Many people believe that Stonehenge, the famous collection of rock monoliths in Great Britain, is an early form of a calendar or astrological calculator. The abacus, which appeared in the sixteenth century BC, was developed as an instrument to record numeric values and on which a human can perform basic arithmetic.

In the middle of the seventeenth century, Blaise Pascal, a French mathematician, built and sold gear-driven mechanical machines, which performed whole-number addition and subtraction. Later in the seventeenth century, a German mathematician, Gottfried Wilhelm von Leibniz, built the first mechanical device designed to do all four whole-number operations: addition, subtraction, multiplication, and division. Unfortunately, the state of mechanical gears and levers at that time was such that the Leibniz machine was not very reliable.

In the late eighteenth century, Joseph Jacquard developed what became known as Jacquard’s loom, used for weaving cloth. The loom used a series of cards with holes punched in them to specify the use of specific colored thread and therefore dictate the design that was woven into the cloth. Although not a computing device, Jacquard’s loom was the first to make use of what later became an important form of input: the punched card.

It wasn’t until the nineteenth century that the next major step was taken in computing, this time by a British mathematician. Charles Babbage designed what he called his analytical engine. His design was too complex for him to build with the technology of his day, so it was never implemented. His vision, however, included many of the important components of today’s computers. Babbage’s design was the first to include a memory so that intermediate values did not have to be reentered. His design also included the input of both numbers and mechanical steps, making use of punched cards similar to those used in Jacquard’s loom.

Ada Augusta, Countess of Lovelace, was a romantic figure in the history of computing. Ada, the daughter of Lord Byron (the English poet), was a skilled mathematician. She became interested in Babbage’s work on the analytical engine and extended his ideas (as well as correcting some of his errors). Ada is credited with being the first programmer. The concept of the loop—a series of instructions that repeat—is attributed to her. The programming language Ada, used largely by the U.S. Department of Defense for many years, is named for her.

During the later part of the nineteenth century and the beginning of the twentieth century, computing advances were made rapidly. William Burroughs produced and sold a mechanical adding machine. Dr. Herman Hollerith developed the first electro-mechanical tabulator, which read information from a punched card. His device revolutionized the census taken every ten years in the United States. Hollerith later formed a company known today as IBM.

In 1936, a theoretical development took place that had nothing to do with hardware per se but profoundly influenced the field of computer science. Alan M. Turing, another British mathematician, invented an abstract mathematical model called a Turing machine, laying the foundation for a major area of computing theory. The most prestigious award given in computer science (equivalent to the Fields Medal in mathematics or a Nobel Prize in other sciences) is the Turing Award, named for Alan Turing. The 2014 movie The Imitation Game is based on his life. Analysis of the capabilities of Turing machines is a part of the theoretical studies of all computer science students.

In the mid to late 1930s, work on building a computing machine continued around the world. In 1937, George Stibitz constructed a 1-bit binary adder, a device that adds binary digits, using relays (see Chapter 4). Later that year, Claude E. Shannon published a paper about implementing symbolic logic using relays. In 1938, Konrad Zuse built the first mechanical binary programmable computer. A biography of Konrad Zuse is presented in Chapter 6.

By the outbreak of World War II, several general-purpose computers were under design and construction. In London in 1943, Thomas Flowers built the Colossus, shown in FIGURE 1.4 and considered by many to be the first all-programmable electronic digital computer. In 1944, the IBM Automatic Sequence Controlled Calculator was given to Harvard; it was subsequently known as the Harvard Mark I. The ENIAC, pictured in FIGURE 1.5, was unveiled in 1946. John von Neumann, who had served as a consultant on the ENIAC project, started work on another machine known as EDVAC, which was completed in 1950. In 1951, the first commercial computer, UNIVAC I, was delivered to the U.S. Bureau of the Census. The UNIVAC I was the first computer used to predict the outcome of a presidential election.⁴

The early history that began with the abacus ended with the delivery of the UNIVAC I. With the building of that machine, the dream of a device that could rapidly manipulate numbers was realized. Some experts predicted at that time that a small number of computers would be able to handle the computational needs of mankind. What they didn’t realize was that the ability to perform fast calculations on large amounts of data would radically change the very nature of fields such as mathematics, physics, engineering, and economics. That is, computers made those experts’ assessments of what needed to be calculated entirely invalid.⁵

After 1951, the story becomes one of the ever-expanding use of computers to solve problems in all areas. From that point, the search has focused not only on building faster, bigger devices, but also on developing tools that allow us to use these devices more productively. The history of computing hardware from this point on is categorized into several “generations” based on the technology they employed.

