Brief introduction:Why do we write? For many technical students, writing is much more difficult than coding, and it’s not a good experience to spend hours with a computer and find that you can’t write anything decent.

The author | | juven xu sources alibaba cloud native public number

The writing motivation

Why do we write? For many technical students, writing is much more difficult than coding, and it’s not a good experience to spend hours with a computer and find that you can’t write anything decent. Even for someone with some experience, I estimate that it would take more than six hours to write a four-thousand-word article of decent quality, which is not even the time consumed by the accumulation of materials.

So much work, why do it? I think there is great value in this, and there are two levels of value, which I’ll call surface value and deep value.

Surface value is extremely utilitarian. For example, a student wants to be promoted, and one indicator of promotion is personal influence, so writing articles can improve personal influence. For example, a Team Leader wants to recruit, how to let others know about you and your Team, writing articles is also a good method; Then there are articles written for superiors or stakeholders, similar to project reporting. The core concern of surface value is not the article itself, but the person behind the article. The author’s expectation of readers is often not that the readers agree with the content of the article, nor do they expect the readers to participate in the discussion of the content, but only that the readers quickly agree with the author.

Writing only about surface values is putting the cart before the horse. It’s like writing a popular science article on PM 2.5. If you start with the motive of selling your air purifier, the stench will soon show through.

In contrast to surface values, I think any article should start from deep values. This so-called deep value is the content of the article, the article point of view, the article needs to be as objective as possible, to the attitude of academic truth approximation. I write an article because I have my own thoughts on an issue, have a detailed understanding of many people’s thinking, have found some conflicts of opinions, and will not cater to others’ opinions for political correctness. To the best of my ability, take the ideas that I think are valuable, summarize them, and spread them to others in a clear and interesting way; I can feel the passion of writing, which comes from the pleasure of thinking, from the collision of ideas. During this writing process, my thinking has grown. Through a lot of logical thinking, my thinking has been improved. Secondly, the articles that are written are of great value to readers, because valuable knowledge is disseminated.

Another motivation for writing is to spread valuable technology. For example, Evangelist Josh Long of Pivatal has written a number of introductory articles on technology, as well as some great presentations. I asked him why he did so well that he was named one of the 20 most influential people in the Java field in the world over the years. His answer was this:

I think that people don’t trust technology, they trust people. So, while it is possible that the spring team could just publish good documentation and leave it for the world to find, It’s far more compelling when u feel u can ask questions of someone. And u can see that they’re having fun.I love Spring because it has made millions of lives easier. It makes me happy to think about its application, to see people happy with its possibilities.

An article is written because the author loves a technology, recognizes the value of technology from the heart, and believes in the potential of technology. Or because the author has something to sell. The difference in motivation between the two will soon be noticed by the more careful reader. Of course, these motivations are often mixed together, but if the main motivation for writing is on the surface, then the article is basically worthless.


Josh is one of the world’s most influential people in the Java field. He has influenced millions of Java programmers around the world through lectures, books, and blogs, and has helped promote excellent technologies such as Spring around the world. I wrote a book about Maven, which has sold tens of thousands of copies. In addition, the number of pirated PDFS has influenced more than 100,000 Java programmers in China, which has helped the promotion of Maven technology in China. I have been with Alibaba for more than 8 years, and have published over 60 articles on ATA, our internal technology community, with more than 50,000 cumulative views. I believe these articles have made some small positive changes to Alibaba’s technology. From these perspectives, writing good writing takes a lot of effort, but because it’s so easy to share and spread, it can affect a lot of people quickly, and the impact can last.

Of course, times are changing. In the past, due to the limitation of Internet technology, the transmission of words was more convenient, and the production and transmission costs of videos were relatively high. Therefore, the actual influence of speeches was much weaker than that of articles and books. Today’s Internet and video technology is so mature that making high-quality video may be easier to distribute.

