaled up & more powerful version of grafitti suitable for desktops...'

I really think that the graffitti interface is a bit unsuitable for getting real work done. Have you actually used a palm, it takes an hour to sit down and learn how to write the letters so the machine'll understand what you're writing, and then you can never get above 20 words per second, it's just a pain in the ass to use, and much too slow for getting stuff done.

The only reason (IMO) that the graffiti interface took of with the palm, is that it is the only interface that can fit in that limited space wihout cramping the user.

If this were implemented with the oxygen system, how would it work, would you just carry around a plastic tablet or would there be communcal tablets or could there just be a gyro-stick?

Besides, oxygen's aim is to make little machines be able to fly around and interface directly with the human mind and the physical environment, all in all, a grafitti interface is much to clunky.
------------------------------------------

'Hello sir, you don't look like someone who satisfies his wife.'
[ Parent ]

20 words per second!!!! Who taught you to type? (none / 0) (#24)
by Anonymous 242 on Thu Jun 22, 2000 at 09:42:40 AM EST

I really think that the graffitti interface is a bit unsuitable for getting real work done. Have you actually used a palm, it takes an hour to sit down and learn how to write the letters so the machine'll understand what you're writing, and then you can never get above 20 words per second, it's just a pain in the ass to use, and much too slow for getting stuff done.

First off, I bet you meant to say 20 words per minute. Though, I could be wrong.

Second, given that it takes an hour or so to learn grafitti, most people will be faster at grafitti than they would be at touch-typing, given the same amount of training.

I know people writing entire essays and books on their Palms. Grafitti is uncannily intuitive. I have friends that fear computers, but given a Palm they go to town. Hand writing is a much more intuitive interface than typing.

Of course, there will always be room for typists. Someone with the motivation to sit down and learn touch typing and the motivation to constantly push themselves to get faster and faster will be able to input at much higher speed than a grafitist. However, this is the realm of the power user and of the specialized worker. For the vast majority of people, grafitti is not only sufficient, but is also a qualitatively better ui than a keyboard.



[ Parent ]
Re: 20 words per second!!!! Who taught you to type (3.00 / 1) (#46)
by feline on Thu Jun 22, 2000 at 10:07:16 PM EST

'Hand writing is much more intuitive interface than typing.'

I don't know about y'all, but I type a whole lot faster than I can write, and when I do write, my hand always ends up hurting. I think writing is a bit of a 'legacy' system, and not fit to go into such an innovative sounding technology such as Oxygen.
------------------------------------------

'Hello sir, you don't look like someone who satisfies his wife.'
[ Parent ]

I don't think that you've grasped my point. (2.00 / 1) (#48)
by Anonymous 242 on Fri Jun 23, 2000 at 04:31:04 PM EST

To reiterate:
Second, given that it takes an hour or so to learn grafitti, most people will be faster at grafitti than they would be at touch-typing, given the same amount of training.

Typing is a specialized skill. Do I type faster than I can write? You bet! Years of working with computers with keyboards has given me enough practice so that I can type characters far faster than I can pen the same characters. But how many people do the computer thing for a living? The vast majority of jobs out there do not require facing a terminal with a keyboard all day. Mechanics, wait staff at restaurants, pilots, taxi drivers, delivery personel, and more are much better off with a more intuitive interface than the keyboard. Even home users and students are likely to be more productive with a stylus on a writing pad than sitting at a keyboard. The typical computer user sucks at typing.

Faster is not always better. A Formula One race car can go much, much faster than a station-wagon, but which is better for the average driver? Does this mean that there is no place for fast cars? Certainly not, but sedans and wagons are much more ubiquitous than Formula Ones, for good reason.



[ Parent ]
Yes! (none / 0) (#23)
by Anonymous 242 on Thu Jun 22, 2000 at 09:33:08 AM EST

I would really like to see a grafitti style interface for desktops. I know a tremendous number of people that can have grafiti speeds far higher than their typing rate. I think the only thing that astonishes me more than the number of computer users that can't touch type is the speeds at which some people can put down grafiti. And grafiti has the advantage of being much more intuitive. Most (not all, I certainly prefer typing) people would much rather write than type.

I would think that such a thing would be relatively easy to implement. I already use a touch-pad instead of a mouse. If PC's were as easy to use as palms, they would instantaneously become much more popular. But as it is, all the interfaces are unweildy and unintuitive (with the possible exceptions of the mouse and the graphics tablet).

