) Programming experts

On the mit bbs, the Microsoft Phone interview question is "Who do you think is the best coder, and why ?". I think it's quite interesting. Ranking in no particular order.

Bill Joy, former Chief Scientist of sun, presided over the development of the earliest BSD version at Berkeley. He is also the author of VI and CSH. Of course, CSH programming considered harmful is another topic. It is said that he wants to see if he can write an operating system. In three days, he wrote his UNIX, the predecessor of BSD. Of course it is a legend, but it shows his skill. Another legend is that in early 1980, DARPA asked BBN to add the TCP/IP code developed by BBN in Berkley UNIX. At that time, the graduate student B was angry and refused to add the BBN TCP/IP to the BSD because he thought the BBN TCP/IP was not well written. As a result, Uncle B took the shot and quickly wrote the high-performance Berkeley TCP/IP. At that time, BBN and DARPA signed a huge contract to develop TCP/IP stack. who knows their code is not as good as that of a graduate student. So they held a meeting. I saw B's T-shirt appear in the conference room (t-shirt was not as casual as it was now ). Q: How did you write this? B's answer: simple. You can read the protocol and then program it. Even though Mr. B decided to develop in the industry after graduation, sun, who had only one office at the time, then he designed the world for the company... A cool man like this kind of hardware and software can't do anything without admiration. According to a colleague of Bill Joy, Uncle B always casually reads a bunch of magazines during a meeting. However, it is often critical that Uncle B speak directly and put forward beautiful ideas to let his colleagues crash completely. By the way, he is also one of the main authors of Java spec And Jini.

John Carmack, founder of id software and lead programmer. He chatted with a graphics engineer last month. He didn't even know John Carmack, and even fainted. However, it may be a bit different from the actual research. Everyone who likes the first-person shooting game knows J's brother. At the beginning of 1990s, when we were able to make a small animation on the PC, we were amazed at it. J's brother launched the amazing castle Wolfstein, And Then Yan Li, doom, doomii, quake... every time we push the 3-D technology to the pole
Limits. On his resume, J's brother said that his specialty is "Exhaust 3-D technology". I am a good man. A person like J's brother is very happy, because the major graphics card manufacturers will "Contribute" to him as soon as they have a new product. Otherwise, if his game does not support this kind of card, what kind of card will basically die. At the beginning, Ms direct3d had to listen to his opinions and modified a lot of APIs. Of course, J's brother has been programming for more than 14 hours every day for the first ten years of his marriage, and even ordinary people have nothing to do with it. By the way, J's brother, senior high school student (?), It can be said that it is self-taught. However, it is a big mistake to use this example to defend yourself from poor learning. Leonardo da Vinci is still self-developed (people are selfish and cannot go to school ). There are differences between common people and geniuses. By the way, it is quite wrong to call "da Fen Qi", because Vinci is a place name, and Da Vinci is the meaning of people who come from Vinci. In other words, Leonardo da Vinci means "Leonardo from Vinci. I don't know what Da Vinci is like. Well, it's too far. Stop.

David Cutler, chief designer of VMS and Windows NT, is the best kernel developer in Silicon Valley before Microsoft. He and his team wrote a bootable kernel with basic functions in Microsoft within a week and said, "who can't write an OS in a week? ", Also said. By the way, when Grandpa D arrived at nt3.5, he managed 1500 developers. He also designed and Programmed himself, without changing the coder color.

Grandpa D is naturally angry, and he prefers to slam the table with both hands in a strong manner when arguing with others. Daily Conversation F-word is left blank. When interviewing the secretary, he must ask: "What do you think of the word 'fuck '? ", Let countless beauty brake Yu. One day, an equally popular girl blurted out the question: "That's my favorite word ". So she was admitted to work for Grandpa D and released nt3.5.

Don knuth. Grandpa Gao doesn't need to say much. I don't know how to learn programming. It's like I don't know about physics, Newton, I don't know about mathematics, I don't know about music, I don't know how to learn Delphi, I don't know about Anders hejlsberg, or Linux does not know Linus Torvalds. To make the article complete, let's say a few more words. Grandpa Gao started to write a variety of strange compilers for the company. He charged 1000 or 2000 US dollars for selling it to others. The companies who took the code and processed and sold it for tens of thousands or even 100,000. However, I have never seen Grandpa Gao feel uncomfortable, and the scholar said. Think about it. In the early 1960s S, Grandpa Gao wrote a lot in the compiler, and then produced an Attribute Grammar and LR (k), which greatly benefited future generations. Grandpa Gao is always the first in the Caltech Programming Competition (with Alan Kay winning experts), and the Code is freeze from Tex to 86 years, it also comes with a reward of 2 ^ n cents, and so on.