Whether it’s an article or a video, influence should be used to do the right thing. Many of today’s gateways to influence have a new power over what sounds are heard, what images are seen, what words are read, and what attitudes are felt. The formation of power turns into rent-seeking power, and then it’s often not about the right thing to do.

I think the best examples of using influence to do the right thing are these two articles by Zhang USefully, the first is “Thoughts on THE formation of API Design Best Practices” and the second is “Vigilant the Complexity Dilemma: Thinking about Software Complexity.” The content of the article are engineers happy about software design of the in-depth analysis and thinking, but I think in this case, is more important than the content in these two articles has a very important information, that is, “in alibaba, even the researcher level, but also someone watching technology in more conscientious, To hell with those lower level managers who are already out of the loop, who talk about technology as a detail.” (Of course, this information is only my personal opinion and cannot represent Gu Pu)

Writing methods

Our basic education to cultivate the students’ writing, seemed there is a problem, in my impression, common those good selection on the market, tend to be pay attention to form and routines, tend to lack logic point of view, is to use an idiom to summarize “moaning whinge-bags”, what are the high lyric, view and the end result is a big deal. Everyone’s overall foundation has been weak, and the students engaged in technical work, often in the reading of the Language is still a weak, so the result can be imagined. Written things, to the logical logic is not rigorous, to the form of the form is messy, basic segmentation, punctuation, words are more problematic. The good news is that writing isn’t something you can only train in school. It can also be trained on the job, and there are clearly rules to follow.

1. The reading quantity

Zheng Ziying (name: Nanmen) emphasizes the importance of “reading more” in his article. He recorded reading more than 50 books a year on Douban, which is equivalent to at least one book a week on average, an astonishing number. I looked back at my own douban record and found that I read about half as many books as he did, about 25 books a year. The benefits of reading needless to say, I want to emphasize the point here is that constantly reading quality books can improve their own text appreciation, read more books, pick up a leafing catalog, which to find a few paragraphs to read, for the quality of the book, the in the mind probably have a spectrum. Most of the time when you’re writing, you’re actually looking at the structure and method of existing examples, and if you’re learning from a lot of good examples, you’re not going to get any worse.

Before I wrote Maven Practical, I read a lot of O ‘reilly, Pragmatic Bookshelf, and Manning’s computer books, and most of the books they publish are very high quality, they have a clear structure when it comes to technology, This is complemented by well-sized cases and theoretical analysis that is not boring. I wrote my own book, so to speak, based on the structure and approach of great books, and my book was generally well reviewed.

In addition to the benefits of writing methods, reading benefits more importantly from content. Again, for example, I was in “how to do technical TL” this article mentioned the TL of reading a lot over the years he’s doing this on management of books, including “win”, “driving force”, “the secret of the door”, “nonviolent communication” and so on, these books help I added my own thinking on the great degree. Writing is thinking, and thinking does not come out of thin air. Thinking needs raw materials, and the common raw materials are our actual work experience. However, one’s experience is very limited after all, and reading about the experience and thinking of others, in a humble manner to provide more raw materials for one’s own thinking, is obviously wise.

“How to do A good Technology TL” received an interesting comment that read:

Personally, I think all books related to management are chicken soup, so I don’t need to read them. If you need to rely on “how to do management” books to do management, then this person is not suitable for management.

The implication is that some knowledge, such as that of management, cannot be passed on and can only be learnt by nature. For this comment, I can only say that the ignorant are fearless, we must not write articles to hold this state of mind, even if their ideas are how unique, how original, also must not build fences, sit in a well.

2. The material

I recently has been a year to do the related work of cloud native period I’ve been thinking, what is the cloud native architecture, the current interpretation of the word there are so many in the industry, but all these explanations can’t give to my satisfaction, they are all too lay particular stress on technique and the Angle of cloud vendors, and the lack of application architecture perspective. With this in mind, I wanted to write something about the concept of cloud native architecture based on the work I’ve done over the past year, and in the future, to add to the concept and help people use the technology better. The article hasn’t been written yet, but the motivation was there.