I would also like to see a pointing device that works by tracking where I'm looking on screen.

But, the interface that I'm waiting for is a one-way neural interface where I can think the words I want to appear on screen. Such an interface would blow voice recognition out of the water and with the advances in neuro-science, I don't think its that far off. But the thing is, its gotta be one way. I don't trust anyone's code (free software, open-souce, or proprietary) running in my head.



[ Parent ]
Why? (none / 0) (#22)
by error 404 on Thu Jun 22, 2000 at 09:32:32 AM EST

Buckminster Fuller voted i on this story

The site doesn't seem to go into that, other than a couple of application examples that seem to be almost afterthoughts.

The project appears to be driven by a vision of the future, rather than the needs or desires of the people served. But then it doesn't go into the question of how that future world will be better than what we have now. I'd be impressed with the level of thought if the authors also described how it would be worse in some ways. But it doesn't, it just says this is how it must be. Now I could see an argument that the envisioned future is inevitable, but hten the project is redundant.

Anyway, cool possibilities, but I won't be doing much to advance the project.
..................................
Electrical banana is bound to be the very next phase
- Donovan

Re: your rant. Short sighted and ignorant (1.00 / 1) (#31)
by Anonymous Hero on Thu Jun 22, 2000 at 11:01:44 AM EST

There has been immense disscusion regarding ubiquitous computing over the last several years. Never has anyone said, "Ubiquitous computing technologies will completely replace current standard interfaces". I dont believe that has ever been the goal. Rather, researches are trying to develop all pervasive intelligent environments that will recognize and accomadate our needs and wants. You begin your rant with, "But do people really want to talk to the walls like they do on Star Trek? I don't think so." Ok Sherlock, whatever. Even in Star Trek people use keyboard like input devices (stylus as well), I always presumed for privacy reasons. The fact is, we will have keyboard and mice until we can directly (and securly) command computational devices with our very thoughts. Your rant leads me to deduce that this article must surely be your first introduction to all pervasive ubiquitous computing. I would very much like to further explain why your little rant is flawed, but dont have the time(posting from work). I suggest you do a little research, it will open you eyes to one of humanities many possible, and more grand futures. -John BTW- My spelling and grammar suck, I know.

User Interface (none / 0) (#32)
by Anonymous Hero on Thu Jun 22, 2000 at 11:21:19 AM EST

I think that too many people involved in ubiquitous/wearable computing research get caught up in the "isn't this awesome I can talk to my <fill in the blank here>" or "look I can play Quake while watching a movie on the couch with my girlfriend" By getting caught up in the whizz-bang niftiness of it all they miss the original point of all these electronic devices that surround us these days, they're supposed to improve/simplify our lives not just be "cool" In the development of these types of computers the most important paradigm to develop is not the technology that allows you to speak to a computer, nor the technology that allows a bunch of small computers in walls to be networked together, nor any other technology really, what needs to be developed first and foremost are the paradigms for a user interface to these technologies. The example given in the initial article explains this well. Speech technology will never work in a library/office/shared computer lab setting where multiple voices will conflict with one another, the keyboard may seem (and is in some ways) clunky but it does have the distinct ability to make the machine know that a user is communicating with it and not a coworker or the cofee machine etc. This is of course not an argument for the keyboard, but rather an argument for the one-to-one input to machine relationship that the keyboard enables. A keyboard is a terrible user interface, for example for an onboard car audio system. Here a small touchpad or even voice recognition are useful since the most important user interface paradigm is hands free operation. Like anything else the people who design these systems need to ask not "how cool is this" but "would my mother want to use this" before they start shoving yet another silicon wonder drug down our throats.

Re: User Interface (none / 0) (#53)
by Anonymous Hero on Sun Jun 25, 2000 at 11:03:15 PM EST

Good point! but i must make fun of your example: "'I think that too many people involved in ubiquitous/wearable computing research get caught up in the "isn't this awesome I can talk to my <fill in the blank here>" or "look I can play Quake while watching a movie on the couch with my girlfriend"'" BUT WHO WOULD PLAY QUAKE WHILE WATCHING A MOVIE WITH HIS/HER GIRLFRIEND! THAT GUY/GAL IS A MORON!

[ Parent ]
the really interesting thing (none / 0) (#37)
by eries on Thu Jun 22, 2000 at 03:56:42 PM EST

I actually think the really interesting thing is the communications aspect of all this. I hate having to give out twelve forms of contact info for people who want to reach me. I can't even fit them all on one business card anymore (luckily I can easily get more for free).

