China DOS Union

Required reading for programmers:

Bill Gates, October 28, 1955: Bill Gates was born in Seattle, Washington, in the northwestern United States. His father, William Gates, was an outstanding lawyer, and his mother was a director of First Interstate Bank. While he was in school, Gates was sent to the strictly run Lakeside private school in Seattle. In 1973 Gates entered Harvard University, and in 1975 he automatically withdrew, writing a version of the BASIC computer language together with Paul Allen. Soon after, the two formally founded Microsoft, and he served as Microsoft's chief executive officer. He is now Microsoft's chairman and chief software architect.

There are many students and programmers invited to take part in this event today, and it reminds me of when I first began learning programming as I was growing up. The first computer I encountered when I was 13 was at my high school. It was not a real computer; it was actually just a terminal connected by telephone line to a large GE computer. This was back in 1968. At that time computer time was very expensive, so people were afraid to use it, worried something might go wrong. In fact there was a teacher then who created an infinite loop and as a result spent 200 dollars. After that, no other teachers at our school were willing to have anything to do with computers, so only I and a few other students continued using them. This also inspired me to think: how could computers be improved? How could they be made into a tool that everyone could use? Of course, we knew there were nowhere near as many programmers then as there are now, and competition in the computing industry was not nearly as intense. I was greatly encouraged, took part in many programming contests, and thought about the miracle of chip capacity doubling every two years, and the powerful tools that suitable software could produce on top of that. This was the true beginning of the personal computer in its real sense, and also a dream that Paul Allen and I had 33 years ago.

A few years after we had this dream, the first 8008 chip was born. Its capabilities were not especially powerful, but it was very clear that the future it pointed to was unbelievably miraculous. Microsoft was founded at the very beginning to write BASIC for machines using this chip. Programming has already gone through a long journey. At first the BASIC we wrote ran on 4KB machines, so optimizing every single part was extremely important. Of course, by using the hardware resources we have today, we can do things far more complex than that.

What you will have in the future is this: your programming will make business more effective, make work more efficient and more interesting, and even change the whole world. We call this era the digital age. Why do we say that? Why do we call the decade from 2000 to 2010 a very unique era? We say this because at the beginning of this era, people's use of personal computers was very limited, but by the end of this era it will be completely different. At the beginning of this era, there was only simple document processing and email. There was very little commercial activity on the Internet, at most reading things or recording meetings. Things people are very familiar with, like paying bills, taking photos, and music, were not handled by computers. But by 2010, people will naturally use computers to arrange schedules, read information, or take notes, so there will be tremendous changes. And this change will come through the continuous development of hardware and software, all built on the foundation created in the 1990s. That key foundational period we call the era of the personal computer. The number of personal computer users grew from several million to tens of millions, and by the end of that era we had 500 million people using personal computers. Among these, the MS-DOS standard for the first time ran uniformly across hardware from different companies, so MS-DOS marked the beginning of the software industry. In the 1990s it was widely accepted, and the history of software began to flourish from the zero starting point of the 1980s. In the early 1990s Microsoft's Windows graphical interface was accepted by people, and the appearance of Windows95 in 1995 was an important milestone. By the late 1990s, we had another new surprise: the Internet became the mainstream standard way for interconnecting computers. Internet standards are very important. Everything we do now is based on these standards, such as TCP/IP HTML; all of this has been established. But this foundation alone is not enough. We need to build a richer platform, secure and self-managing, so that you can focus on developing more software and achieve unprecedented efficiency. So this requires everyone to invest in building the platform, and that is very exciting.

