The Next Fifty Years - 4
I consider this one to be a very tough one to call. There have been anticipations of personal vertical takeoff and landing aircraft that will land in and take off from your backyard for the past fifty years. There are various problems with this scenario such as the dangers of wind shear and bad weather, traffic control challenges, the perceived dangers of aircraft falling in the neighborhood, the problems of certifying flightworthiness at all times, noise abatement, and fuel and other costs of personal aircraft. If this were to happen, it would probably be limited to lightweight passenger craft. Truck-type functions would probably remain the province of ground-based transportation systems. Computerized operations, robotic inspection of aircraft, extensive use of sensors, and so forth might enable such a development to occur at some point in the future, but I'm leery of guessing at a date this far in the future.
This is another prognosis that's hard to call, since it
depends partly upon technology and partly upon non-technical factors.
I think that changes in commercial air travel hinge upon breaking the flight access bottleneck.
speed-power product vs. thermal noise
There is another kind of limitation that will restrain computer miniaturization at some point: the average room-temperature thermal energy (= 1/40th of an electron-volt = 1/40 X 1.6 X 10-19 joules). We might select a lower bound of 1 X 10-19 joules on the energy required to switch from one state to another. At that level of switching energy, 1 transistor in e25 or about 1 transistor in 1011 would switch spontaneously as a result of thermal noise.
This choice would set an upper bound upon the speed-power product in terms of the number of transistors that could switch per second.
For example, if we set a chip's clock speed at 100 terahertz (1014 Hz) and we set an upper limit on a chip's power dissipation of 100 watts = 100 joules/second, then if 10,000,000 (107) transistors switched states per second, they would generate 100 watts of power.
If Intel's plan to achieve a transistor count of 10 billion (1010) transistors per chip running at a clock speed of 100 gigahertz (1011 Hertz) comes to fruition by 2010, and if all the transistors on the chip switch states 100 billion times a second, then we would be right at the 100-watt minimum speed-power set by thermal noise limitations. Of course, all the transistors on a chip may not average anything like 100 billion switches per second. But speed-power products may sooner or later set a fundamental physical limit upon circuit shrinkage and simultaneous speed increases. (I hope not.)
One way to beat this limit would be to operate at cryogenic temperatures, but that's awkward, and impractical for personal computers.
Organizations can be expected to embrace grid computing in which idle capacities on their networked computers are used for computationally intensive calculations when they're not engaged in foreground computing for their local operators. Local corporate networks can be run at very high speeds. Later, as bandwidths increase in the world at large, grid-computing arrangements may greatly expand the effective processing power of your PC. Most of us are using our PC's most of the time for word processing, or at most, surfing, so that our computers are idle most of the time.
"Grid" or cooperative computing may become a major way to boost computer speeds by networking innumerable private computers, many of which would be available upon demand for remote users. The key to this will be broadband data links that will permit the rapid transfer of information to and from remote computers.
How will computers be used?
Smarter devices and networking of devices has already been anticipated. Low-cost machine vision and recognition will probably change human/computer interactions. I would expect to see PDA's running in wide-band, wireless mode. The key issue here is that of the interface. Whether cell phones and PDA's will be embodied in the same instrument or whether cell phones and PDA's are separate is something I would hate to try to guess. Longer-term, there are dark horses in the race, such as more direct brain/computer interfaces. These might not necessarily have to be embedded in the brain. It might be feasible to use muscle signals to "read" the brain.
Telepresence and virtual reality displays
I'm still waiting, more or less patiently, for telepresence and virtual reality displays. Large screen, high resolution displays are becoming available and reasonably affordable. Desktop displays already have the size and resolution to provide telepresence and virtual reality imagery for individuals. For some reason, desktop displays aren't being used in this way except, perhaps, for video games. Cable modems and MPEG4 should permit reasonable quality teleconferencing and video telephony, but for some reason, that isn't showing up, either. I don't how much of this is sheer inertia, and how much of it is technical inadequacy. Since users must have the proper hardware and software at both ends of the telephone line, you can't do it by yourself. There has to be someone cooperating at the other end. One problem is that most people don't know how to go about it. Another problem may be the packet switching character of the Internet, which makes it difficult to use the Internet for long-distance telephone calls. Still, in fifty years, that should be under control.
Sharp's decision to sell 3-D displays may finally open up the 3-D market. I've been anticipating 3-D displays for ten years, and they still aren't here. When you see it in a properly designed theater, the effects are spectacular.
It seems to me that there's a lot that could be done that hasn't been tapped.
For example, I enjoyed scenes in the game Riven that gave you sights and sounds that made you feel as though you were in a real-world situation. I would pay for a screen-saver or partial-screen display that would display a restful, slightly dynamic scene, including sound effects. People have offered static scenes of beautiful locations, but how about a slightly dynamic scenes with sound effects.... for example, something with a few buzzing , humming insects, and bird calls, and a flying bird or two? A three-minute high-resolution, digital video recording of a beach scene, with several yachts, and maybe a shot of the beach with several good-looking bathers, or a recording of a beautiful waterfall would sell well. (You can buy picture frames that will display digital pictures, but they're very expensive. How about software that would accomplish the same thing using your desktop or laptop? You could harness old computers for this application at a much lower cost than a digital picture frame.)
The development of large, cheap, high-resolution displays will probably occur between now and 2052, and may lead to major changes in our worlds. So far, this hasn't really been tapped, and could be the open-sesame to a new world of telepresence and virtual reality.
Lightweight headset displays offer another possible route to augmented reality and virtual reality.
One-petaflops supercomputers are already in the works, both at IBM and at Cray Computers. Such supercomputers should permit more accurate weather forecasts and all sorts of simulations.
The development of local networks, and possibly even remote linkages capable of multi-terabaud data rates could revolutionize distributed computing
Today's 3 GHz desktop computers are as fast as most of the 1992 supercomputers.
With one-petaflops computers already under development, it's hard to imagine what kind of encore will be available in 2052. I would suppose that 1,000 petaflops computers might be in the offing, although future computing strategies may make such a simple measure of computing speed inapplicable.
3 4 5