Commercial computers in the first generation (from approximately 1951 to 1959) were built using vacuum tubes to store information. A vacuum tube, shown in FIGURE 1.6, generated a great deal of heat and was not very reliable. The machines that used them required heavy-duty air conditioning and frequent maintenance. They also required large, specially built rooms.

The primary memory device of this first generation of computers was a magnetic drum that rotated under a read/write head. When the memory cell being accessed rotated under the read/write head, the data was written to or read from that place.

The input device used by these computers was a card reader that read the holes punched in an IBM card (a descendant of the Hollerith card). The output device was either a punched card or a line printer. By the end of this generation, magnetic tape drives had been developed that were much faster than card readers. Magnetic tapes are sequential storage devices, meaning that the data on the tape must be accessed one after another in a linear fashion.

Storage devices external to the computer memory are called auxiliary storage devices. The magnetic tape was the first of these devices. Collectively, input devices, output devices, and auxiliary storage devices became known as peripheral devices.

The advent of the transistor (for which John Bardeen, Walter H. Brattain, and William B. Shockley won a Nobel Prize) ushered in the second generation of commercial computers. The transistor replaced the vacuum tube as the main component in the hardware. The transistor, as shown in FIGURE 1.7, was smaller, faster, more durable, and cheaper.

The second generation also witnessed the advent of immediate-access memory. When accessing information from a drum, the CPU had to wait for the proper place to rotate under the read/write head. The second generation used memory made from magnetic cores, tiny doughnut-shaped devices, each capable of storing one bit of information. These cores were strung together with wires to form cells, and cells were combined into a memory unit. Because the device was motionless and was accessed electronically, information was available instantly.

The magnetic disk, a new auxiliary storage device, was also developed during the second computer hardware generation. The magnetic disk is faster than magnetic tape because each data item can be accessed directly by referring to its location on the disk. Unlike a tape, which cannot access a piece of data without accessing everything on the tape that comes before it, a disk is organized so that each piece of data has its own location identifier, called an address. The read/write heads of a magnetic disk can be sent directly to the specific location on the disk where the desired information is stored.

In the second generation, transistors and other components for the computer were assembled by hand on printed circuit boards. The third generation was characterized by integrated circuits (ICs), solid pieces of silicon that contained the transistors, other components, and their connections. Integrated circuits were much smaller, cheaper, faster, and more reliable than printed circuit boards. Gordon Moore, one of the co-founders of Intel, noted that from the time of the invention of the IC, the number of circuits that could be placed on a single integrated circuit was doubling each year. This observation became known as Moore’s law.⁸

Transistors also were used for memory construction, where each transistor represented one bit of information. Integrated-circuit technology allowed memory boards to be built using transistors. Auxiliary storage devices were still needed because transistor memory was volatile; that is, the information went away when the power was turned off.

The terminal, an input/output device with a keyboard and screen, was introduced during this generation. The keyboard gave the user direct access to the computer, and the screen provided an immediate response.

Large-scale integration characterizes the fourth generation. From several thousand transistors on a silicon chip in the early 1970s, we had moved to a whole microcomputer on a chip by the middle of this decade. Main memory devices are still made almost exclusively out of chip technology. Over the previous 40 years, each generation of computer hardware had become more powerful in a smaller package at lower cost. Moore’s law was modified to say that chip density was doubling every 18 months.

By the late 1970s, the phrase personal computer (PC) had entered the vocabulary. Microcomputers had become so cheap that almost anyone could have one, and a generation of kids grew up playing Pac-Man.

The fourth generation found some new names entering the commercial market. Apple, Tandy/Radio Shack, Atari, Commodore, and Sun joined the big companies of earlier generations—IBM, Remington Rand, NCR, DEC (Digital Equipment Corporation), Hewlett-Packard, Control Data, and Burroughs. The best-known success story of the personal computer revolution is that of Apple. Steve Wozniak, an engineer, and Steve Jobs, a high school student, created a personal computer kit and marketed it out of a garage. This was the beginning of Apple Computer, a multi billion-dollar company.

The IBM PC was introduced in 1981 and was soon followed by compatible machines manufactured by many other companies. For example, Dell and Compaq were successful in making PCs that were compatible with IBM PCs. Apple introduced its very popular Macintosh microcomputer line in 1984.