The issues we discuss today will have a very great impact and will drastically change the way people do many things. In fact, it can entirely be said that the development of information technology has reshaped the whole world. For example, some jobs that could only be done inside a company can now be done outside the company, and some work that could only be done in the needed country can now be done anywhere in the world through the Internet. This progress is especially important in China. China has so many highly qualified and highly skilled people; they can create job opportunities not only for China, but also for the whole world. We can already see this now in China, India, and some other countries. But compared with the future, these are only the very surface-level changes. The shape of computers themselves will also change. I think many of you here have seen the latest Tablet-PC, which Microsoft launched in the second half of 2002. This kind of computer is very suitable in size for carrying around. Our research found that in order to make reading information from a screen as convenient as reading from paper, we not only need to increase resolution, we also need to put it in people's hands. When we read documents, we can change our reading angle, and we do not need to care where around us it comes from. It is not like when we look at a CRT or LCD; they are always in a fixed position. So together with our partners we invented this kind of Tablet-PC. For the development of computers, this was also the first time there was the concept of digital ink. You can use a pen in place of a mouse, and also use it to take notes, send messages, and record documents. There is also digital ink recognition technology, and this concept is very interesting. Let us imagine we are discussing sales and shipment quantities in a meeting. If it is on an ordinary sheet of paper, that is dead information, but if it is on a screen, it is live information; you can study it and share it with others. So in the end, activities such as reading newspapers, textbooks, magazines, and scientific journals will all be done online through devices such as the Tablet-PC.

When people want to watch TV, no matter what program they want to watch, or view or organize their photos, or enjoy the music they like, they will do all of this on computers. Communication will also not be only voice communication. We can share screen content with partners, and have video-image communication, so the relationship between computers and telephones will change, and there will be more complementarity between the two. Now pen and ink have already been incorporated into computers, and we will also soon realize computer speech recognition. Research on the latter is crucial for achieving a truly natural interface. With this technology, we will be able to conveniently record meetings, locate what was being said at a certain moment based on the sound, and quickly rewind as something very natural. Microsoft's research centers, including the one here in Beijing, are working hard on just this kind of extraordinary research. Companies will digitize the informal business procedures and steps they have long had, thereby making it possible to track the flow of information and formalize communication among employees and among application software systems. Therefore this will be a huge change. Some of you may ask whether such a vision is similar to the dreams people had for the Internet in the late 1990s. Yes, but at that time the people dreaming had not recognized the necessity of a new software platform: there needed to be standards for exchanging information among different companies, better software to help realize e-commerce and e-government, and stronger security, reliability, and scalability for these new application software systems.

Therefore these requirements are now being put before us today, including setting standards such as XML Web services. The XML Web services architecture was a big bet Microsoft made six years ago. We decided that all our products would adopt this architecture. This is not the first time we have made such a determination. In fact, we were the same way 10 years ago when we put our full effort into developing the graphical interface. At that time many people thought it was crazy, saying machines could not run that fast and that it would not bring any real benefits. Now people naturally use graphical interfaces in any system. Likewise, the development of XML Web services now, and the “·NET” platform based on XML Web services standards, will also be used naturally by people. But at present we must reconstruct this platform. We need to inspire a new generation of developers. We need to form the habit of thinking and programming in terms of XML Web services. For us, that means changing Windows Database Microsoft Office and so on, everything we do. We need industry standards for Web services to be widely accepted. When we made this big bet in 1997, we took an enormous risk. But that is no longer the case today. Leading companies in the IT industry have all recognized the necessity of XML Web services for realizing e-commerce and e-government, have shown great interest, and are jointly working to develop industry standards.

Of course, protocol standards are independent of any specific software. But these protocols need the best programs to implement them, so competition will be very intense. Microsoft will invest more than other companies in this area, but the whole industry also recognizes its importance for e-commerce and is working on the development of standard protocols. So we will be able to use the “·NET” architecture to connect information, connect people, connect application software systems, and also various different devices. This architecture can complete all kinds of connections. It may seem a little unbelievable, but it truly is fundamental progress, such as the classification of information, the description of information, and standard information formats. For example, what is a meeting? What is a medical report? What is an address in a country, and so on. How is this information represented? Some of these standards will be done directly by Microsoft, but many specific industry standards will be completed by groups distributed around the world. Microsoft is only working together with partners to ensure that standards can be established. It is by this method that we move from connecting bytes through the Internet to truly connecting information through these standards. This leap is something computer science has long been looking forward to. Much past work in databases also had this goal, but only after the investment and learning of the last ten years have we finally been able to realize it. For example, someone needing a certain kind of service can find all providers of that service online, check their reputation, and carry out transactions with them, even though the buyer's programmers and the seller's programmers have never met or spoken. This kind of thing will soon become possible.

