[英]《奇思妙想》作者Dennis Shasha：小品電腦科學（圖靈訪談）

圖靈訪談之三十四：專訪《奇思妙想》作者Dennis Shasha

Dennis Shasha，《奇思妙想：15位計算機天才及其重大發現》的作者，紐約大學柯朗數學研究所電腦科學教授，先後獲得耶魯大學理學學士、雪城大學理學碩士和哈佛大學哲學博士學位。他的研究領域包括生物計算（包括實驗性設計）、模式識別以及樹和圖示的查詢、時間序列的模式探索、加密檔案系統等。

此次圖靈社群為了不耽誤電腦科學的發展，只問了薩沙教授五個問題。

iTuring: Some people accused that science study is too far-fetched from industry development, and their value is hard to be seen. So how the value of science study can be evaluated by non-scientists? And as a computer scientist yourself, do you have any experience and story dealing with this kind of problem?

Dennis : There is much to be gained from practice in industry. Workers and engineers understand what the real problems are, instead of having to imagine what the problems are. However, think how much is gained from pure science: Galileo was worried about the movement of planets; Newton found a simpler formulation than Galileo and in the process found the laws of force and motion; practicing mechanical and civil engineers use those every day. Within computer science, the discoveries of Rabin and Scott were done out of mathematical interest, but "regular expression searching" is the basis of search engines. Number theorists found curious properties of primes that have become the basis for modern cryptography and secure e-commerce.

iTuring: The nature of computer science seems to be rather stable since it was ever invented, and some people say that the recurrence and reinvention of software development is a waste of time of the potentially most genius people on earth, what's your opinion about it? In return, what do you think computer science is trying to achieve, solving people's problem or helping them get closer to the absolute truth?

Dennis : I agree that there is a lot of wasted effort when for example a perfectly good piece of software is replaced by another one in the "latest" language. However, to say that all that has happened is reinvention would be to ignore all the great software-based devices we use every day from cellphones to websites to smarter airplanes. Computer science is a flowering subject trying to achieve so much: helping to analyze data to help researchers understand the natural world better; robotics to help old people; learning patterns to do better science, finance, and marketing; cryptography to keep secrets; graphics to display information; numerical analysis to design better airplanes.

iTuring: Some subjects, such as physics, have developed into a certain stage, that few breakthroughs are made. Will computer science be like this someday? And in your opinion, what is the most characteristic feature about computer science comparing to other science subjects?

Dennis : Physics is a science in that it attempts to understand nature. Nature reveals its secrets only through great effort on our part these days (spacecraft and accelerators). Computer science is a science too in its study of computational complexity, but it is also an engineering discipline. As such it is much more dynamic, because while nature is fixed, the human imagination can open to billions of possibilities.

iTuring: In the coming decade, what do you think is the most promising subject in computer science? And do you have any advice for those who are interested in having a career in computer science?

Dennis : Scientifically, I think the most promising subject is machine learning, because almost all disciplines are generating data at increasing paces and being able to analyze that data and learn from it will lead to great new science. Monetarily, jobs that involve finding the best ways to move large sums of money around always seem to pay the best on the average.

To enjoy computer science, you must love to learn new skills. The reason is simple: if you stick with old skills, you will become obsolete. So start by learning the basics: algorithms, programming, and management of data. Then find a job with people whom you respect and who respect you. Then keep learning.

iTuring: Lots of people wonder what computer science may be able to do one day. In future, is it possible for computer to imitate human intelligence? And is it possible for computer to improve human intelligence?

Dennis : Yes, computers already imitate or even exceed human intelligence in many areas: computer software imitates stock traders for example; computers play better chess than most people (though not GO, yet); computers do a better job managing warehouses than people; and robots are taking over many factory jobs. Computers certainly improve human intelligence by taking care of part of the work. If your computer trades stocks for you, then you can think of how to manage risk for example.

---

更多精彩，加入圖靈訪談微信！