In the mid-1980s, larger, more powerful machines were created; they were referred to as workstations. Workstations were generally meant for business, not personal, use. The idea was for each employee to have his or her own workstation on the desktop. These workstations were connected by cables, or networked, so that they could interact with one another. Workstations were made more powerful by the introduction of the RISC (reduced-instruction-set computer) architecture. Each computer was designed to understand a set of instructions, called its machine language. Conventional machines such as the IBM 370/168 had an instruction set containing more than 200 instructions. Instructions were fast and memory access was slow, so specialized instructions made sense. As memory access got increasingly faster, using a reduced set of instructions became attractive. Sun Microsystems introduced a workstation with a RISC chip in 1987. Its enduring popularity proved the feasibility of the RISC chip. These workstations were often called UNIX workstations because they used the UNIX operating system.

Because computers are still being made using circuit boards, we cannot mark the end of this generation. However, several things have occurred that so dramatically affected how we use machines that they certainly have ushered in a new era. Moore’s law was once again restated in the following form: “Computers will either double in power at the same price or halve in cost for the same power every 18 months.”⁹

Although computers that use a single primary processing unit continue to flourish, radically new machine architectures began appearing in the late 1980s. Computers that use these parallel architectures rely on a set of interconnected central processing units.

One class of parallel machines is organized so that the processors all share the same memory unit. In another class of machines, each central processor has its own local memory and communicates with the others over a very fast internal network.

Parallel architectures offer several ways to increase the speed of execution. For example, a given step in a program can be separated into multiple pieces, and those pieces can be executed simultaneously on several individual processors. These machines are called SIMD (single-instruction, multiple-data-stream) computers. A second class of machines can work on different parts of a program simultaneously. These machines are called MIMD (multiple-instruction, multiple-data-stream) computers.

The potential of hundreds or even thousands of processors combined in one machine is enormous, and the challenge of programming for such machines is equally daunting. Software designed for parallel machines is different from software designed for sequential machines. Programmers have to rethink the ways in which they approach problem solving and programming to exploit parallelism.

In the 1980s, the concept of a large machine with many users gave way to a network of smaller machines connected so that they can share resources such as printers, software, and data. Ethernet, invented by Robert Metcalfe and David Boggs in 1973, used a cheap coaxial cable to connect the machines and a set of protocols to allow the machines to communicate with one another. By 1979, DEC, Intel, and Xerox joined to establish Ethernet as a standard.

Workstations were designed for networking, but networking personal computers didn’t become practical until a more advanced Intel chip was introduced in 1985. By 1989, Novell’s Netware connected PCs together with a file server, a computer with generous mass storage and good input/ output capability. Placing data and office automation software on the server rather than each PC having its own copy allowed for a measure of central control while giving each machine a measure of autonomy. Workstations or personal computers networked together became known as LANs (local area networks).

The Internet as we know it today is descended from the ARPANET, a government-sponsored network begun in the late 1960s, which originally consisted of 11 nodes concentrated mainly in the Los Angeles and Boston areas. Like ARPANET and LANs, the Internet uses packet switching, a way for messages to share lines. The Internet, however, is made up of many different networks across the world that communicate by using a common protocol, TCP/IP (Transmission Control Protocol/Internet Protocol).

Paul E. Ceruzzi, in A History of Modern Computing, comments on the relationship between Ethernet and the Internet:

Some of the most recent changes in the overall management and use of computing hardware are based on the increasing reliance on cloud computing, which is the use of computer resources on the Internet instead of relying on devices at your physical location. The computing hardware is somewhere “in the cloud” instead of on your desktop or even in your building.

Hardware problems make computing difficult, especially for businesses and other large-scale endeavors. People trying to solve problems don’t want to have additional problems related to the maintenance of computers and the need to upgrade to more advanced technology as it becomes available.

As networking became more available, businesses would increasingly rely on data centers to support their computing efforts. A data center is generally a third-party solution—a separate business providing computing hardware that another business would use. The personnel at the data center set up and configure the machines needed by the business.

The use of a data center offloaded a big part of the challenges of managing the necessary computing infrastructure, but the process could still be painful because of the communication necessary between the business and the data center. It would often take several days to set up a new computer (a server) needed by the business.

Eventually, the idea of a data center evolved into the concept of cloud computing, in which a business could “spin up” a new server in a matter of minutes using some commands through a web interface. The process is so much more streamlined now that many businesses don’t even consider owning and managing their own computing hardware.

Cloud computing is discussed further in Chapter 15.

The first programs were written using machine language, the instructions built into the electrical circuitry of a particular computer. Even the small task of adding two numbers together used three instructions written in binary (1s and 0s), and the programmer had to remember which combination of binary digits meant what. Programmers using machine language had to be very good with numbers and very detail oriented. It’s not surprising that the first programmers were mathematicians and engineers. It’s also not surprising that programming in machine language was both time-consuming and prone to errors.

