Slashdot Mirror
The Computer of 2010
nostriluu writes "
With the assistance of award-winning firm frogdesign (the geniuses behind the look of the early Apple and many of today's supercomputers and workstations), Forbes ASAP has designed and built (virtually, of course) the computer of 2010."
I don't know why everyone thinks we want to talk to computers. I want to talk to my computer about as bleeping much as I want to talk to my television. I can't talk 100WPM but I can get close to that on a keyboard - and I don't know why you'd want to change that. Even thought recognition would be a pain in the ass. I can type almost without thinking about it - which might explain some of my posts, ha ha, but surely you must know what I mean; Thoughts flow easily to keyboards that might not to voice. Maybe that's conditioning, but writing down thoughts is something that goes back for all of recorded history and I think it's more than just me.
Computers of the future will be optical. They'll run at 100's of Ghz. They'll have stupid huge hard drives. Hell, they might even think. But you won't be talking to them - because it's plain not efficient compared to other input techniques, like computers. Do you know how sore your vocal chords would be after dictating all day?
Arrggh. That's my rant for the day.
..don't panic
1. You show me a way to enter C code with a voice system and _then_ I'll throw out my keyboard. I could just see it: "up, up, up, left brace..." Screw that.
2. Processors don't have to spend 2/3 of their time waiting around for data. Real ones at least. I have a 533MHz alpha that does 980 MFLOPs, don't tell me it's waiting around most of the time.
3. I doubt that anyone will want to use Lithium batteries in ten years because fuel cells will have been out for 8 years.
4. If we have a quarter terabyte of main magnetic memory, what is the terabyte of optical disk for? It's the only moving part in the computer, what the hell do we need it for? Magnetic memory is static.
5. What about the network connection? OC-192? Better? I'd personally vote for some type of ATM, especially if we're going to use it for all of our communications. QOS is important, I don't want to lose frames on my movie just because someone calls..
6. They think that absolute security relies on thumbprints? Give me a break (or break-in). What we really need is to make sure that IPv8 is double-key encrypted at all levels.
7. There's nothing that they describe that is going to take a Cray to process. What does the typical secretary need with a supercomputer? A voice activated webpad is about enough. Gamers are another story entirely. Immersive VR is going to take more than they've got scheduled anyway.
In short, the forbes article is a fluff piece.
At least you didn't say they worked well. Hey, let's look at some input device "theory" shall we?
1. You store information in your brain. It's chemical. It's analog.
2. You want that information in your computer. It's electric. It's digital.
3. Can it possibly be that the best way to bridge these two qualitative gaps is by wiggling physical limbs over hard plastic nubbins?
4. Depressingly, the answer appears to be "yes"...
5. So now it's down to a matter of appendages, nubbins and how you wiggle them (feel free to make porn jokes now)
6. Alternate WAN (wiggling appendages over nubbins) techs have risen and fallen. The mouse is a popular WAN... but the guy who came up with the mouse idea (you know, whats-his-name who worked at SRI) also had this bizarre "chord playing" device for input as well... sorta like using an one-handed accordian.
7. Text. We want text input because we're slaves to alphabetic, pseudo-phoenetic written languages.
8. WAN techs must not only be efficient but be acceptable by people as well...
9. So, we need a WAN. It must be text-oriented, efficient and have a high acceptance rate among people.
10. You're answer to that is the keyboard. I work with a guy who turns blue under the eyes without his stylus.... the bottom line is:
We have WANs now that do the job, but we have seen new WANS (mouse, stylus) come along and there is no reason to think that WAN evolution will stop just because we like our F-keys and Num Lock. In 1983 I would never have imagined a mouse. But it happened.
2 1337 4 u!
I mean, really, why do people want to do away with keyboards?
Keyboards are quick and efficient. This article says that you'll instead use a 3D interface, and simply touch with your hands what you want to do.
Is it me, or does that sound rather slow and clunky? Do I really want to be waving my arms around just to open a damn program?
Face it: keyboards are still around after all these years because THEY WORK. They might not look futuristic or uber-high tech, but THEY WORK.
-[Blaine]- "'Oh dear,' says God, 'I hadn't thought of that,' and promptly vanishes in a puff of logic."