The demand for this kind of software expert is enormous. Our company's employees are developing software to simplify this process. Because our company is purely a software enterprise, our attention is mainly focused on giving the platform various functions. At present the platform's functions are already far greater than before. Functions that platforms used to have, such as queuing, transaction processing, event management, extra-overhead upward scaling, and fault tolerance, could formerly be realized only on proprietary hardware platforms. Now we are integrating all of these functions into the Windows platform and offering them at a lower price. This idea of a rich platform offered in volume at low cost is our company's contribution to social development. In doing this, we simplify the programmer's work. We do not buy many different software packages, and programmers do not need to face packages based on various different architectures. They only need to focus on developing their own business logic. It can be said that the goal we care about is being realized little by little. That is, connecting all kinds of different devices. In China now, mobile phones are very widespread, and of course there are other digital devices too, such as smart TVs, car computers, advanced computers, and information technologies like the Tablet-PC. These can connect the above-mentioned devices. Microsoft even wants to put chips into watches, so that through the watch screen you can display the functions of digital devices. You can see the computer screen in your pocket, the Tablet screen, the desktop screen, the wall-mounted screen. All of these are needed; they will not replace one another. They are simply connected to the same network, running the same applications, sharing the same data without requiring user intervention. We are amazed by the SPOT watch project. The microprocessor inside is worth only six dollars, yet its processing power is five times that of IBM's original personal computer. (A 28MHZ chip, 512KB of ROM, 384KB of writable memory.) Just these alone can drive the 120ⅹ92 pixel display on your wrist. All this is very different from the computing conditions when I was growing up.

Another key aspect of the platform lies in its openness. It can adopt different development methods: JAVA, C, VISUAL BASIC, COMBO, FORTRAN, and other new languages that may be invented. We are adding XML capabilities into all kinds of languages, and there will continue to be new exploration in this area. We can realize support for languages in an environment very rich in resources.

“·NET” means supporting all of these languages within a single framework. It uses the same development environment and the same debugger. Another very important advantage is that what we are talking about is not redeveloping existing applications. An important example is that the government can publish information to the public through government networks. The government cannot possibly redevelop those medical software, tax software, and land management software systems, but it can reintegrate them through XML. We have already shown governments powerful software that in just a few years can be efficiently developed through project initiation, and this new method can use XML to integrate previous applications and then connect them with current Web services. An important challenge the software industry faces in achieving this goal is effective reliability. This is what Microsoft has proposed as trustworthy computing: using reliable systems to transmit bills, arrange production schedules, and so on. All this requires special reliability, just like the reliability required in power systems. Up to now this is something all software has had to do but has not done, whether Mainframe UNIX, Windows, or any of these systems—they have all failed to do it. Most software industry R&D departments are studying trustworthy computing. How do we test code? How do we verify it? What is reasonable? How do we update code? How do we improve systems? How do we make hardware and software adapt to the system? How can failures be made as few as possible? And even if they occur, how can the time they last be made as short as possible? We have many tools to analyze software. One important tool is tracking software throughout its lifecycle so that it can be monitored while applications are running. All of these will be included in the tools and platform, so that these tracking tools can be used whenever needed.

Changes in programming tools will mainly move toward model-based programming. At the beginning, models are built graphically to create various data elements and various process fragments, then they are implemented through programming. The development model will be one of interconnection. It is unlike current systems in which two things are isolated from one another. You can even observe program execution and testing through various views of this model. Many of the functions we are now building are to enable programs running on different platforms to communicate well.