Because writing in machine code is so tedious, some programmers took the time to develop tools to help with the programming process. Programmers began writing programs in assembly language, which used mnemonic codes to represent each machine-language instruction.

Because every program that is executed on a computer eventually must be in the form of the computer’s machine language, the developers of assembly language also created software translators to translate programs written in assembly language into machine code. A program called an assembler reads each of the program’s instructions in mnemonic form and translates it into the machine-language equivalent. These mnemonics are abbreviated and sometimes difficult to read, but they are much easier to use than long strings of binary digits. Assembly language was the first step toward the languages we use today to program computers.

The programmers who wrote these tools to make programming easier for others were the first systems programmers. So, even in first-generation software, there was the division between those programmers who wrote tools and those programmers who used the tools. The assembly language acted as a buffer between the programmer and the machine hardware. See FIGURE 1.8. Sometimes, when efficient code is essential, programs today may be written in assembly language. Chapter 6 explores an example of machine code and a corresponding assembly language in detail.

As hardware became more powerful, more powerful tools were needed to use it effectively. Assembly languages certainly presented a step in the right direction, but the programmer still was forced to think in terms of individual machine instructions. The second generation saw more powerful languages developed. These high-level languages allowed the programmer to write instructions using more English-like statements.

Two of the languages developed during the second generation include FORTRAN (a language designed for numerical applications) and COBOL (a language designed for business applications). FORTRAN and COBOL developed quite differently. FORTRAN started out as a simple language and grew as additional features were added to it over the years. In contrast, COBOL was designed first and then implemented. It has changed little over time.

Another language that was designed during this period is Lisp. Lisp differs markedly from FORTRAN and COBOL and was not widely accepted. It was used mainly in artificial intelligence applications and research. A dialect of Lisp called Scheme is one of the languages used today for programming modern artificial intelligence systems.

The introduction of high-level languages provided a vehicle for running the same program on more than one computer. As with assembly language, a program written in a high-level language has to be translated into machine instructions to be executed. Such a translator is called a compiler, and it also checks to ensure that the syntax of the high-level language is being followed correctly. A program written in FORTRAN or COBOL wasn’t tied to a particular type of computer. It could run on any machine that had an appropriate compiler.

At the end of the second generation, the role of the systems programmer was becoming more well-defined. Systems programmers wrote tools like assemblers and compilers; those people who used the tools to write programs were called applications programmers. The applications programmer was becoming even more insulated from the computer hardware as the software surrounding the hardware became more sophisticated. See FIGURE 1.9.

During the third generation of commercial computers, it became apparent that the human was slowing down the computing process. Computers were sitting idle while waiting for the computer operator to prepare the next job. The solution was to put the computer resources under the control of the computer—that is, to write a program that would determine which programs were run when. This kind of program was the first operating system.

During the first two computer software generations, utility programs had been written to handle often-needed tasks. Loaders loaded programs into memory and linkers linked pieces of large programs together. In the third generation, these utility programs were refined and put under the direction of the operating system. This group of utility programs, the operating system, and the language translators (assemblers and compilers) became known as systems software.

The introduction of computer terminals as input/output devices gave users ready access to computers, and advances in systems software gave machines the ability to work much faster. However, inputting and outputting data from keyboards and screens was a slow process, much slower than carrying out instructions in memory. The problem was how to make better use of the machine’s greater capabilities and speed. The solution was time sharing—many different users, each at a terminal, communicating with a single computer all at the same time. Controlling this process was an operating system that organized and scheduled the different jobs.

For the user, time sharing is much like having his or her own machine. Each user is assigned a small slice of central processing time and then is put on hold while another user is serviced. Users generally aren’t even aware that there are other users. However, if too many people try to use the system at the same time, there can be a noticeable wait for a job to be completed.

As part of the third generation, general-purpose application programs also were being written. One example was the Statistical Package for the Social Sciences (SPSS), which was written in FORTRAN. SPSS had a special language that allowed the user, who was often not a programmer, to describe some data and the statistics to be computed on that data.

At the beginning of the computer era, the computer user and the programmer were the same person. By the end of the first generation, programmers had emerged who wrote tools for other programmers to use, giving rise to the distinction between systems programmers and applications programmers. However, the programmer was still the user. In the third generation, systems programmers were writing software for nonprogrammers to use. For the first time, there were computer users who were not programmers in the traditional sense.

The separation between the user and the hardware was growing wider. The hardware had become an even smaller part of the picture. A computer system—a combination of hardware, software, and the data managed by them—had emerged. See FIGURE 1.10. Occasionally programmers still needed to access the computer at lower levels, but the tools available at higher levels changed the situation greatly.