Now, let's talk about the computer of the future I imagine. First of all it will be a half dissasembled box with various optical cables coming out of it and a little bit of dust gathering on the exposed parts. The processor is of course tweaked in some way as to make it 1.5-2 times as fast if occasionally unstable.
The computer is hooked up via a wireless VPN to a bunch of my hacker friends all over the world where we share our thoughts, and our music in secrecy. Of course I've got a high bandwidth Internet connection. It's perfect for serving up movies, music, and games, but it's still not quite enough to handle some of the latest technologies (some things never change).
I've got several of my older computers hooked up on the other end. Sure, they are slow and primitive, but it's fun! Needless to say these are all in a state of semi-disarray, with cables in a giant spaghetti mess on the floor.
Sure, I've got one of those cool mega-displays that display everything in photographic quality in a screen the size of a desk, but I've got some throw backs. I've of course got a keyboard since those virtual keyboards are cludgy at best. I've got a scrolling LED display I found in a junk yard and managed to hook up to my box. If somebody tries to hack my box a bunch it displays a message on the LED to let me know what's happening.
Now, that sounds like my dream computer of the future! Maybe it would be nice to have something portable to go with it, but I want a box I can hack and play with.
---
This sig has been temporarily disconnected or is no longer in service
There would have been no way that they would have predicted the importance of the Internet - or something like Slashdot. In 1990 the communications capability of computers was only known and appreciated by a very few geeks; most people had local call modem access to bulletin boardsif they had anything. (Please don't post how you had access to the Internet in 1983 - that just proves YOU are a geek and nothing else. Who could an average person have used as an ISP in 1990?)
In 1990 very few prognosticators would have predicted anything like a noticeable percentage of people running a Unix style operating system. Nor would they have predicted anything like Windows 2000 or an iMac.
One of the most interesting things about this article is that they had almost nothing to say about the - external to your house - communication capability of the machine. I suspect this will be one of the most important aspects of that machine.
One of the reasons that I bought OS/2 Warp 4 was the voice recognition capability built into the OS. I wound up using it very little. Not because it didn't work, it did. The reason I didn't use it much was that in order to activate it I had to say the word 'desktop'. For me at least 'desktop' is a VERY difficult word to pronounce properly. The 'k' sound at the end of one syllable followed by the 't' sound at the start of the next is just tough to say. When I thought about it I realized that I pronounced it 'destop' as do many of the people who say it in normal speech. The computer didn't know what a 'destop' was.
'Desktop' is a minor stumbling block, but it is the sort of thing that keeps voice recognition from being utilized as much as it could be. One of the keys to a useful voice command computer is to use words in the command structure that people can pronounce.
There is also a slight misconception in the article; the good thing about optical communication between computer subsections is not the speed of light vs the speed of electrical pulses - the good thing is that optical communications can switch on and off faster; you can obtain higher frequencies.
The article also gets it a little wrong when it blames the electrical interconnect for causing delays in main memory fetching. The problem is that dram speeds have only grown about 10 times faster since the days of the Z80 while processor clock speeds are up by a factor of 250 or so. Unless there is a real breakthrough in memory speeds that trend will continue.
Let's take a little look at this proposed computer of 2010:
SECURITY
The PC will be protected from theft, thanks to an advanced biometric scanner that can recognize your fingerprint.
Now all they need is biometric scanners on screwdrivers too.
INTERFACE
You'll communicate with the PC primarily with your voice...
This should make university computer labs interesting, especially for people writing code. And how about when your friend Bob pops into your office to say hello:
The Desktop as Desk Top
In 2010, a "desktop" will be a desk top ... You won't need a keyboard because files can be opened and closed simply by touching and dragging with your finger.
Be careful when drumming your fingers.
Your Home
The PC of 2010 plugs into your home so your house becomes a smart operating system.
"Open the refrigerator door, HAL." "I'm sorry, Dave, I'm afraid I can't do that."
The master control is made up of no fewer than 50 redundant processors, each smaller than a millimeter cube and much faster than a modern computer, though they vary somewhat in their contruction, scattered about my reinforced skull. There is no way for all of them to be damaged while I remain alive.
The workhorse of the system responsible for archiving and heavy processing, a solid-state homogenous massively parallel connection machine about the size of my fist, is tucked away in my chest cavity between my oxygen-storage organ and my 2nd and 3rd backup hearts (cylinders about the size of a thumb).
