Science Fiction Conventions: Talking Computers

[Scattercat]

Why?  Why talking computers?  Why is that considered to be so standard a progression that science fiction stories now regularly feature conflict, complications, and plot twists related to talking computers? 

Even ignoring the unbelievably difficult task of getting computers to the point where they'd understand human speech, what on Earth (or outer space) is the point?  Doing stuff by audio takes longer (witness the fact that a ten-thousand word story takes me fifteen minutes to read, but an hour to listen to on EP) and requires a lot more effort from working memory.  The only advantage it has is the hands-free nature of it, which I'll agree would be a nice feature to have... as an option rather than the apparent default mode.  Maybe for cars, or a recipe book in a kitchen, or at ATMs or other places where simulating a customer-salesperson interaction would be preferable in the consumer's mind. 

I'm frustrated by this apparently nonsensical convention (or more specifically, by thoughtless or unquestioning use of the convention), but I'm also genuinely curious.  Where did it come from?  Is it just Star Trek, where it was easier to have Magic Voice talk to the captain than to try and animate a "future" computer screen (with predictably laughable results, as many eighties movies will demonstrate)?  Shouldn't text be less driven by the needs of visual special effects and be free to have crazy stuff like, I dunno, that pin-model thing in X-Men that showed the model of their target?  Sure, that was goofy and ridiculous too, but at least it made more sense than having ordinary computers require verbal commands as the default interface.   Who started this "futuretech = talky bits" meme?


(I'm not specifically blaming podcasts, you understand, but...)   ;-)

[Heradel]

Thing is, audio can be pretty information dense. I can read an IM from my girlfriend that says "Hey sweetie", but if I hear it I can tell she's had a really quite horrible day from the way she's rushing through it before her voice catches. When we invented writing we didn't suddenly go 'What idiots we were with all of that modulated waves of sound business, lucky that's over'. Even now that a lot of humans spend a lot of time IMing friends instead of calling them, who hasn't had horrible misunderstanding because the tone didn't carry over ('you look horrible in those jeans'), or because you took too long to write a response. Sure, text is faster to read, but the speech carries information beyond the words.

[kibitzer]

Is it a given that speech commands must take longer? I'd have thought interface commands and styles would necessarily evolve, were speech readily understood by computers. Also, it doesn't have to be an either/or. I imagine some things would still be accomplished more "traditionally" whilst others would be well-suited to speech. Imagine trying to code a program through speech!  I'm sure it can be done but it'd be... painful. For me, at least.

It's also about eliminating barriers between humans and computers, whether that be to bring us closer to an efficient interface, or further anthropomorphism of the blank screen. Why'd Soong make Data, for example? Isn't he a computer in another form?  

[Boggled Coriander]

As far as Star Trek and similar TV goes, I assume it's just that a human talking to a computer comes across as marginally more interesting on screen than a human sitting in front of a terminal.  And they didn't have the budget or imagination to come up with a visual interface that would be interesting and make sense to the viewer.

[Scattercat]

To clarify further, this has been a general trend that often irks me, but the particular incident that raised my hackles was a science fiction story in which one character asks another if he can use the computer in their shared room after lights-out.  The reason permission was needed was because the whole talking back and forth thing would potentially wake up a sleeping roommate.  The roommate was flabbergasted when he was informed that, thanks to super-android-tech, the computer guy wouldn't need to TALK at all!  Amazing!

RE: Heradel

That's certainly true... for human-to-human communication.  Facial expressions and body language are even more essential for full comprehension.  However, for interfacing with my computer during daily tasks, text is vastly superior to any kind of audio, especially something where I'm not reading closely and need to skim.  Likewise, you can do a lot more work more quickly with a touch-based interface than by issuing voice commands.  Think about, say, resizing a window, or adjusting the brightness, even beyond any actual tasks at hand.

---

And yeah, I'm sure Star Trek did it because it was better for actors and cheaper on the effects.  Is Star Trek the origin of the Talking Computer meme?  Are any science fiction scholars on here better versed in the roots of the genre and able to pinpoint when that became a really major trope?  If it coincided with science fiction as popular television entertainment, then that would make some sense.

[MacArthurBug]

I actually have something of a half arsed answer - but only on a personal level. I'd love to have a computer that not only understands speech but can duplicate it in a natural sounding way. My Elder daughter is dyslexic. She has a VERY hard time reading, writing, or communicating via computer/text. I've been searching for a program that would allow her to speak aloud her thoughts and translate them into written format for a VERY long time. She grew up with some truly horrible text to speech programs, which I usually dumped and ended up reading things aloud for her. This is getting aggravating we both age. She wants to read/ listen to/ write as independently as possible. A computer or computer program that allowed her to do this would make my life, and hers a  heck of a  lot easier. I'm sure there are other people with similar or comparable disabilities that would benefit from such a system as well.