Oxygen offers the opportunity for a single point of contact. Your smart agent calls up my smart agent, and they negotiate for the best way to get ahold of me. If you're a hot model, it'll take you to my private cell phone. If you're my lawyer, you can get my fax. If you're a spammer, you can go straight to /dev/null.
Promoting open-source OO code reuse on the web: the Enzyme open-source project

Alright. (none / 0) (#39)
by 3than on Thu Jun 22, 2000 at 04:31:58 PM EST

OK-here's where I attack it. This whole thing sounds pretty cool in a way. And in another way, it's not really cool at all.
Let's face it-their 'vision' of the future isn't anything new. Everybody's said it already-Star Trek.
But remember that scene in Star Trek 4 when Scotty gave the engineering firm the info to make transparent aluminum? When the 20th-century guy pointed at the mouse, Scotty gave it an indulgent smile and said 'how quaint,' thinking it was a voice interface device.
After he figured out what it was, he input everything on the keyboard. Really fast.
Let's face it folks-for a lot of things, the mouse actually slows the interface down. I've slowly been learning keyboard shortcuts for GUI stuff because I've realized the amount of time that mousing takes. And let's face another fact-GUI's have a lower learning curve, but they are a lot slower for many, many tasks.
Voice recognition is even slower. By far. I type many more words a minute than I can say. Now, it's because I've learned to type. Let's imagine in the future that voice rec takes over, and people no longer routinely learn to type. Their relationship to info has slowed very considerably. There's no way that typing as a skill will be eliminated by voice rec. Maybe by direct brain connection...
So voice rec is a little fluffy. And I sure don't want it listening to everything I say. I hope that at the very least, there's an entry word. What I'd like right now is this: I say, 'console' and get voice-driven bash that will read me my email, etc, and I can do stuff while watching tv or whatever. But that really doesn't need MIT's resources. A bunch of 15-year-old perl hackers could do that well enough. That's really my only problem with the project-it seems like an over-appropriation of resources. Everything that they want to do is possible with traditional computing architecture. All it really entails is a user-friendly shell and lcd's to be cheaply available. I hope that their 'vision' is just a starting point-I think that MIT's vast resources should be able to come up with something better and newer. And most of all, I hope that they have the common sense to make their products freely available. Imagine this project as an insular, inbred, overly academic exploit. Now imagine it as an open-source project with daily downloadable releases and tons of user feedback. I hope that someone at MIT realizes how important users are when designing a UI.
A project like this has great potential to be a good thing for everyone-as long as it's open source. If not-well, they're just talking to themselves, really.

Are we really slowing down interaction? (Re: Alrig (none / 0) (#43)
by GoRK on Thu Jun 22, 2000 at 05:26:00 PM EST

Given:
Voice recognition is slower than mousing.
Mousing is slower than typing.

Assume:
Typing is slower than the machine instinctively knowing what you want it to do.

What is the definition of speed? When people get in a room and communicate, a whole lot of information gets exchanged. Would the communication between two people in the same room be faster if each person typed and e-mailed their ideas, or would it be slower? The idea with computers (currently) is that you are given an interface which you may operate with only one apparatus at any given time. In sci-fi like star-trek, this is much the same -- only you are correct -- the mechanisms for human/computer interaction are more inefficient. The only thing that really makes it seem more efficient is that it takes LESS input (more automation) to accomplish things in Star Trek systems vs. the systems of today. Thus, it is a progression in systems design while it is a regression in human/computer interface design.

Oxygen is taking the human/computer interface to the forefront as the focus of the project and is willing to develop the necessary systems to support this. They are attempting to create a ubiquitous environment for interaction that is quite unlike anything that we are currently used to.

Imagine a quake match with three equally skilled players. One uses only a mouse for control, one uses only a keyboard, and one uses both simeotaneously. Who has the advantage? Obviously it is the person using both the keyboard and the mouse -- they have controls that form a more natural interaction with the game environment.

"Using a computer" in the future will no longer mean "operating a computer," it will mean "interacting with a computer." And our idea of the computer itself will change.

Honestly, how many of you enjoy sitting down at a terminal and typing, mousing, etc. your commands onto the machine? Really, think about this. The computer screen is absurd. The keyboard is absurd. The mouse is absurd. They are unnatural for interaction. They cause repetitive-strain disorders. So, we keep improving them. We make them ergonomic; we reduce the glare; in general we make them more "touchy-feely" versions of themselves.

