第1章コンピューターサイエンス入門

The term computer science was popularized by George E. Forsythe in 1961. A mathematician who founded Stanford University’s computer science department, Forsythe defined computer science as “the theory of programming, numerical analysis, data processing, and the design of computer systems.” He also argued that computer science was distinguished from other disciplines by the emphasis on algorithms, which are essential for effective computer programming.³

Computer science is not only about the study of how computers work, but also everything surrounding computers, including the people who design computers, the people who write programs that run on computers, the people who test the programs to ensure correctness, and the people who are directly and indirectly affected by computers. In this way, computer science is as much about people and how they work with computers as it is about just computers.

Not everyone agrees with this definition. Some people argue that computer science is more about computers or software than the people it affects. However, even if we were to study just the “things” of computer science, the people are still there. When someone designs a computer system, they are thinking about what kinds of programs people might want to run. Typically, effort is made to design the computer system so it is more efficient at running certain kinds of programs. A computer optimized for calculating missile trajectories, for example, won’t be optimized for running social media apps.

Many computing innovations were initially developed for military research and communication purposes, including the predecessor to the Internet, the ARPANET (Figure 1.3).

1 S. Golomb, “Richard E. Bellman 1920–1984,” n.d. https://www.nae.edu/189177/RICHARD-E-BELLMAN-19201984 2 Geeks for Geeks, “Dynamic Programming or DP,” 2024. https://www.geeksforgeeks.org/dynamic-programming/ 3 D. E. Knuth, “George Forsythe and the Development of Computer Science,” Communications of the ACM, vol. 15, no.8, pp. 722–723. 1972. https://dl.acm.org/doi/pdf/10.1145/361532.361538

While computer science is about much more than just computers, it helps to know a bit more about computers because they are an important component of computer science. All computers are made of physical, real-world material that we refer to as hardware. Hardware—which has four components, including processor, memory, network, and storage—is the computer component that enables computations. The processor can be regarded as the computer’s “brain,” as it follows instructions from algorithms and processes data. The memory is a means of addressing information in a computer by storing it in consistent locations, while the network refers to the various technological devices that are connected and share information. The hardware and physical components of a computer that permanently house a computer’s data are called storage.

One way to understand computers is from a hardware perspective: computers leverage digital electronics and the physics of materials used to develop transistors. For example, many of today’s computers rely on the physical properties of a brittle, crystalline metalloid called silicon, which makes it suitable for representing information. The batteries that power many of today’s smartphones and mobile devices rely on lithium, a soft, silvery metal mostly harvested from minerals in Australia, Zimbabwe, and Brazil, as well as from continental brine deposits in Chile, Argentina, and Bolivia. Computer engineers combine these substances to build circuitry and information pathways at the microscopic scale to form the physical basis for modern computers.

However, the physical basis of computers was not always silicon. The Electronic Numerical Integrator and Computer (ENIAC) was completed in 1945, making it one of the earliest digital computers. The ENIAC operated on different physical principles. Instead of silicon, the ENIAC used the technology of a vacuum tube, a physical device like a light bulb that was used as memory in early digital computers. When the “light” in the vacuum tube is off, the vacuum tube represents the number 0. When the “light” is on, the vacuum tube represents the number 1. When thousands of vacuum tubes are combined in a logical way, we suddenly have memory. The ENIAC is notable in computer history because it was the first general-purpose computer, meaning that it could run not just a single program but rather any program specified by a programmer. The ENIAC was often run and programmed by women programmers (Figure 1.4). Despite its age and differences in hardware properties, it shares a fundamental and surprising similarity with modern computers. Anything that can be computed on today’s computers can also be computed by the ENIAC given the right circumstances—just trillions of times more slowly.

How is this possible? The algorithmic principles that determine how results are computed makes up software. Almost all computers, from the ENIAC to today’s computers, are considered Turing-complete (or Computationally Universal, as opposed to specialized computing devices such as scientific calculators) because they share the same fundamental model for computing results and every computer has the ability to run any algorithm. Alan Mathison Turing was an English mathematician who was highly influential in the development of theoretical computer science, which focuses on the mathematical processes behind software, and provided a formalization of the concepts of algorithm and computation with the Turing machine. A Turing-complete computer stores data in memory (either using vacuum tubes or silicon) and manipulates that data according to a computer program, which is an algorithm that can be run on a computer. These programs are represented using symbols and instructions written in a programming language consisting of symbols and instructions that can be interpreted by the computer. Programs are also stored in memory, which allows programmers to modify and improve programs by changing the instructions.

While both hardware and software are important to the practical operation of computers, computer science’s historical roots in mathematics also emphasize a third perspective. Whereas software focuses on the program details for solving problems with computers, theoretical computer science focuses on the mathematical processes behind software. The idea of Turing-completeness is a foundational concept in theoretical computer science, which considers how computers in general—not just the ENIAC or today’s computers, but even tomorrow’s computers that we haven’t yet invented—can solve problems. This theoretical perspective expands computer science knowledge by contributing ideas about (1) whether a problem can be computed by a Turing-complete computer at all, (2) how that problem might be computed using an algorithm, and (3) how quickly or efficiently a computer can run such an algorithm. The answers to these questions suggest the limits of what we can achieve with computers from a technical perspective: Using mathematical ideas, is it possible to use a computer to compute all problems? If the answer to a problem is yes, how much of a computing resource is needed to get the answer?

Clearly both humans and computers have their strengths and limitations. An example of a problem that humans can solve but computers struggle with is interpreting subtle emotions or making moral judgments in complex social situations. While computers can process data and recognize patterns, they cannot fully understand the nuances of human emotions or ethics, which often involve context, empathy, and experience. On the flip side, there are tasks that neither computers nor humans can perform, such as accurately predicting chaotic systems like long-term weather patterns. Despite advancements in artificial intelligence (AI), computer functions that perform tasks, such as visual perception and decision-making processes that usually are performed by human intelligence, these problems remain beyond our collective reach due to the inherent unpredictability and complexity of certain natural systems.

Theoretical computer science is often emphasized in undergraduate computer science programs because academic computer science emerged from mathematics, often to the detriment of perspectives that center on the social and technical values embodied by applications of computer technology. These perspectives, however, are gradually changing. Just as the design of ARPANET shaped the design of the Internet, computer scientists are also learning that the physical aspects of computer hardware determine what can be efficiently computed. For example, many of today’s artificial intelligence technologies rely on highly specialized computer hardware that is fundamentally different at the physical level compared to the general-purpose programmable silicon that has been the traditional focus of computer science. Organizations that develop human-computer interaction (HCI), a subfield of computer science that emphasizes the social aspects of computation, now host annual conferences that bring together thousands of researchers in academia and professionals in the industry. Computer science education is another subfield that emphasizes the cognitive, social, and communal aspects of learning computer science. Although these human-centered subfields are not yet in every computer science department, their increasing representation reflects computer scientists’ growing desire to serve not only more engaged students, but also a more engaged public in making sense of the values of computer technologies.