Discussion is an important part of scientific and technological academic papers. The purpose of the discussion is to explain the phenomenon, expound the viewpoint, explain the significance of your investigation/research results, and make suggestions for the follow-up research. Its main function is to answer the questions raised in the introduction and explain how the research results support your answers and how these answers are consistent with the existing relevant knowledge on the subject. Discussion is usually regarded as the "heart" of a paper, which can best reflect the author's mastery of literature and his understanding and understanding of an academic issue. Discussion is a weak link in many papers by China authors, and it is also a gap with international authors. The discussion part is difficult to write, so the author should attach great importance to it, devote due time and energy, and write the discussion carefully. The quality of discussion writing often determines the depth of an article and is also a sign of its academic level. The author should focus on the theme, discuss it pertinently, avoid repeating data results, avoid repeating summaries and introductions, and avoid lengthy text piling up. The discussion should be based on the research results, and the author should not be afraid to draw conclusions, nor should he rush to draw conclusions. The discussion part should emphasize the new important achievements and conclusions of the research, and explain the value and limitations of the research; The discussion should be combined with the research purpose to avoid the conclusion that the research results are not supported; If you don't do economic analysis, you shouldn't draw a general conclusion about cost and benefit; Avoid emphasizing and implying the importance of unfinished work. If you are sure, you can put forward new assumptions and suggestions for further study. The discussion of many authors' papers is not systematic and in-depth, and one of the important reasons is that the relevant literature is not quoted correctly, so the research results are not discussed in a broader context. Many authors don't quote or systematically quote relevant documents during the discussion, some think it is unnecessary, some don't find them, and some deliberately avoid quoting them to highlight the "novelty" and "value" of their research. Although some authors quoted relevant literature, they did not combine their own research and discussion, that is, they did not discuss the previous research results with their own research results. This kind of discussion leads to the Balkanization of the paper, and readers can't understand your research results systematically and deeply. An honest or mature author should have an objective evaluation of the research results. It not only emphasizes the success and advantages of the research, but also points out the limitations and shortcomings of the research. In order to make your information clear, the discussion section should be as brief as possible on the premise of comprehensively expounding, supporting, explaining and demonstrating your answers and discussing other important and directly related issues. It must be pointed out that the purpose of the discussion is to comment on the results, not to repeat them. You shouldn't dwell too much on unimportant side issues, because it will obscure your important information. The phrase "no paper is perfect" is endless. The key is to let readers know what is certain and what is just hypothetical. The structure of the discussion section is very important. Before you start writing, you should draft an outline to organize your thoughts in logical form. Chain diagrams, problem trees, numbers or other organizational structures can be adopted. Drawing lessons from predecessors' experience and combining my ten years' editing experience, this paper summarizes ten steps of writing "exposition" in order to help sort out ideas. 1 Organize "discussion" with the logical structure of "from special to general": from your discovery to literature, to theory, to practice. When writing "discussion", you can discuss anything, but it must be concise, short and clear. Start by reiterating your hypothesis and answering the questions in the introduction. When asking questions in the introduction, use the same terms, the same verb tense (present tense) and the same point of view. 3 Support the answer with the test results. Explain how your results are related to expectations and existing literature, and clearly explain why your research results are acceptable, and how they are consistent or consistent with the knowledge of published literature on the subject. Discuss all the results related to the questions raised, briefly summarize the main significance of the research results, regardless of the statistical significance of the results. The models, principles and relationships revealed by each major discovery/achievement should be described in the correct logical order. The logical order in which these messages are expressed is very important. State the answer first; Secondly, provide relevant results; Then quote other people's research results. If necessary, point out illustrations or tables to readers to deepen their understanding of the story. 6 demonstrate your answer. If necessary, explain why your answer is satisfactory and others are not. Only by providing both positive and negative arguments can you make your explanation more convincing. 7 Contradictory interpretation of discussion and evaluation results. This is a sign of a good discussion. Many domestic authors usually avoid discussing content that is inconsistent with the conclusion. ) discuss any unexpected results. When discussing an unexpected discovery, start with discovery and then describe it. Point out the potential limitations and shortcomings, and comment on the importance of these factors to the interpretation of your results and how they affect the correctness of the research results. When pointing out these limitations and shortcomings, avoid using the tone of apology. Provide suggestions for further research (two at most). Don't make suggestions that should have been mentioned in this study, which will show that you haven't fully tested and explained the data. 10 to make clear how important the results and conclusions of this study are and how they affect our knowledge or our understanding of the studied problems, the discussion of scientific papers needs to construct the reader's IMRaD structure of the paper (introduction, methods, results and discussion; Introduction, methods, results and discussion) are quite familiar with, consciously or unconsciously know the role of each part. Similarly, readers are familiar with structured abstracts, which contain more information than unstructured abstracts. Some publications require special papers to use special structures, for example, papers reporting randomized trials use CONSORT structures. Now we propose that the discussion part of scientific papers should also be structured, because it is often the weakest part of the paper, and it will turn into chatter after careful explanation. Old papers often lack new data, but the discussion is eloquent. The role of discussion seems to be to convince readers that the author's interpretation and speculation of the data are correct. This is not a fair evidence test. Times have changed, and people pay more attention to methods and results, because methods are more complicated and scientific. But we still find that the discussion of many papers seems to be "selling" ourselves. Richard Horton, editor-in-chief of The Lancet, and others have written about how the author plays with rhetoric in the discussion of papers. Some authors' discussion articles are very miscellaneous and have no subtitles; Heavy general, light special; The narrative is bloated, and the results are selective and repetitive, which exaggerates the importance and universality of the discovery and leads to deviation. Draw inappropriate conclusions regardless of the collected evidence. Suggest the structure used in the discussion part of scientific papers: 1. State the main findings. Advantages and disadvantages of this research. Advantages and disadvantages compared with other studies; In particular, we should discuss the differences in the results. 4. The significance of this research. 5. The unanswered questions and future research directions. At the beginning of the discussion, we should reinterpret the main findings, and it is ideal to express them in one sentence. Then fully explain the advantages and disadvantages of this study, both of which can not be neglected. In fact, editors and readers are most concerned about the shortcomings of research, which is inevitable in all research. Once editors and readers find the shortcomings of the research, but the author does not discuss them, their trust in the article will be shaken, and they will have doubts: Are there any other weaknesses that they and the author have not found? Secondly, the research is related to the previous work, not to show off how good your work is, but to compare its advantages and disadvantages. Compare with other studies and avoid covering up your own shortcomings. It is important to discuss why the conclusion is different from others, and the author can speculate freely; However, if you don't know why your research results are different from others', it is not convenient to make such a guess, and you should not assert that your research results are right and others' are wrong. Then we will discuss what our research "explains", how to explain our research findings, and what is the significance for decision makers? At this moment, the author's situation is dangerous, and most editors and readers can understand the author's caution and don't exceed the limit of experience. It is up to readers to judge the significance of this research: they will do it. The author can even point out that the research results can't prove anything, lest readers draw excessive and untrue conclusions. Finally, we should point out which questions have not been answered and the work to be done. Obviously, neither editors nor readers like exaggeration. In fact, the author often writes this part of the paper in a mess. Although the author cannot be prevented from writing an article full of speculation, the evidence must not be destroyed by speculation. The discussion section may sometimes need other subheadings, but we think the structure proposed now is suitable for most research papers. Although it is difficult and even limited to unify the structure, we believe that this structure will reduce the total text length, prevent inappropriate speculation and repetition, reduce the deviation of the report and improve the overall quality of the report. This idea can stand the test completely.