Just the flip side of the coin there please don't hit me with anything TOO heavy.

[Heradel]

I actually have something of a half arsed answer - but only on a personal level. I'd love to have a computer that not only understands speech but can duplicate it in a natural sounding way. My Elder daughter is dyslexic. She has a VERY hard time reading, writing, or communicating via computer/text. I've been searching for a program that would allow her to speak aloud her thoughts and translate them into written format for a VERY long time. She grew up with some truly horrible text to speech programs, which I usually dumped and ended up reading things aloud for her. This is getting aggravating we both age. She wants to read/ listen to/ write as independently as possible. A computer or computer program that allowed her to do this would make my life, and hers a  heck of a  lot easier. I'm sure there are other people with similar or comparable disabilities that would benefit from such a system as well.

Just the flip side of the coin there please don't hit me with anything TOO heavy.

There are the Kurzweil Reading Machines, but I'm given to understand that they're not cheap. Certainly do that text to speech thing with physical books though. For the input side, one of my professors likes Dragon Naturally Speaking.

RE: Heradel

That's certainly true... for human-to-human communication.  Facial expressions and body language are even more essential for full comprehension.  However, for interfacing with my computer during daily tasks, text is vastly superior to any kind of audio, especially something where I'm not reading closely and need to skim.  Likewise, you can do a lot more work more quickly with a touch-based interface than by issuing voice commands.  Think about, say, resizing a window, or adjusting the brightness, even beyond any actual tasks at hand.

Ok, but most of the time the talking computers verge on sentience rather than just being terminals, though those exist too.

I'm just not sure text always would be faster. Star Trek is the default answer here (though I'm not sure it's the first), and they certainly use LCARS a lot. I think there's a convenience factor in here too we're glossing over. It's easier to say "Computer, Tea, Earl Grey, Hot" when walking over to the replicator than punching through a bunch of choices on a touchscreen.

[jrderego]

She has a VERY hard time reading, writing, or communicating via computer/text. I've been searching for a program that would allow her to speak aloud her thoughts and translate them into written format for a VERY long time.

Dragon Systems: Naturally Speaking. That's the program you want.

And yeah, I'm sure Star Trek did it because it was better for actors and cheaper on the effects.  Is Star Trek the origin of the Talking Computer meme?  Are any science fiction scholars on here better versed in the roots of the genre and able to pinpoint when that became a really major trope?  If it coincided with science fiction as popular television entertainment, then that would make some sense.

Because in 1965, when they were shooting Star Trek, the computer interface was a paper tape, or tractor paper print out spooled through a specially equipped electric typewriter attached to a huge punch card or paper tape driven monstrosity that was very good at calculating numbers and nothing else. Plus, there was no precedent for visually explaining what a computer was and how it worked or how someone might use it. Also, most of the time the computers in Star Trek were used sort of like search engines more than anything else, again, because no one really knew what computers would be good for.

Inputting data via a stack of 5000 punch cards wouldn't have made for interesting TV. And there was no real paradigm for human/computer interaction at the time anyway. Interfaces that regular people would sort of understand didn't really happen come into the real world until the Apple 1 was being sold for $666 dollars in the mid 1970s.

The visual style and hardware paradigms set up in Star Trek were carried over to The Next Generation and other shows too moreso because they way the interactions happened in The Old Series were comfortable and contingent with the look and feel that The Next Generation wanted to maintain.

Even ignoring the unbelievably difficult task of getting computers to the point where they'd understand human speech, what on Earth (or outer space) is the point?  Doing stuff by audio takes longer (witness the fact that a ten-thousand word story takes me fifteen minutes to read, but an hour to listen to on EP) and requires a lot more effort from working memory.  The only advantage it has is the hands-free nature of it, which I'll agree would be a nice feature to have... as an option rather than the apparent default mode.  Maybe for cars, or a recipe book in a kitchen, or at ATMs or other places where simulating a customer-salesperson interaction would be preferable in the consumer's mind. 

I've had a voice interface on all of the computers I've had since 1999 or so as it's been a regular part of the Mac OS since OS 9, and while it's not great, it does allow me to check mail or navigate between open windows, even do a little editing in Word or Text Edit,  while typing and not having to take my brain off of the task that in which my fingers are engaged. Does it always work? No. Is it good for much much than fetching email, or reading back a text message, or very light editing? Not really. But in some cases the technology that works in computing gets applied to other places, like hands-free cell phone software that allows dialing by spoken name.

[MacArthurBug]

Thanks for the reccomend to Dragon systems. We'll give it a go and see if it's as good as advertised. I'm pretty sure we can write it off tax wise as educational- but even if we can't it'd more then be worth the cost. You've officially made my day

[Sandikal]

Even ignoring the unbelievably difficult task of getting computers to the point where they'd understand human speech, what on Earth (or outer space) is the point?  ... The only advantage it has is the hands-free nature of it, which I'll agree would be a nice feature to have... as an option rather than the apparent default mode.  Maybe for cars, or a recipe book in a kitchen, or at ATMs or other places where simulating a customer-salesperson interaction would be preferable in the consumer's mind. 

