My views on how to write a scientific paper well (2) Min's research motivation published yesterday is the first in a series of blog posts on how to write a scientific paper well, and I want to share with you my experience in reading and writing papers. What I'm going to talk about today are the types of scientific papers. Count them as the second! This is by no means a classification, because the papers I have read and written are limited, so I can't generalize. (1) new concepts It is not easy to introduce new concepts into the paper. For example, claude elwood shannon (1916–2001) was an outstanding scientist, who created many things. He and E. F. Moore published the article "Reliable Circuits Using Unreliable Relays" in J. Franklin Institute in 1956. The innovation of this paper is not only to define the reliable circuit, but also to prove that any reliable circuit can be composed of unreliable components. This is very new. Because it is generally believed that the more components, the less reliable the system. However, he doesn't think so. This proof is certainly difficult. Some of our articles, although keen on introducing new concepts, have only general and imprecise narratives, and after defining them, they are finished, with neither theorems nor conclusions. Such a concept is suspected of playing with the concept. Because you are inconclusive about this concept, what's the point? There is no point in defining without a theorem. (2) New methods There are many articles that put forward new methods. For example, there are many articles about routing algorithms under wireless conditions. When proposing a new method, be sure to explain: Where is your new method? What are the advantages of this new method? Otherwise, this new method is meaningless. You may need to compare with existing methods. When comparing, it should be a fair comparison You can't show yourself by making other people's methods very bad and deliberately lowering the performance of other people's methods. (3) Improve and popularize this kind of articles the most. In other words, the basic idea belongs to others. However, if you find something wrong with other people's articles, you need to improve them, or you can promote them (including concept promotion, method promotion, application promotion, etc. Some people satirize this as "using literature". I don't think so. What can improve and promote others is also good. This is often the driving force to form the mainstream of scientific research. Scientific research should be good at following the mainstream and braving the wind and waves in rivers. (4) System implementation In the field of engineering, you can also write articles on system implementation. However, your system implementation must have considerable workload, innovation and experimental indicators. A bought system, linked together, doing an experiment, is not good for writing articles. You made up some programs yourself, just transferred them. It is not easy to write an article without testing, experimental results and comparison with similar systems. (5) Overview Some of our graduate students want to write a summary after reading several articles. Actually, it's not a summary, it's a reading note. It is an article that guides the research in a certain field after having a unique opinion on it. Therefore, a good magazine will invite top scientists in a certain field to write a summary, and a good conference will invite outstanding experts to give special reports. (6) Popular science articles Popular science articles should also be counted as part of scientific papers. Because popular science articles preach science and technology to the public from a higher angle. It's not easy. Some popular science articles are purely mystifying and even promoting pseudoscience. Everything else is parrot-learned, nothing new. When you want to write a scientific paper, I suggest you think about it. What kind of article do I belong to?