With this motivation, and some of my own personal interests, I’ve been accumulating material over the past year that’s been very interesting. For example, I was directly and deeply involved in the projects of International Middle Platform and Koala, whose architecture was based on cloud-native technology and many Products of Ali Cloud, and they made a great upgrade. I also learned in detail how Cainiao built a platform on the cloud to serve his partner companies; Learn how dodo and IoT, which sell services on the public cloud themselves, are building on the cloud. I communicated with relevant students through IM, phone or face to face, and they were all very friendly and told me everything. Then I sorted out relevant materials and analyzed the same pattern from them. The human brain loves the stimulation of pattern recognition, and I enjoy doing it. In addition to practical cases, the opinions of industry veterans are also the source of materials I usually collect. For example, Lin Hao (name: Bi Xuan) recently wrote an article titled “Further visualization of Cloud Native”. Usually, their articles do not follow what others say, but if you read them carefully, you will find some original opinions.

In addition to the above materials that are closely related to the target theme, I also found that interdisciplinary reading often brings me unexpected harvest. Take cloud native as an example. Although I work in this field, I have been reading some economics books on and off in the last year or two, including Principles of Economics by Mankiw. (I guess a lot of people have bought this book, but few non-professionals have really read it.) When learning the thinking mode of economics, I want to explain the cloud native from the perspective of economics. The original self-built technical infrastructure is gradually evolving cloud-based infrastructure in the cloud era. For businesses, how should the choice behind it be analyzed from the perspective of economics? What are the costs and benefits of self-building? What is the cost and utility of the purchase? After breaking down a complex business technology architecture into layers, it may be possible to analyze each layer and make an economically optimal choice at each layer.

Mind mapping

I first learned about mind mapping from reading Pragmatic Thinking and Learning, where the central use of mind mapping is to write everything that’s relevant to a topic, both important and minor, in a single map, lines that grow and stretch, not to form a clear structure, It’s to show the brain where it can expand. Therefore, the way to draw mind maps should be as open and free as possible, and the goal is to show all the contents related to the theme, rather than to establish a clear logical structure. So when I see a lot of drawn mind maps that are actually catalogs, I think they’re basically wrong.

Writing articles (and giving speeches are similar) is a process of opening up and then convergence. In the early stage, continuous accumulation of materials, use of mind mapping to expand thinking and establish the association of materials are the stages of opening up. This is the stage where logical thinking can be put back in the background, where the eyes are open, where there is no evaluation (building structure, prioritizing is actually evaluation), and where there is an empty cup mentality to listen and observe the real world and the thoughts of others. Before I give a speech or write an article, I always find a quiet place, prepare coffee, open MindNode, or just take out a piece of A4 paper and a pen, and give myself half an hour to an hour to draw a mind map related to the topic. A quiet environment, a lack of time pressure, and an energetic brain are all combined to take the focus off the task at hand and allow the unconscious awareness in your brain to surface and drop onto the page.

Structure of 4.

A large number of materials piled together to create an article is not natural, the work must have a clear result. In architecture, we are familiar with the column structure of ancient Greek architecture, and gothic architecture with pointed arch, rib arch, flying buttress and other elements as the core of the structure; In the program, common MVC, layer, microkernel and other patterns are also clear structure. The structure of the article can help the author express his ideas clearly and guide the reader in a clear path. When I have a mind map, I usually extract the most appropriate structure from the complex material, and then organize the material based on that structure.

