To gain an appreciation of how one builds optical fibers for different communications applications, look at Figure A-1.8.1, which illustrates three different types of fibers.

Each of these types is distinguished by its refractive index profile. The refractive index profile is a key part of fiber design because it determines the product's ability to propagate short pulses and maintain their shapes over long distances.

The topmost illustration shows the simplest stepped-index profile fiber (also called multimode). This is the least costly type to manufacture, but it has the poorest pulse propagation characteristics. It is used for transmission over short distances.

Single mode fiber with a tailored stepped refractive index profile, shown in the second illustration, is much better at keeping the pulse shape intact, but it is costlier. This type of fiber is used for high-data-rate communications over long distances.

Finally, in the illustration at the bottom of the figure, we have a multimode, graded-index fiber that falls between the two profiles above it. This fiber has applications in intermediate distance transmission.

In all three examples the glass fibers are encased in claddings. Large numbers of individual cladded fibers are enclosed in cables that protect them from external forces and the elements. Special techniques exist for creating cables that can survive undersea deployment and other demanding environments.

In Chapter 1 we pointed out the difficulty of using vacuum tubes for large-scale digital data processing.

This observation, while perfectly true, begs the question, why shift to digital in the first place? We live in the natural world, which is defiantly analog. It seems logical that analog systems would process information from the analog world more efficiently than their digital counterparts.

On the surface they do seem more efficient. In analog electronic systems a smoothly variable electrical quantity (such as voltage, capacitance, or current) is used to represent information. The value of the electrical property is directly proportional to the magnitude of the physical signal being processed or transmitted.

Suppose, for example, you want to measure the intensity of light in a room. You place a photodetector (an analog light sensor) there. When light hits the device, it responds either by generating an electrical current or by permitting such a current to flow. As the light gets stronger, the current flowing in the system becomes greater as well.

The current increases in a linear fashion, allowing you to track and record the level of light falling on the photodetector. The term “linear” is used to describe analog systems because, under ideal conditions, their electrical information varies in line with the received signal.

Human beings are analog systems, too. We respond in a linear fashion to external analog signals such as temperature, sound, light, and images.

The fundamental physics of laser operation are similar for all media capable of laser operation, whether gases, insulators, or semiconductors. However, the operating details are vastly different in the various materials.

The early lasers were made using special solid-state materials or gases confined within glass envelopes. For example, important gas lasers use helium-neon or argon gas mixtures enclosed in a large glass tube 10 centimeters or more long. These are obviously bulky but useful as light sources when certain emission colors or high power are required – but not for the major applications that awaited semiconductor lasers.

The basic attraction of semiconductor lasers is their high atomic density, which makes possible laser operation in a volume which is ten million times smaller than in a gas. The idea of using semiconductors was discussed theoretically in the 1950s, but it took until 1962 for a laser diode device to be demonstrated.

In semiconductors, lasing is possible only in certain classes of materials endowed with special natural properties – direct bandgap semiconductors. Lasing occurs as a result of a highly complex interactive process between very high densities of electrons and holes confined within a region where the intense radiation released by their recombination is also confined.

Laser emission has been obtained in semiconductor laser diodes in a spectral region ranging from the blue to the far-infrared.

To get a feel for object-oriented architectures, consider a very simple program designed to update the contents of a particular master file. Figure A-2.3.1 shows the “old” way of architecting such a program, by logical decomposition of the process into steps that are sequentially linked to produce the program. To produce this program, each element of the process must be coded in tightly-linked sequences.

By contrast, in the object-oriented architecture shown in Figure A-2.3.2, the program is divided into abstractions of the key elements. Instead of steps, the problem is now defined in terms of autonomous interconnected objects. Each object has its real-world mission defined, and is given its operational orders and required variables through the bus.

Program development using object-oriented technology can be divided into four interdependent elements.

• The first critical task is the definition of program requirements. This requires close communication with the intended users and an understanding of their applications.

• Program architecture is developed on the basis of these requirements. The skill of the architect determines the ultimate quality of the program and the time required to complete it.

• The production of the program proceeds, with coding by programmers who write within tight guidelines provided by the program architects.

• Finally, the completed program must be tested. Note that testing can account for between 40 and 50 percent of the cost of developing a large program.

Program definition and architecture require the most skilled and experienced people. These two steps produce virtually all of the program's intellectual property value.