The interface can take any form, since the system has full access to all nerves leading into my brain, and has plenty of power to simulate a believable environment; it can be superimposed atop real-world data or it can be fully immersive. The failsafes are carefully trained (in a process taking months of daily feedback training) direct neural connections to the master control, which can be used to cut off any problematic computer functions and reconnect my mind to my body; a spare nervous system is tatooed across my skin so there is no single point of failure. I'd have to be almost fully decapitated to lose control of my body (not likely to occur accidentally, thanks to carbon microfiber reinforcements). I'd certainly be dead before my connections were disabled, unless a very careful surgery was undertaken.
Connections to other computers can be made by many different forms of electromagnetic transmission, or by tiny electrical currents through contact on any of hundreds of points on my body.
Power is available through the main storage battery of my body (a distributed system with surprising capacity), but essential functions, such as the master control, can be supported by generation of minute amounts of power from the glucose and oxygen in my blood.
---
Despite rumors to the contrary, I am not a turnip.
For some reason people are obsessed about guessing our future, and determining how our lives will be affected and what technologies will be in use. The simple truth of the matter though, is we are really, really bad at doing it.
In the 1950's and 1960's a lot of predictions (many of which you have probably seen yourselves) were made as to how we would live in the year 2000. Now, I don't know about you guys, but I'm not wearing silver jump boots and driving my hover car to work yet, and I still have to cook all by myself. It was foolish to try and predict 40 or 50 years into the future, and we've learnt our lesson. So now, a design team want to get some publicity and attempt to predict where our computers are going to be in 10 years time. There are some flaws though.
1. Why do I want a computer shaped like a frisbee in the first place? Too big to carry around, too small to make it look like it deserves a space in my office.
2. All of these technologies are still on drawing boards and we won't see prototypes until 2010 at the earliest. In addition better and more useful technologies are likely to emerge between now and 2010 meaning some of these components may never get the R&D required to develop them.
3. Desks as screens is the most stupid thing I've ever heard in my life. After a weeks use, the cramp and pain in my neck would probably become unbearable. It's also an "illegal" position for visual display equipment to be placed in under the Health and Safety at Work Act in the UK. But hey, that nasty dude with the supercomputer in Tron had one, right?
4. More powerful computers doesn't mean smarter computers. The article implies that because this machines will have equivalent power to something like a Cray J90 or somesuch that it will therefore make our environments "smart". So, does that mean they are predicting advancements in artificial intelligence as well? Funny that, because for the last 2 decades people have been saying that in 10 years time we'll have smart machines.
In short, I think that this is possibly the worst article I have ever seen concerning the future of computers over the next ten years. Seeing as it's completely publicity-generating pie-in-the-sky and not clearly thought out by anybody who understands these issues, why am I even bothering writing this reply? Because I've got nothing better to do? In that case, I think I'll go and design my house of the future (*yawn*).....
I think that the future of input devices will not be an all or none soluation, but rather a combination of input devices and methods.
As it is, there are many things I want to tell my computer to do, for example.
Save this file then shutdown for the night.
Saying that one sentence is much faster then waiting for the file to save, then checking to make sure that the OS actualy shuts down OK (granted, more reliable OS's would solve half of this problem.)
On the other hand, there is no way in the world that you are going to seperate me from my Rhino3d command bar, its a lovly thing. Abbreviations are much easier to type then say, and often times written words form patterns that spoken words do not. Thus the reason that I can type my name two or three times faster then anything else, my fingers have become acostum to typing it. On the other hand I still stutter while saying my name (ironic, eh?)
I do like the idea of a virutal desktop though, there are many times in life that I just wanted to be able to play around with somthing in real 3d. In addition, it is easier to tell a computer to do somthing like:
Undo that second to last spell check
Then it is search through a document for the word that you accidently corrected (spell checkers don't understand most industry buzz words, go figure!)
Of course there is one major problem with voice reconization technology right now, namly
THE DAMN THING CAN'T UNDERSTAND A WORD THAT I SAY.
Sure sure, I've read the manuals "may need 3 or more hours of training in order to work as advertised"
I've also spent half an hour trying to get the trainer to understand a single word that I am saying.
The reason?
Simple.
I have spent so much time reading, that most of the words that I know the definition of, I have never actualy pronounced!