The 1970s saw the introduction of better programming techniques called structured programming, a logical, disciplined approach to programming. The languages Pascal and Modula-2 were built on the principles of structured programming. BASIC, a language introduced for third-generation machines, was refined and upgraded to more structured versions. C, a language that allows the user to intersperse assembly-language statements in a high-level program, was also introduced. C++, a structured language that allows the user access to low-level statements as well, became the language of choice in the industry.

Better and more powerful operating systems were being developed, too. UNIX, developed at AT&T^™ as a research tool, has become standard in many university settings. PC-DOS, developed for the IBM PC, and MS-DOS, developed for PC compatibles, became standards for personal computers. Apple capitalized on research done at Xerox PARC by incorporating a mouse and point-and-click graphical interface into the operating system for the Macintosh, which ushered in an important change to computer–user interaction on personal computers.

High-quality, reasonably priced applications software packages became available at neighborhood stores. These programs allow the user with no computer experience to perform a specific task. Three types of application software introduced during this period are still crucial to computer use today: spreadsheets, word processors, and database management systems. Lotus 1-2-3 was the first commercially successful spreadsheet that allowed a novice user to enter and analyze all kinds of data. WordPerfect was one of the first word processors, and dBase IV was a system that let the user store, organize, and retrieve data.

The fifth generation is notable for three major events: the rise of Microsoft^® as a dominant player in computer software, object-oriented design and programming, and the World Wide Web.

Microsoft’s Windows operating system emerged as a major force in the PC market during this period. Although WordPerfect continued to improve, Microsoft Word became the most used word processing program. In the mid-1990s, word processors, spreadsheet programs, database programs, and other application programs were bundled together into super packages called office suites.

Object-oriented design became the design of choice for large programming projects. Whereas structured design is based on a hierarchy of tasks, object-oriented design is based on a hierarchy of data objects. Java^™, a language designed by Sun Microsystems for object-oriented programming, began to rival C++.

In 1990, Tim Berners-Lee, a British researcher at the CERN physics lab in Geneva, Switzerland, created a set of technical rules for what he hoped would be a universal Internet document center called the World Wide Web. Along with these rules, he created HTML, a language for formatting documents, and a rudimentary, text-only browser, a program that allows a user to access information from websites worldwide. In 1993, Marc Andreesen and Eric Bina released Mosaic, the first graphics-capable browser. To quote Newsweek: “Mosaic just may have been the most important computer application ever.”¹⁶

At this point, there were two giants in the browser market: Netscape Navigator (derived from Mosaic) and Microsoft’s Internet Explorer (IE). Microsoft bundled IE with its Windows operating system, which made IE the winner in the browser wars. This bundling led to a monopoly lawsuit filed by the U.S. government, the 2001 settlement of which required Microsoft to be more open with its competitors. Netscape’s future became uncertain after America Online purchased it in 1998, and it was eventually discontinued. Microsoft eventually replaced Internet Explorer with a browser called Edge.

Firefox, a web browser that retained some of the flavor of Mosaic, was released in 2004. Around the same time, Apple released its Safari browser. In 2008, Google announced its browser called Chrome. All of these now claim their share of the browser market.

Although the Internet had been around for decades, the World Wide Web made it easy to use the Internet to share information around the world. User-generated and user-edited content became the norm. Online blogging allowed anyone to become an author or a critic. Social networking has revolutionized how we interact with other people. Websites such as the online encyclopedia Wikipedia provide rich content that anyone can add to and edit.

The fifth generation must be characterized most of all by the changing profile of the user. The first user was the programmer who wrote programs to solve specific problems—his or her own or someone else’s. Then the systems programmer emerged, who wrote more and more complex tools for other programmers. By the early 1970s, applications programmers were using these complex tools to write applications programs for nonprogrammers to use. With the advent of the personal computer, computer games, educational programs, and user-friendly software packages, many people became computer users. With the birth and expansion of the World Wide Web, web surfing has become the recreation of choice, so even more people have become computer users. The user is a first-grade child learning to read, a teenager downloading music, a college student writing a paper, a homemaker planning a budget, and a banker looking up a customer’s loan record. The user is all of us.

In our brief history of hardware and software, we have focused our attention on traditional computers and computing systems. Paralleling this history is the growing use of integrated circuits, or chips, to run or regulate everything from toasters to cars to intensive care monitors to satellites. Such computing technology is called an embedded system. Although these chips are not actually computers in the sense that we are going to study in this book, they are certainly a product of the technology revolution of the past 55 years.