Flat-screen displays based on organic materials dominate the display device field. Chief among them for its commercial importance is the liquid crystal display, or LCD.

Liquid crystals are a class of organic materials with very interesting properties. Their molecular alignment (and hence their optical properties) change reversibly with temperature, as well as under the effect of moderate electric or magnetic fields.

At high temperatures these materials exhibit the properties of liquids, while below a critical temperature they assume more of the properties of solids. For example, as a result of this transformation, a liquid crystal film can change reversibly from a milky fluid, opaque to light, to one which is transparent to visible light.

The discovery of liquid crystals dates to the nineteenth century. However, the materials languished without serious applications until the late 1960s, when research at RCA Laboratories found an application for them in flat-screen displays. The researchers achieved this by applying electric fields to thin liquid crystal films to change the optical properties of the materials.

This discovery made two types of display applications possible as materials were developed. First, applying an electrical field to certain classes of liquid crystals changes them from light absorbers to good reflectors. Second, in other materials, the applied field can make the film either opaque or transparent to visible light; hence the film can become a “light valve.”

Flat-screen display technology has proven to be crucial to the mass acceptance of sophisticated electronic systems.

This book is about the creation and consequences of digital electronics.

Over the last half-century no other area of technical innovation has so drastically altered the way we live, work, and interact. Digital electronics give us instant access to a whole universe of information, and powerful tools to process it. They also equip us with an unprecedented ability to communicate instantly with people anywhere in the world.

I was privileged to contribute to these developments, first as a scientist engaged in creating new electronic devices and systems, and then as a venture capitalist involved in creating successful new businesses.

From these vantage points, I watched digital electronics grow from a technical specialization into a revolutionary force that generates new industries and transforms developing economies into global competitors, and does both seemingly overnight.

To sense the full extent of its impact, first visualize a contemporary scenario. A programming group in India is working on a project from the US. They are using computers built in China, with LCD displays made in Korea, driven by software from Silicon Valley and Germany. They communicate with colleagues in the US and other countries over the Internet, by e-mail and voice. It all seems quite normal today.

Now try to picture the same situation twenty-five years ago. You can't. There was little global sourcing of technical expertise then. High-tech manufacturing in developing Asian economies was just getting started. The major sources of equipment were Japan, the US, and Europe.

In Chapter 1, we described the three basic logic gates used to build computers. Here we will illustrate a combination gate – a NAND built with MOSFETs. As shown in Figure A-1.2.1, it takes three transistors of the same n-channel type in this configuration. A CMOS configuration for the same gate function requires four transistors.

The NAND is a combination of an AND gate and an inverter. The top MOSFET is connected to a voltage source of 10 volts which biases it in the on (conducting) state.

The truth table gives the logical combinations of this gate.

If both voltage A and voltage B are positive (1), then the bottom two transistors are turned on and hence conducting. This effectively connects the current flow from the voltage source through the top transistor to ground, and the output voltage V is 0. Hence it takes two positive (1) inputs to produce a (0) output.

On the other hand, if either of the two voltages at A or B is 0, then the current path is connected to the output, resulting in 10 volts at the terminal, which is defined as a “1.”

Most of the revolutionary innovations in semiconductors and software explored in the preceding two chapters came out of the corporate R&D organizations of AT&T, General Electric, IBM, Corning, RCA, and other large companies.

That doesn't mean these breakthroughs reflect a corporate mindset. Nothing could be further from the truth. The scientists who created them were following their own paths, not some managerial directive or mandated marketing strategy. As someone who began his career in this environment, I can attest to its creative ferment.

Truly creative people tend to be free spirits. They are the essential spark plugs of innovation. At the same time, their drive for discovery is often in conflict with practical considerations such as company direction or fiscal constraints. This trait presents corporate R&D managers with endless challenges.

Perhaps that is why the talk today is all about “small, innovative companies,” not big, powerful corporate laboratories (labs). But the fact is that for much of the twentieth century those big labs were the primary generators of new technology, because they had the resources and the funding necessary for research. They were the vehicles innovators needed to carry their work forward.

We've begun our study of digital electronics by looking at how the technology works, and how it affects people and businesses.

Semiconductor light emitting devices are grouped into two categories:

• Light emitting diodes (LEDs), which are basically general-purpose replacements for vacuum tube-based light emitters.

• Laser diodes that emit sharply focused monochromatic light beams.