Honestly, how often does the word defiled or the phrase uber mage come up in daily conversation?
The fact is that the written english languages deviates strongly from the spoken english language, heck, ask almost any english lit teacher if you don't believe me.
Many words that people use in their daily spoken language that they have perfected the pronounciation of, they have never actualy written down, and many words that they write down, they may have never pronounced in their entire life (I know I have gone 5 or more years without actualy speaking a word but using it almost weekly in my writting.)
The basic fact is that it is not natural to dictate a document. Most authers who dictate documents (do any writters still actualy do that anymore?) have somebody else go through and eliminate the standard mannerisms of speech.
(this also explains why those medieval fantasy novels sound so emmensly lame when read out loud).
Yet there is hope for integratiing voice reconization into the modern and everyday computer. Many long winded or multiple step commands would fit very well into a voice reconiziation programmed enviorment. Granted, all the Linux nuts out there have script files to do most of the work for them if the job exeeds 5 or 6 long winded tasks, but for other commands, such as the given example of save then shutdown, this would be the perfect system.
Imagine somebody being able to give the following command:
Delete all games that I have not played in the last 5 months and then defragment my hard drive.
The best part about the above command is that it is not technoloicaly impossable. Heck, a file system which keeps track of "last date accessed on" and that allows for catagorizing of directories (games, apps, pr0n) is all that is needed.
Whats more, it would save me the 5 or so hours it takes for me to sort through my games directories.
It has even another bonus, the average user could understand it! Granted, I'm not a big fan of the "average user" (*cough* newbie *cough*) but it doesn't exactly take a rocket scientist to understand the idea of deleting everything that hasn't been used in at least 5 months. Just follow up by (attempting to) explaining defragmenting a HD (you have to do it, shut-up, end of story,-- my perfered method) and you have a very easy to use and often needed command that is actualy easier to use through voice reconization then it is to do by keyboard or mouse
.
Examples of other commands that are easier to use through voice recon then traditional commands:
Save this file as (document name)final.txt then delete all other previus backup copies of this file.
Goto visited websites and rotate all passwords
remove all non-program essential image files
DELETE THAT DAMN DIRECTORY ALL READY (for those people are are stuck with windoze and uninstall shield, you know what I am talking about!)
Find some of all items in colum C and place it in the next avaible row of colum C
Run virus program, clean up temp files, then run defrag
Partional off the newly installed 20gig HD into two 10gig segments labled D and E
Copy all user created text files to the CD-RW that was just inserted
Delete all MP3's that have only been listened to once
Collect all non-program attached midi files and move them to usr/music/midi
(yes yes, some people here still listen to midi files, get over it!)
Anyways, just my, err, uh
2 and a half bits I guess, heh.
Need help treating your acne? Come here!
I recently purchased an used book called Computer Systems in Business: An Introduction recently. Published in 1986, it features computer systems that seem pretty quaint almost two decades later.
One clever thing the book does is discuss the concept of P, a relative-cost measure for dicussing how much computers cost in light of the opposing forces of inflation and Moore's Law which make costs difficult to forecast in actual currency. P is equal to the average cost or value of a standard microcomputer system, which at the time consisted of...
- CPU
- Main memory (256K)
- Keyboard
- Monitor (they specified monochrome)
- Two diskette drives
- Dot matrix printer
The book does mention some things that seem pretty quaint these days: main memory for some larger micros exceeded 2 megabytes in 1985, two diskette drives would give you a total on-line storage capacity of a megabyte, and my favourite line: "A Winchester is sometimes called a hard disk."The books goes on to discuss minicomputers (6 megs memory with 200 ns access time, 8K cache and 1.8 gig of hard disk -- er, Winchester -- space. Only when the book gets to mainframes do the machine specs seem somewhere in line with a machine you could probably buy at your local appliance shop...
The diagram, which I can't include in this posting, shows 35 disk drives of 800 megabytes each (28 gigs total). In P units, the cost of such a beast was 1,043, or about $3.6 million dollars.If we assume that the rate of tech progress over the next 15 years is roughly the same as the past 15, we have a starting point for visualizing what the average personal computer of 2015 could be like -- simply look at machines that cost about a thousand times more than the Celeron 600 at Circuit City.
Then, in the best /. traditon, imagine a Beowulf cluster of them!