Therefore a rich messaging mechanism becomes especially important. This also supports parallel programming. You have all heard of grid computing, which requires connecting many computers together and concentrating their computing power for use. Web services provide the key infrastructure technology that makes solving the difficult problem of grid computing possible. So Web services have affected all these cutting-edge areas of computing. When Microsoft was founded, every company developed its own hardware and software, and of course this market was occupied by IBM and six other large companies. When we entered the IT industry, we naturally adopted a completely different computer architecture. The more specialized a computer architecture is, the more opportunities it has for participation worldwide. So when hardware and software can develop independently in parallel, the personal computer standard emerges. So now we can see that the personal computer industry has already become a global industry. In Asia we can see that Japan, South Korea, Taiwan, and mainland China have already become major members of this industry. Previously the center of development for the entire personal computer industry was in the United States, but now it is developing on a global scale. All of these technologies will become globalized: wireless networking technology, flat-panel display technology, chip development, system development, software applications, and so on. By removing the barriers of various architectures, we make the whole industry more competitive and more efficient. Even things like software consulting and software outsourcing services are done in large proportion in Asia.

China will play a major role in this. China has reasonable costs, high-quality schools, and talented people. For example, the research work carried out in universities. I know that in Chinese research institutions, standards work has only just begun, but convincing achievements have already been made. There are some people in the Chinese government who are very interested in China becoming a software giant. This requires long-term work, as well as creative work, transforming our abilities through changes in hardware, software, and system architectures, so that these companies can participate in valuable work in the software industry.

What I want to say now is that I am very optimistic about the future, very optimistic about the job opportunities this industry will create, and very optimistic about the development of hardware. You may think this optimism of mine is strange, because many IT companies, Microsoft excepted, are cutting expenses. Because the economies of the United States, Europe, and Japan are sluggish, IT development has slowed down, partly because what was used before were expensive hardware and software systems, such as UNIX systems, while now servers are based on the Windows operating system and Intel chips. Servers are powerful, because these are more advanced, and there is another shift as well: people insist on systems being more efficient, which also determines how we must launch a new Web services platform. Our investment in R&D increases by more than 20% every year, and the same growth rate applies here in Beijing as well. We have many devices that need to become intelligent. For example, trends in telephones are always changing: needing to display a map of your location on the phone, or take photos, or save your music and video; alarm clocks needing to be set with consideration of traffic conditions in the morning and their effect on your schedule. Watches can check the sports news or stock prices you are interested in, or display short messages, or carry out various settings within short messages.

We can integrate these intelligent devices with the help of software—sorry, I have to mention software and Web services again—so that they coordinate with one another and very efficiently achieve the personalized communication purposes you want. We have only just begun the software that needs to be done. We need natural human-machine interfaces, just like I am giving a speech to you now. We need to be able to obtain information, and we need to be able to store information in only one place; we call this unified storage. We still have many things to do. We need to improve the quality and efficiency of software development tools, to make communication easier and organization easier. Even people who have psychological barriers to these technologies, or those who are interested only in the Internet, can very conveniently use the conveniences brought by this software. These are all potential functions of software. For each such possibility, people are striving to realize it within a few years. We will build such a platform to provide convenience for software developers of various potential functions, so that these software products can realize their functions to the greatest extent possible. Here we have a very good virtuous cycle. Academic research done at universities contributes to commercial companies like us, and in turn we contribute to them. Profit-seeking commercial companies pay taxes, create job opportunities, and invest in education, and thus a very good virtuous cycle is formed. This kind of close cooperative relationship with academia is very important to Microsoft, and I also recommend that other companies adopt this approach. Microsoft is also making unremitting efforts to maintain and continuously improve cooperative relations with all kinds of scientific research. We have many such activities in China. We provide them with training and knowledge. This is the Great Wall Plan that we are implementing in cooperation with the Ministry of Education. We will first implement this plan in China's five best universities, and then expand it, so that China will no longer be limited in technical skills.

I am very happy to take part in today's “·NET” competition. All the works have been shown to me, showing the progress software development has made in recent years. All the contestants have left a deep impression on us, and I think their works are indeed very good. The most critical factor in this leading industry is software. Microsoft is committed to developing such a platform, so that the next generation of products can be widely promoted and applied, and to establishing a unified system architecture. Many software developers are also entering this industry. We do this because from conversations with many university professors and software developers, we have received this feedback. They have also told us the same thing: now is a very good time to do software development, a very good time to enter the software industry. Because this industry is changing the whole world, and the changes it brings to this world exceed those of any other industry, and especially so in China, because China is the fastest-developing country in the world. I am very proud that we can be part of it, and I am full of expectation for China's success!