By the way, Master Gao is an undisputed master of writing. The preface he wrote to concrete mathematics is a good example of the preface. His technical article is also a breeze, meticulous, well-explained, and without learning, without losing the air. I remember reading concrete mathemathics a few years ago and laughing from time to time. This made my mom very depressed and thought I was helpless when nerdy got home. In fact, the child is not a fish, but the joy of the fish, I do not know that it is the credit of Grandpa Gao. Speaking of writing experts, Stephen A. Cook cannot be mentioned. His article was highly recommended by our writing teacher, known as a sample of elegance. Grandpa Ku has a silver hair and a long figure. He always smiled modestly and said that he exactly matches his fairy tale.

Grandpa Gao is still a pioneer in the open source movement. Although he hasn't moved around like Richard Stallman, he has contributed many works and can see them online, such as the famous mathematical writing, mmixware, And the Tex Book, needless to say, it is enough to make him feel better at Tex.

Ken Thompson, author of C language predecessor B and inventor of Unix (the other is Dennis M. riche leader, honored as DMR), one of the authors of Belle (a powerful chess Program), the main author of the Operating System Plan 9 (the other is Rob Pike, recently, it was dug by Google ). Grandpa Ken is also a man in the history of computer science. In the prehistoric era of computer science in 1969, ordinary people thought that only the mainframe could run general-purpose operating systems, and minicomputers could only run on a mountain basis. As for operating systems written in advanced languages, it is even more a joke. Grandpa Ken is not a thing in the pool, so he and DMR angry, in 1969 to 1970 with assembly on the PDP-7 to write the first unix version. They do not know that a vigorous UNIX legend kicked off. In 1971, Grandpa Ken re-wrote UNIX with C, so c became a hero's dream and glory in the next 20 years.

Grandpa Ken also has a piece of story: Installed UNIX PDP-11 was first installed in Bell lab for your daily use. Soon, you will find that Grandpa Ken can always access their accounts and obtain the highest permissions. The scientists in Bell lab are all at a higher level, and of course they are very depressed. As a result, the Experts burst out and analyzed the Unix code, found the backdoor, modified the code, and re-compiled the entire UNIX. When everyone thought that "the world is clean", they found that Grandpa Ken still had to easily get their account permissions, but they had to continue to be depressed after thinking about it. Who knows this is depressing, it will be 14 years, until Grandpa Ken said the reason. Originally, there was indeed a backdoor in the code, but the backdoor was not in the Unix code, but in the C compiler that compiled the Unix code. Every time the C compiler compiles UNIX code, the backdoor code is automatically generated. The entire Bell lab team uses the C compiler of Grandpa Ken.

(6) Rob Pike, former member of technical staff of at&t Bell lab, now Google is studying the operating system. He was the pioneer in UNIX and was the first UTF-8 designer to develop UNIX with Ken Thompson and Dennis M. ritche at Bell Labs. He also made a small face on the evening program of David Letterman, the famous American mouth, to help a fat man brag. What I admire is that he is also the silver medal winner of archery in the 1980 olympic games. He is still a very powerful amateur astronomy. The design of the Lamma ray telescope was almost used by NASA on a space shuttle. He is also one of two classics, the Unix programming environment and the practice of programming. If you want to improve programming, read these two books. They all have Chinese versions. He also wrote the first Bitmap-based Window System in UNIX and was the author of the famous blit terminal. Of course, he is also the main author of the innovative operating system plan9. Unfortunately, plan9 does not attract much attention. Luobo was furious and wrote out the arrogant systems software research is irrelevant. He scolded the current system development for its own shortcomings. Although this article is quite excited, it does show the helplessness of system development: the development cycle is getting longer and more expensive, users are unified to a few systems, and more activities are being measured and repaired, with fewer and fewer real innovations.

Just as Uncle Luo was very depressed, Google was eager to seek help. If there is another public company that is making every effort to push system development to the extreme, Google is happy. Just look at the results of Google. GFS, a distributed file system with strong fault tolerance and load balancing capabilities (now it can build a giant distribution system with 100,000 low-cost PCs, and the system for efficient and inexpensive management is not Doha ), large-scale Machine Learning System (spelling check, AD matching, pinyin search... Not to mention the various Google services that process massive parallel computing. Rob says at System Software Research is irrelevant that no one cares about the cutting-edge results of system research. I cannot think of him as wrong. I should be concerned about Google. Google has made a large number of achievements, always trying to learn from the latest results of system research. Presumably Rob Pike is happy at Google. May he make a better system.