There are some common structures for writing technical articles. Here are some paradigms:

  • Problem solving structure. This paradigm usually focuses on solving a specific problem, and the common logic is: background introduction -> raise the problem -> discuss the way to solve the problem -> solve the problem -> value summary. This structure is probably the one most familiar to engineers, because everyone is good at solving specific problems.
  • Knowledge introduction structure. This paradigm is usually used for the introduction of new technologies, the common logic is: industry background -> technology proposed -> simple Demo -> Core Conecpts -> concept depth and Demo (1-N times) -> prospect analysis. Many of the technology introductions you see often use this structure. The advantage of this structure is that it allows you to introduce new technologies and concepts from the ground up.
  • View output structure. This paradigm is more challenging than the previous two. The common logic is: overall view -> sub-view 1 -> sub-view 1 exposition -> sub-view 2 -> sub-view 2 exposition… Summary – >. This kind of structure is very powerful when written, but very challenging to write, because the point of view requires rich material and rigorous logical reasoning, the slightest mistake can be nonsense.

Of course, we don’t have to stick to these paradigms when we write, and we can think about our own paradigms, but there is always a logical structure behind any paradigm. There is a book called “Pyramid Principle”, which I hear many people praise highly (especially in performance season). I read the introduction and evaluation, which is about how to make people understand you in an efficient way. I haven’t read this book, but it should be about the structure and logic of writing.

5. View

Not all articles have opinions. For example, if you write an article summarizing how you solved a performance problem, you don’t necessarily have opinions. Introducing a technology like Rust doesn’t have to be an expression of opinion. But articles that express their own ideas are often more attractive, such as the importance of understanding queuing theory when solving performance problems. Introducing Rust by asserting its future dominance in performance-sensitive scenarios makes it easier to make the technology compelling. Of course, an idea needs to be demonstrated, and its robustness is proportional to the investment of your argument. Logical deduction, data support, and case analysis are all very good means of argument.

We also see some articles full of opinions, but they are mostly quotes, one is Jobs, one is Zhang Xiaolong, one is Ma Yun and so on. Of course, it’s more common to quote company executives, what so-and-so said when, and so on, and then use that to back up your own material. I think occasionally quoted innocuous, always quoted can only show that they do not have an opinion, or the point of view is weak, need strong support for their own.

More courageous, can reflect their rich material, thinking depth of view, but those who dare to say the emperor’s new clothes. Technical articles should have the spirit of science, science is based on the continuous development of the past say no, technical articles should also dare to express the opinion of say no, without fear of offending people, we should understand that the correct technology/architecture/scheme, should be able to withstand questioning, because if the technology is wrong, Over time, reality makes mistakes exponentially more expensive, even if no one dares to question the circumstances. Therefore, articles that dare to say no to the current situation and put forward their own opposing views deserve more praise than those that are full of correct nonsense.

Stories of 6.

Strictly speaking, a full story that embellishes its own content does nothing to make a logical argument. However, to make your essay/presentation compelling, a story element is essential. As humans evolved today, the brain’s response to logic is very slow, and it takes training to understand logic, but the response to stories, as young as three or four years old, is there, and the spirits of early humans, such as Greek mythology and the Bible, are full of wonderful stories. To this day, no matter on the Intranet of the company or on weibo, people’s enthusiasm for eating melons can be ignited by logical reasoning. Stories are easy to elicit empathy, to make people associate themselves with others, and then the cerebral cortex is easy to get high, neural networks are activated, all kinds of hormones are released…

This is the reality of human physiology, so we should respect (exploit) this reality when we write. I want to talk about the poor performance of the program that caused my colleague to wake up with a call at 3:20 a.m. The supermarket cash register was smashed when something went wrong, the introduction of a new programming language had to show a piece of code, and by the way, Java didn’t work; Introduce Mesh by saying that you used to do this for middleware, you don’t need to do this anymore, etc.

There’s a true story I like to tell when I introduce myself to Maven Combat:

A few years before the publication of my book, due to my vanity, I particularly liked to check the reviews on various booksellers’ websites, such as Amazon, Douban, jingdong, etc. When I read the reviews one by one, I was happy to see the reviews with five stars, and angry to see the reviews with one or two stars. Of course, most of the reviews were very positive. Until one day, I got a 1-star review on I was sad and opened the review, but after reading it, I was happy. The reader gave me a 1-star review because he said, “Let your boss make milk tea, bad review!” It turned out that liu Qiangdong milk tea broke the news of love, hurt the reader’s heart, and then I was innocent.