We have already discussed laser diodes. Here we will look briefly at LEDs, which have come to dominate important segments of the lighting industry.

Light emitting diodes consist of p-n junctions in a special class of semiconductors that are capable of emitting light of many colors ranging from the infrared into the violet. Furthermore, of great practical importance is the fact that by combining LEDs of red, blue, and green, or using special phosphor coverings, white light can be produced, with conversion efficiencies from electrical input to useful light output that are very competitive with conventional vacuum-based devices.

When we discussed p-n junctions, we noted that one of their most valuable properties is their ability to inject carriers across the p-n interface. This is the principle behind LEDs. When forward-biased, they emit various colors of light, ranging from the blue into the infrared, depending on the semiconductor materials used.

Semiconductors useful as LEDs and lasers have special structures. As noted earlier when we discussed lasers, they are direct bandgap materials. Within them the recombination of an injected electron with a hole (in effect the capture of the free electron by an empty spot in the atomic outer orbit) results in the release of energy in the form of a photon. A photon is a unit of light energy.

Made in China. Or in Malaysia, Sri Lanka, Singapore – the list goes on.

Discovering that your clothing was sewn in a newly industrializing Asian country no longer raises eyebrows. But now it seems that even the most advanced electronic products are produced in a developing country, and that the manufacturing job is an endangered species in the US and Europe.

This is an exaggeration, of course. Yet the kernel of truth within it is undeniable: there has been an extraordinary migration of manufacturing, most notably of electronics, out of developed countries since the 1980s. What's more alarming is that the trend shows no signs of slackening.

To a large extent the rapid globalization of manufacturing is a consequence of digital technologies reaching maturity. Technological changes have altered the nature of the competitive advantage derived from manufacturing.

To take just one prominent example, contract manufacturing as a stand-alone business has become a major global activity. Developed in Asia, it represents a total reversal of the historical business structure, under which integrated manufacturing was considered a competitive advantage for practically all product suppliers.

Furthermore, as we discuss later, even software writing has become part of the outsourced “manufacturing” sector, a development made possible by the remarkable advances made in the technology of software development.

This chapter will trace the trajectory of outsourcing through product sectors and whole countries, revealing the pattern of exporting low-skill assembly jobs to foreign shores in search of cost reductions.

Semiconductor devices for sensing light levels are in wide use, and their properties are tailored to specific applications. They range in complexity from the simplest light sensors, used in door openers, to sophisticated devices that detect ultra-short, very low-level communications signals that have traveled hundreds of miles over an optical fiber.

As mentioned in the main text, light sensors are also the core sensing element in imaging devices, where millions of sensors are built into a single chip. Finally, solar energy converters (solar cells) are in the same class of device.

We learned in Chapter 2 that a photodetector is usually a reverse-biased p-n junction. The type of semiconductor material used for the junction is determined by its intended application, particularly with regard to the wavelength of the light to be detected and the response speed desired.

The operational principle is that a photon of incident light absorbed in the vicinity of the p-n junction, and within the depletion zone, generates free carriers that give rise to a current proportional to the intensity of the incident light. Figure A-1.7.1 shows the p-n junction detector in the dark and with incident light. The current through the device increases with light intensity.

Silicon photodetectors are widely used to capture visible light, whereas III-V compound devices are used to detect signals in the infrared portion of the spectrum.

My career as a researcher gave me a front-row seat during the period when electronics quickly advanced from primitive prototype transistors and lasers to incredibly tiny devices, sophisticated software-driven systems, and vast computing and communication networks. It was an exciting time to be a scientist, and to make my own contributions to the progress of the field.

As a venture capital manager, I've had an equally fortunate vantage point. I saw digital technology become a global force, delivering new services and creating opportunities where none had existed before. I also had the opportunity to arrange funding for innovative companies that helped move the technology forward.

No one in the 1950s and 1960s, when the fundamental discoveries were rolling out of US corporate laboratories, could have imagined that digital electronic technology would help to lift whole countries from isolation and poverty into participation in the global economy. But I've seen first hand how these revolutionary innovations have improved the human condition and even realigned the economic order among nations.

Today we live in a global village sustained by digital computing and communication technologies. It is a much more complex world, technologically, socially, and politically. To conclude this study of the nature and influence of digital technology, we will look at what the globalization of technology means for developed and developing nations alike, and how innovation and manufacturing can build prosperity at both stages of economic progress.

