Giving the Globe a Networked Skin
n the 20th century, scientists at Bell Laboratories invented modern marvels like the transistor and the laser, but that is ancient history to the engineers working there today for Lucent Technologies Inc. In fact, they do not really have much use for the present.
In a windowless office at Bell Labs' headquarters in Murray Hill, N.J., Gerald Butters, chief of Lucent's optical networking group, gestures dismissively at his magazine-thin laptop computer and the network cord dangling beside it.
"Look," he says, "to get onto the network today, I've got a skinny wire here and I'm limited by the speed of the modem and I'm limited by the network to something surely less than the processor speed in this computing unit."
In the Bell Labs vision, that wire will soon disappear. For that matter, the laptop computer may disappear as well. In their place will emerge a world in which the ability to connect -- with wires or without -- to a global network at lightning speeds will become almost ubiquitous. For when you think about it, which is what Bell Labs spends billions doing, the communications revolution so far has been limited to a mere handful of devices -- the phone, the fax, the modem, the television.
Moreover, consumers are generally stuck with only the services that big carriers want to offer. If caller ID is not sold in your area by the local phone company, well, too bad. Want to set up an eight-way conference call for the whole family? It's not generally possible at home. Want your e-mail automatically forwarded to a fax at your hotel? Good luck.
The upshot is that for the last few decades, more technology has meant more complexity -- learning a new computer, understanding voice mail instructions, programming a VCR.
The challenge that Bell Labs has set for the next few decades is to make ever more capable communications systems as easy to use as they are powerful, to make the technology sufficiently advanced that it becomes transparent.
One big part of that is making the global communications network ubiquitous, rather than something that is accessed from a computer at a desk, or a phone jack in the wall. Bell Labs' world is one covered by a skin of communications.
Take the changing role of fiber-optic cable. For almost two decades, optical fibers have been widely used to transmit torrents of data between fixed points employing pulses of light rather than pulses of electricity.
Now, Bell Labs engineers are trying to perfect a concept largely articulated by British Telecommunications P.L.C.: small wireless communications devices -- essentially antennas -- attached to fiber-optic lines every few hundred feet. Anyone within range of these transmitters could wirelessly tap into the fibers' vast capacity with their camera, or their video unit, or their living room remote control as easily as people use cell phones today.
With thousands of miles of optical fiber being deployed around the globe every day, if not every hour, according to many analysts, the industrial world could soon be encased in a literal web of communications.
But such a web is of little use if people still must link to the network with clunky, largely fixed devices like today's computers and telephones.
That is why Bell Labs is developing new devices and new ways of interacting with them.
In Holmdel, N.J., for instance, researchers at Bell Labs have developed what they call the world's smallest camera. It is on a microchip the size of a fingernail. Coupled with a lens and some supporting circuitry, an entire camera unit could be sewn into a lapel, perhaps as a security device. Were you mugged? The offender's picture could be on its way to the police within moments.
Similarly small circuitry, coupled with the power of ubiquitous high-speed wireless connections, could allow doctors to continutally monitor the health of chronically ill people, using wireless biomonitors in a wristwatch.
All of this is meant to facilitiate communication with people. Currently, communicating with someone who is not in the same room requires linking a set of electronic codes -- home phone numbers, work phone numbers, cellular numbers, fax numbers, pager numbers and e-mail addresses -- with the time to find the right code for the right moment.
In a few years, you may be able to tell your watch, "Call Bob," and not have to worry about where Bob is or what sorts of devices he has handy. Using technologies like those being developed at Bell Labs, the network will handle those chores -- finding Bob, if Bob wants to be found.
Sound like science fiction? That's what many would have called the Internet just a decade ago.
Can Molecules Beat Moore's Law?
It's an elaborate parlor trick. But this is an elaborate parlor: I.B.M.'s Almaden Research Center, where Mr. Chuang, a physicist, is one of hundreds of people searching out the next frontiers of computation. It's a submicroscopic realm where the rules can be confounding and the results uncertain.
Mr. Chuang's quantum computer, for example, is based on strange physics that apply when it is possible to control the behavior of individual molecules and atoms. In contrast to conventional computers, which store and retrieve information represented as ones and zeros, quantum machines can simultaneously compute upon a potentially vast number of pieces of information in the space of the same one or zero. It is a through-the-looking-glass world in which yes and no can be true at the same time, and where events can happen simultaneously in multiple places.
The potential is enormous. The reality is that Mr. Chuang's device has not shown that it can solve useful scientific problems. And the need for such a breakthrough is being felt with urgency by the computer industry.
In a range of fields that place demands on the components of a modern computer -- processing logic, memory and permanent storage -- researchers are facing the challenge of how to continue creating faster and more powerful computers once the limits of Moore's Law have been reached.
In a phenomenon first observed in the 1960's by Gordon Moore, the co-founder of the Intel Corporation, the semiconductor industry has been able to double the number of transistors on a single piece of silicon every 18 months, leading to a still accelerating increase in computer processing power. The industry thinks this will continue until 2014 when, if not sooner, the basic processes underlying semiconductor electronics will simply stop working. Devices will be so small that they will be plagued by the very subatomic forces that Mr. Chuang and his colleagues here are hoping to harness.
In his laboratory, Mr. Chuang holds a thin test tube filled with a bright yellow fluid. Suspended in this liquid crystal solution are about 10 septillion (10 to the 18th power) molecules that make up the fundamental component of a quantum computer, known as a qubit. These are molecules chosen because in an intense magnetic field their electrons can be precisely oriented. Computing takes place when the molecules are "programmed" with a burst of radio waves, causing the rotation of their electrons to change and a quantum calculation to occur.