This story actually has nothing to do with the technology I cover in the book, but I love telling it because it makes the audience laugh and remember that I wrote a book about Maven.

What does bad writing look like

With so much talk about how to write good articles, I also want to talk about what bad articles look like. A bad article is one that has little or no positive value to the reader, or even worse, not only no positive value but negative value. Here is my summary:

  • Personal notes into an article. I solved a technical problem, made some records, and then wrote an article and sent it. This kind of article can only be called material at best, because it is not summarized and refined, and it is written completely from a person’s perspective. When readers read it, they can not only feel the system, but also have little value.
  • PPT paste into an article. The author made a speech somewhere, because he wanted to spread it to more people, so he arranged the article in the way of stickers. Kindly, he added some explanatory words in the middle of the picture, and then it was written. I assume that the quality of the speech itself is good, with logic, ideas and cases, but even so, the reading experience of such so-called articles is very poor. In a speech, POWERPOINT is an aid to “speaking”. If there is only POWERPOINT, the real core of “speaking” is lost. Therefore, it is more responsible for the speaker to write down what he or she is saying in a clear and rich way, rather than just a few dry PPT pages.
  • Write articles in the manner of propagandising their military exploits. It’s not uncommon to see articles like this in organizations, with titles that include the usual words “summary,” “year,” “system,” “reflection,” and “outlook.” Often, these articles have neither system nor reflection, and their main purpose is to claim credit. The basic routine is to write that you have done a lot of things in a year, and the results are very good. Add a few similarlooking frame “structure” pictures, and more frankly, a group photo. This type of post usually gets a lot of likes, but little discussion. What is it? The author came to take credit, you say he did not do well to offend people? From the point of view of spreading knowledge and promoting thinking, this kind of article has almost zero value.
  • All kinds of stickers into articles. Compared with PPT map, there are a variety of other pictures pasted into the article, mind map directly pasted, design directly pasted, flow chart directly pasted, monitoring map directly pasted, a look down is not to see a few lines of words. Good picture, is indeed a picture worth a thousand words, but it should only be the finishing touch, if the so-called article is all pictures, you can not see the focus, also can not see the system. Words are very powerful. They can pull readers into the author’s thinking system and guide readers’ thinking with logic and rhetoric. Or induction, or derivation, let the originally hidden knowledge show; Or eloquence, or humor, make the author’s point shine. Writing should give full play to the power of words. It is not only pictures that provide structure, nor is it only pictures that embody thinking. On the contrary, pictures are often used by many people to hide their lack of thinking because of their imprecision.
  • Correct nonsense into articles. Bureaucratic jargon, often abstract to a very high height, from the company strategy, a detailed analysis, is full of the recently criticized words such as “top-level design”, “bottom logic”, “enabling”, “grasp” and so on; If you can’t find clear logic, just do 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, anyway, the major version number plus 1 is better than the previous one, as to why the minor version number has never been used, I don’t know, anyway, N 0 is right, direct N is not, direct N 0.0 is not ok. After reading an article like this, you can’t tell what went wrong. The only thing you know is that it’s worthless and you’ve wasted a few more minutes of your life.

Copyright Notice:The content of this article is voluntarily contributed by real-name registered users of Ali Cloud, and the copyright belongs to the original author. Ali Cloud developer community does not own the copyright, and does not bear the corresponding legal responsibility. For specific rules, please refer to the “AliYun Developer Community User Service Agreement” and “AliYun Developer Community Intellectual Property Protection Guidelines”. If you find any suspected plagiarized content in this community, please fill in the infringement complaint form to report. Once verified, the community will immediately delete the suspected infringing content.