Aren't computers almost to the point where they can understand human speech now?  My cell phone has no problem understanding me when I say "Call Home".  I didn't even have to "train" it like I did the cell phone before it.  I also spend a lot of time talking to computers on the phone.  I place a call to my credit card company, my insurance company, or a utility and a computer asks me a bunch of questions.  It sometimes understands my answers better than real human beings.  I think we're 5-10 years away from voice-recognition computer being as commonplace as they are on Star Trek.  I think there will be a demand for it because of the hands-free aspect of it. 

For the person who asked if Star Trek was first to use the talking computer trope, I think you'll find there's an episode of "The Twilight Zone" with a talking computer that falls in love with its programmer.  I'm pretty sure "The Twilight Zone" episode pre-dates "Star Trek".

[wakela]

Software is designed to be used with the keyboard, mouse, and monitor because understanding and producing speech is so problematic.  But once we get the bugs ironed out (I agree, about 5-10 years) software will become more voice-friendly.  As was mentioned above there will be some things better suited to typing, some things better for voice.  I'd rather program my TiVo with voice, or check my mail while walking or driving.  Also, there will be things we haven't thought of.  Kind of like how mp3 players were intended to store and play music, but users came up with the idea of podcasting. 

But, Scattercat, I love questioning genre conventions, and this is a good one.  The example you give in the dorm, does seem pretty stupid.  Even if the computers all talked, why not use headphones?

How about a future where computers can understand sub-vocalized commands and you can type by slightly twitching your fingers...

[Bdoomed]

type by slightly twitching your fingers...

yeah now I want you to try only slightly twitching your fingers in random patters and quickly.  not easy.  heh plus i guess i'd consider modern typing "slightly twitching your fingers" anyway.  but as for even less movement than that, not so easy to do.

plus anyway there won't be voice commands it'll just be thoughts because all computers will be in our brains DUUUH.

[wakela]

type by slightly twitching your fingers...

yeah now I want you to try only slightly twitching your fingers in random patters and quickly.  not easy.  heh plus i guess i'd consider modern typing "slightly twitching your fingers" anyway.  but as for even less movement than that, not so easy to do.

plus anyway there won't be voice commands it'll just be thoughts because all computers will be in our brains DUUUH.

I don't know, I don't think it's that hard.  You could have a virtual keyboard on the bottom of the monitor and there would be dots to indicate where your fingertips are in virtual space.  So you can move your fingers over the keys and do a tap motion when you want to hit one.  The advantage would be that you could move your virtual fingers a long way with little real effort, like when you move your mouse a little bit and the cursor goes across the screen.  I bet you could even do this with today's technology, but probably not for cheaper than a physical keyboard, and people like the tactile sensation.  But it might be handy with eyeglass monitors, so you wouldn't need a desk.  I'm not saying it's going to happen, or even that's it's a good idea, but it's at least as realistic as the talking computers.

[Alasdair5000]

Just as an aside, Red Planet, which is WAY more fun than I expected it to be (Val Kilmer!  Terence Stamp!  Carrie Ann Moss!  A killer robot!  Basically EVERY standard hard SF plot in two hours!) has the flat out best talking computer moment I've ever seen.  It's about halfway through, after the ship's taken some pretty serious damage and Captain Bowman (Yes indeedy) asks for a damage assessment.  The computer provides it, Bowman mutters:

'That's not good.'

and the computer replies

'No, it's not.'

See, THAT's the kind of talking computer I want, an AI that will at least sympathise with you:)

[Bdoomed]

See, THAT's the kind of talking computer I want, an AI that will at least sympathise with you:)

Watch Outlaw Star, that has a nice AI.
and of course the AI on board the Heart of Gold
and Cortana
and anyone from Reboot

[wakela]

I just thought of something.  How would modern computing be different if voice activation and speech synthesis had been developed instead of typing?  What would that do to programming languages? 

[Heradel]

I just thought of something.  How would modern computing be different if voice activation and speech synthesis had been developed instead of typing?  What would that do to programming languages? 

See: Natural Language processing.

[kibitzer]

I don't know, I don't think it's that hard.  You could have a virtual keyboard on the bottom of the monitor and there would be dots to indicate where your fingertips are in virtual space.  So you can move your fingers over the keys and do a tap motion when you want to hit one.  The advantage would be that you could move your virtual fingers a long way with little real effort, like when you move your mouse a little bit and the cursor goes across the screen.  I bet you could even do this with today's technology, but probably not for cheaper than a physical keyboard, and people like the tactile sensation.  But it might be handy with eyeglass monitors, so you wouldn't need a desk.  I'm not saying it's going to happen, or even that's it's a good idea, but it's at least as realistic as the talking computers.