With a small number of qubits, a quantum computer could perform a vast number of calculations in parallel. One possible use might be to factor large numbers rapidly -- perhaps so rapidly as to undermine the encryption systems that are the foundation for electronic commerce on the Internet.
"Fortunately," Mr. Chuang says, "what quantum computing will take away with one hand it will give back with another."
And indeed, in the same research laboratory, tucked away in the rolling hills south of Silicon Valley, another group of scientists has developed a working data-scrambling system based on the same quantum phenomena. The system might make it possible to exchange information secretly without fear that the communications could be decoded by even the most powerful computer.
Openness to radically new ideas is common among the scientists here. They also seem to be acutely aware that fundamental advances can neither be engineered nor planned for.
For example, in the early 1990's an I.B.M. scientist, Stuart Parkin, stumbled upon the crucial material necessary to create a phenomenon known as gigantic magneto resistance, or G.M.R., when someone in his lab incorrectly filled in a table on a computer screen. This discovery of a thin sandwich of materials led to a new ultrasensitive sensor for disk drives, permitting a vast increase in the amount of information that can be stored and retrieved by a computer. Now Mr. Parkin has set his sights on a new data storage application for magnetism based on the quantum forces Mr. Chuang is exploring. He is working on a new kind of memory chip -- known as magnetic random access memory -- which, if it can be made cheaply, might become the standard computer memory for future personal digital assistants and every other portable electronic device.
Mr. Parkin spends his days searching for the magic combination of superthin sandwiches of molecules that interact in the precisely right way. "It's a huge space to search," he said. "It involves serendipity, luck and hopefully some science."
A New Meaning for Automatic
It will tell him about traffic jams and predict where new ones probably will develop. It will watch for jaywalkers and warn him if he seems not to have noticed a stop sign ahead. It will listen for unusual engine noises, run diagnostic tests and report any problems. And if he gets caught in slow-moving highway traffic, his car can take over the controls, keeping a safe distance behind the vehicle ahead.
"It will be a dialogue," said Mr. Metzler, who heads research on "machine understanding" at DaimlerChrysler here. "The car will be taking information from its surroundings and putting it to use. That is the function of a brain."
Mr. Metzler is at the crossroad of automobiles and computing. It is a clash of different traditions: one, a neo-Newtonian world grounded in heavy metal and seemingly immutable laws of motion, inertia and gravity; the other, a quicksilver world of molecular electronics and algorithms in which the speed limits constantly double.
Cars have been getting smarter for years. On-board navigation systems can map the best route to a destination and tell drivers if they make a wrong turn. Some top-of-the-line Mercedes-Benz models can warn drivers when they get too close to other cars.
But for people like Mr. Metzler, that is just a start. He is convinced that the future of cars is in "drive-by-wire" systems that would fundamentally change the relationship between driver and car.
Consider Daimler's somewhat bizarre "side-stick" car, a sporty yellow Mercedes coupe that has no steering wheel and no pedals for either braking or accelerating. Instead, all controls are on joysticks to the right and left of the driver.
Press forward on either one or both of the joysticks and the car accelerates. Pull backward and the car brakes. Push to the right or left and the car turns accordingly.
Why do it? Daimler engineers say the joystick car is much simpler to control than traditional vehicles because everything can be done in one fluid move of the hand. Moreover, the on-board electronic systems simulate a feeling of the road conditions in the joystick, which moves more freely if the wheels are spinning on ice and more slowly if the car is starting up a steep incline.
"You have the feeling you have full control right in the grip of your hand," said Lutz Eckstein, 31, an engineer who wrote some of the algorithms and has tested the car.
In tests, Mr. Metzler said, 17-year-old novice drivers seemed to handle the car well. But people accustomed to conventional vehicles had a much harder time.
Daimler executives do not know whether they will ever try to market such a car. The idea, they say, is mainly to show that advanced electronics make it possible to think about car design in entirely new ways.
Social and political acceptance will also affect the future of ever more intelligent vehicles. Klaus-Dieter VĂ¶hringer, DaimlerChrysler's managing director in charge of research, is convinced that technology makes it possible to have accident-free traffic in the next 10 or 20 years.
But it may take much longer before drivers or regulators accept it. Cars could see and hear the world around them. They would not only warn drivers about trouble but would be allowed to take partial control.
One glimpse of the future is Daimler's experimental "urban transit assistant." Video cameras produce three-dimensional images of the road.
An on-board computer matches the stream of incoming images with a library of images of pedestrians, traffic signs and potential barriers. It then highlights important images of things to stop for and pay attention to in blue or red.
Many Daimler engineers think it would be dangerous in practice to flash such images in front of drivers. But the system could warn drivers with alarms or spoken words.
The bigger issues are about control and trust in technology. Should cars of the future simply do more to warn drivers or should they sometimes take over the controls?
Daimler is already campaigning for governments to approve what it calls an electronic tow bar, a system that would regulate the speed and distance between trucks in a convoy. The technology would enable trucks to reduce wind resistance and save fuel by bunching together, Daimler engineers say.
Looking further ahead, engineers here are convinced that cars will be able to anticipate collisions. Cars would also continuously monitor road conditions and surrounding traffic. They would beam a constant stream of reports over wireless networks to regional traffic authorities.
Drivers might give up some control. On highways, they could leave it to the cars to keep a safe distance from one another. In a possible collision, cars might even communicate on their avoidance strategy.
All this could detract from the thrill of being in full command of a high-performance car.
But it may happen anyway.
"Driving is fun, and customer preferences must always be on the agenda," Mr. VĂ¶hringer said. "But if there is a situation where the human being is not capable of reacting, then the technology should be able to react to dangers."