We'll begin with an assessment of the current situation.

Technology, like life, evolves in stages. The genesis of digital electronics was a handful of simple yet profound innovations, the silicon counterparts of single-celled organisms in primordial seas. Soon successive generations of the technology, each more complex than the last, appeared at unheard-of speed and spread across the face of the earth.

Unlike the progenitor of life, the creative spirit behind the digital electronics revolution is easy to identify. It is the genius of innovation, embodied in the visionary researchers who developed solid-state electronics, ICs, sensors, and software.

Fifty years ago these innovators were concentrated within the pioneering R&D organizations of a few American corporations. They invented the transistor, fiber optics, the microprocessor, and the semiconductor laser. They created new software structures and languages.

In Schumpeter's sense of the term, these were revolutionary innovations, with far-reaching (and largely unforeseen) consequences.

Digital processing needs lots of logic gates. MOSFET transistors, as noted in the main text, are the preferred way to create them. However, owing to excessive current leakage, early gates built with MOSFETs suffered from high power dissipation. Put millions of these devices on a chip and the constant current drain could cause overheating.

The CMOS architecture provided the solution to this problem, at the expense of more complex chip manufacturing, by dramatically reducing the gate power dissipation, enabling very large scale integrated circuits.

To demonstrate the principle behind CMOS, let's examine the differences between a simple inverter gate built without this architecture and one built with it.

We start with the non-CMOS inverter built with just one kind of MOSFET. As Figure A-1.4.1 shows, two n-channel transistors are interconnected with the source of the bottom transistor connected to the drain of the top transistor. The top transistor is always on since its gate is always connected to the battery.

Look at the left side of the figure. When a value of 10 volts (the input voltage) is applied to the bottom transistor gate, a (1), it turns on, forming a conducting path from battery to ground. Therefore, the output voltage is zero, indicating a (0) state.

Now look at the right side of the figure. The input gate voltage is zero, the bottom transistor is off, and the output is connected to the battery through the top transistor. Hence the output voltage is 10 volts, a (1).

By the early 1970s, most of the basic electronic development that made digital technology feasible was in place. One great burst of innovation over two decades had produced the transistor, the p-n junction, the CMOS process for integrated circuits, magnetic disk storage, heterojunction semiconductor lasers, fiber optic communications, and semiconductor imagers and displays.

The speed with which researchers put the new devices to practical use was almost as remarkable. They quickly developed an array of systems that gather, process, and transmit information.

The most visible example was the computer. In 1982, Time magazine focused popular attention on the computer's importance by naming it “Machine of the Year,” a widely discussed departure from the “Man of the Year” honor the magazine had bestowed for over sixty years.

Since then computers have become a part of our lives. Smaller, more powerful microprocessors and supporting devices, combined with plummeting costs, have sparked an endless stream of new computerized consumer and industrial products.

However, while the computer has been central to the Information Age, the most revolutionary transformation has taken place in communications. Data transmission, once limited to 300 bits per second on analog phone lines, now uses digital links to carry multi-megabit per second data streams. Voice has gone digital and wireless. The Internet has metamorphosed in just ten years from an “insider” network for scientists to an indispensable, ubiquitous communications backbone linking billions of people around the globe.

And both computing and communications have been enabled by advances in software architecture.

Just when global competition forced the big vertically-integrated US corporations to focus their R&D on near-term products, a new support system for innovation emerged: venture capital.

In the early 1980s, the venture capital industry became a major catalyst in bringing to market digital technologies that had been incubated in the corporate laboratories. In the mid-1990s came the rush of venture investments in start-up Internet companies.

Much as we might deplore the excesses of the Internet bubble, the many billions of dollars invested there left a permanent legacy of achievement. It is doubtful that the Internet would have developed as rapidly as it did, bringing with it the benefits that we enjoy today, without venture capital. About $200 billion of funding was provided between 1984 and 2005 to build thousands of new companies.

Having entered the venture capital business in 1983, I had a ringside seat as these events unfolded. I participated in the successes as well as the heartbreaks that come with revolutionary periods in history.

In hindsight it looks obvious that the new digital technologies would find willing investors. That was not the case. In fact, many promising companies struggled in their early years because the markets were so ill-defined.

But trailblazers in every era have a hard time finding investors to fund their exploits. The difficulties Columbus faced in financing his journeys are well known.