I picked it out for reference only: Academician Li Guojie: I think the development trend of computer science is usually divided into three dimensions. One dimension is the "high" direction. The performance is getting higher and higher, and the speed is getting faster and faster, mainly because the main frequency of computers is getting higher and higher. Like the 286,386 we used a few years ago, the main frequency is only tens of megabytes. In the early 1990s, the integration level of integrated circuits has reached more than 65.438+0 billion, and it has entered the ULSI period since VLSI. Moreover, due to the maturity and popularization of RISC technology, the annual growth rate of CPU performance has increased from 35% in the 1980s to 60% in the 1990s. Later, Pentium series appeared, and now Pentium 4 microprocessor appears, with the main frequency reaching more than 2GHz. Moreover, the development of computers to a higher level is not only the improvement of chip frequency, but also the improvement of the overall performance of computers. A computer may use hundreds or thousands of processors instead of just one processor, which is called parallel processing. In other words, there are two ways to improve the performance of the computer: one is to improve the speed of the equipment, and the other is parallel processing. As mentioned above, by inventing new devices (such as quantum devices) and adopting nanotechnology, system-on-chip and other technologies, the device speed can be increased by several orders of magnitude. The innovation and progress of architecture marked by large-scale parallel behavior is another important way to improve the performance of computer systems. At present, the world's highest-performance general-purpose computers have adopted tens of thousands of computers in parallel, and ASCI of the United States plans to complete 12 per second. 3 trillion parallel machines. At present, 30 trillion parallel computers and 100 trillion parallel computers are being developed. Another plan of the United States aims to launch Petaflops computers at around 20 10, and its processors will use superconducting quantum devices, each processor will run at 1000 billion times per second, and * * * will use 65,438+million processors for parallel. The parallelism of special computer is higher than that of general computer. IBM is developing a special computer for calculating protein folding structure, called Blue Gene Computer. A chip includes 32 processors, and the peak speed reaches 1 trillion times per second. It is planned to be realized in 2004. Connecting tens of thousands of computers to form a parallel machine is by no means an easy task, just like organizing thousands of workers to produce a product. The key technology of parallel computer is how to efficiently connect a large number of computers, that is, high-speed communication between processors, and how to effectively manage the coordination of thousands of computers. This is the function of operating system, the system software of parallel computer. How to deal with the contradiction between high performance and universality and the portability of application software is also a technical choice that must be faced in the development of parallel computers, and it is also an important topic in the development of computer science. The other direction is to develop in a "wide" direction. The trend of computer development is so ubiquitous that it is like "no computer". In recent years, the obvious trend is networking, infiltrating into various fields, that is, developing and developing in breadth. Abroad, this trend is called pervasive computing or pervasive computing. For example, how many motors are there in your home? No one can tell. It is found in washing machines, refrigerators and tape recorders, almost everywhere. None of us can count. In the future, computers will exist in all kinds of electrical appliances at home, just like motors now. At that time, I asked you how many computers you had at home, but you couldn't count them. Your notebook and books have been digitized. Including textbooks for primary and secondary schools in the future, after ten or twenty years, students may no longer use textbooks in class, but just a notebook-sized computer, in which all curriculum textbooks, tutorials and exercises for primary and secondary schools are located. Different students can easily find the information they want according to their own needs. Moreover, these computers are integrated with the current mobile phones, and they can surf the Internet and exchange information with each other anytime and anywhere. Therefore, some people predict that computers may be as cheap as paper and can be used once, and computers will become the most commonly used and unnoticed daily necessities. The third direction is to develop in the "depth" direction, that is, to develop intelligent information. There is a lot of information on the Internet. How to turn these vast things into the knowledge you want is an important topic in computing science, and the man-machine interface is more friendly. In the future, you can use your natural language to deal with computers, you can also use handwritten words to deal with computers, and you can even use your expressions and gestures to communicate with computers, making computer communication more convenient and fast. Since its birth, electronic computers have been devoted to simulating human thinking, hoping that computers will become more and more intelligent, not only to do some complicated things, but also to do things that require "wisdom", such as reasoning, learning, association and so on. Since 1956 put forward "artificial intelligence", the pace of computers in the direction of intelligence is not satisfactory. Scientists have failed to achieve the expected goal of artificial intelligence many times, which shows that it is a very arduous task to explore the essence of human intelligence. At present, the way of computer "thinking" is very different from that of human beings, and the interval between man and machine is not small. It is still difficult for human beings to deal with computers in natural ways, such as language, gestures and expressions. The difficulty in using computers has become a huge obstacle to the further popularization of computers. With the popularity of the Internet, the demand for ordinary people to use computers is increasing day by day, which will greatly promote the research of computer intelligence. In recent years, the technology of computer recognition of characters (including printed and handwritten) and spoken language has been greatly improved and has reached the level of commercialization. It is predicted that in 5- 10 years, handwriting and oral input will gradually become the mainstream input methods. Gestures (especially sign language gestures) and facial expression recognition have also made great progress.