That kind of stuff sorta already exists: http://www.virtual-laser-keyboard.com/ or http://www.projection-keyboard.com/store/productInfo.html?productId=22

(I said "sorta" -- I realise this isn't virtual-in-space)

[wakela]

I just thought of something.  How would modern computing be different if voice activation and speech synthesis had been developed instead of typing?  What would that do to programming languages? 

See: Natural Language processing.

Right.  So what would modern computers be like if computer evolution had started with Natural Language Processing instead of typing (improbable as that may be)?  Would we be sitting around in offices mumbling to our computers?  Programing as we know it would be impossible if you had to speak all that nonsense.  Then again, it's not very readable to non-programmers.  What would a spoken programming language be like?

Kind of like wondering what keyboards would be like if the Chinese or Japanese had invented them. 

Quote from: wakela on November 16, 2009, 02:32:01 am
I don't know, I don't think it's that hard.  You could have a virtual keyboard on the bottom of the monitor and there would be dots to indicate where your fingertips are in virtual space.  So you can move your fingers over the keys and do a tap motion when you want to hit one.  The advantage would be that you could move your virtual fingers a long way with little real effort, like when you move your mouse a little bit and the cursor goes across the screen.  I bet you could even do this with today's technology, but probably not for cheaper than a physical keyboard, and people like the tactile sensation.  But it might be handy with eyeglass monitors, so you wouldn't need a desk.  I'm not saying it's going to happen, or even that's it's a good idea, but it's at least as realistic as the talking computers.

That kind of stuff sorta already exists: http://www.virtual-laser-keyboard.com/ or http://www.projection-keyboard.com/store/productInfo.html?productId=22

(I said "sorta" -- I realise this isn't virtual-in-space)

ooooooh yeah.  I remember reading about these things.  I've never met anyone who used one.  I wonder how accurate they are.  But if you used VR gloves instead of a laser keyboard, you could type with much smaller and faster finger movements, maybe even evolving to the point of faint muscle flexes that are not even visible. 

[Yargling]

Even ignoring the unbelievably difficult task of getting computers to the point where they'd understand human speech, what on Earth (or outer space) is the point?  Doing stuff by audio takes longer (witness the fact that a ten-thousand word story takes me fifteen minutes to read, but an hour to listen to on EP) and requires a lot more effort from working memory.

Thats not a reason NOT to have it. After all, AI and FTL travel are 'unbelievably difficult' with current technology and engineering, but we believe they many be possible at some future date.

As for why they are popular in scifi, its the same reason Windows holds 90-ish% of the desktop computer market; because the Linux alternatives are NOT user friendly to get started with. Its the same reason that (sadly) PC gaming isn't really growing whereas console gaming is growing year on year; consoles are easier to setup and use.

With voice controlled computers, it is (in theory) extremely easy to use (unless you don't have the power of speech). Also, if the story also features AI, talking to and with a computer seems the natural way to communicate with them - treating them as people inside a computer.

[Planish]

I've had a voice interface on all of the computers I've had since 1999 or so as it's been a regular part of the Mac OS since OS 9, and while it's not great, it does allow me to check mail or navigate between open windows, even do a little editing in Word or Text Edit,  while typing and not having to take my brain off of the task that in which my fingers are engaged. Does it always work? No. Is it good for much much than fetching email, or reading back a text message, or very light editing? Not really. But in some cases the technology that works in computing gets applied to other places, like hands-free cell phone software that allows dialing by spoken name.

PlainTalk speech recognition came out with the AV Macs and System 7.
Even before that, there was "MacInTalk 2" text-to-speech on the Mac Plus and SE with System 6 in 1988. Also the unofficial MacInTalk for the original first-gen Macs and Lisa.

But in some cases the technology that works in computing gets applied to other places, like hands-free cell phone software that allows dialing by spoken name.

We'll probably see a lot more of stuff like that in automotive applications as jurisdictions pass laws banning the operation by drivers of hands-on and "must look at" devices.

[edit]
Hey, can we invite Vicki to this discussion?

[Planish]

Thing is, audio can be pretty information dense. I can read an IM from my girlfriend that says "Hey sweetie", but if I hear it I can tell she's had a really quite horrible day from the way she's rushing through it before her voice catches.
[snip]
Sure, text is faster to read, but the speech carries information beyond the words.

I heard an interview on the radio with a guy who was a telegraph operator back in the day. He said that they could identify other operators just by how they keyed, which was as unique as their voices. Bear in mind that the only thing that could vary is the time delay between identical-sounding clicks. They could even sense when people were sick, had trouble at home, or were particularly happy, etc.