You can talk all day about "the last mile" of bandwidth, Moore's law, amounts of information, and speeds of computers, but what remains is that after almost 40 years of computing, humans still have to (mostly) sit down and punch buttons and view a screen to work with a computer.

Computers are not interactive. They are unnatural. If you could communicate with a computer the same way you could communicate with another person, then we'd all be much better off and have much better (so called) "tools" to accomplish our objectives. Are we really slowing down human/computer interaction? Doubtful. If you try to visualize this work forced into the current human/computer interface it is slower and less efficient. When you change the entire interface, it really becomes a million times better!

~GoRK

[ Parent ]
It's a cool concept. (none / 0) (#42)
by Anonymous Hero on Thu Jun 22, 2000 at 04:47:20 PM EST

first of all i have to mention that their web page was one of the uglier ones i have seen in months. they obviously had not tried to look at it with a browzer that doesn't use a white background by default.

that said, i am very excited about the concept. they aren't the first ones to think of the ideas but none the less it does sound like a cool progect and here is why.

recently i moved from having a high speed conection to the internet to having a modem. the truth is that i really haven't missed the speed except for when i dist-update my system. and also it would be a problem to download movies. the thing i miss is being perpetually online. i guess i had become addicted to being able to telnet into my computer, to use the "dict " command to access online dictionaries, and to be able to stay on irc overnight logging the channel.

to me this is a parralel with what's going on at the oxygen progect. at first it seems fairly frivolous to have camera's around the room just to open the curtains when you come in. and to play music, make coffee, and adjust the lights for you. and to unlock the doors for you when you come back from work.

but then after a while you begin to like the way that when you come near your computer it logs you in automatically and the screen fades in just how you left it. no more nonsence about your kids messing things up exiting your programs without saving. because when they want to use the computer they can do it under a seperate user. even your 3 year old daughter who can't type a username or password.

of course you won't just have one computer in the house. but when you go to the other computers they'll have the stuff you left on the first computer on it. or they won't. but if you want it they will.

and it will all be connected to the internet. and connected to each other.

and all your shopping will be done online except food. although some will do that. and all your bank statements will be email. and if you get audetted by the irs the you have to type "irs --audit -lsd |lpr -y" . and that's it. your computer makes the appropriate phone calls while you go to the bar.

computers are now are basically blind and deaf. they should know where you are. they should here what you are saying.

they should clean the house when you are not home. and kill intruders too.

it's not about typing vs talking vs making faces and waving your fist at your computer. it's about how many physical things your computer could do if you gave it eyes and ears.



Re: It's a cool concept. (none / 0) (#47)
by tommasz on Fri Jun 23, 2000 at 08:50:56 AM EST

Think of it as a kind of user interface where there is no user interface. Part of the concept is to remove the conscious act of using a computer and replacing it with the actions you'd normally do, just augmented in some way. Xerox Parc has done this work in the past, it included active whiteboards that could be shared and viewed remotely as well as some of the more "gizmo"-like devices. The whiteboards were computers, but you didn't use them any different than regular whiteboards. What I like is that if the computer crashes, it's still a whiteboard :) When your PC goes down, it becomes an expensive paperweight.

[ Parent ]
The Star Trek link (none / 0) (#50)
by Nick Ives on Sat Jun 24, 2000 at 04:59:38 PM EST

Like everyone else, I'm gonna use ST as the example here.
Notice how in ST whenever they want to address the computer, they actually say 'Computer' and they get a little bleep of confirmation? Then they speak to the computer in a certain tone of voice to indicate they are giving the computer an order and not talking to the person stood next to them.
Also, in ST, they dont use speech for *everything*. The thin pads they use for written communication are pretty ubiquitus, and the terminals in each department (engineering, the bridge, etc) are the main form of interface with the ships systems. These terminals *combined* with voice input make for a pretty compelling, easy to use and apparently productive and useful interface.

So to finalise, voice input wont kill off the keyboard. I think we all know this. Just as email didnt kill written letters, voicemail will never really kill typed email. Instead what will happen is all these forms of input will combine and augment each other so maybe in the future I can say to my computer "load up kuro5hin and my email aswell" and a mozilla window will pop up and aswell as an xterm with pine, where I can proceed to type away my responses to my email.

--
Nick

Ubiquitous computing and MIT's Project Oxygen | 53 comments (51 topical, 2 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!