Final Words - Entering the Time Rift to Continue Learning #

Hello, I am Wu Lei.

Time flies by so quickly. Unknowingly, it is time for me to say goodbye to you. First of all, I want to thank you for your support and companionship. Honestly, I am a bit dazed right now, and I can’t believe that this column has come to an end.

Since July, when I started preparing for the column “Getting Started with Spark from Scratch,” rushing, revising, recording audio, and replying to comments have become part of my daily routine. Looking back on the past four months, the first phrase that comes to mind is “surviving in the cracks.”

Why do I say it’s surviving in the cracks? As a middle-aged professional with a family, the dual pressure from work and family makes the already limited time and energy even more stretched. The importance of work goes without saying; it is the foundation of our personal development and naturally requires our full dedication. As Chinese people say, “family affairs come first,” so even the smallest matters at home have a higher priority than anything else.

Undoubtedly, work and family take up most of the time in a day. If we compare time to a wall, then in a 24-hour day, the time I have to concentrate on writing this column is like a crack or fissure in the wall.

The most memorable period was in early November. Due to work reasons (government authorities), my wife had to undergo a two-week isolation period, during which we were completely cut off from communication. Naturally, the responsibility of taking care of our child’s daily life fell on my shoulders. Before the “isolation,” my wife even specially prepared a detailed checklist for this “clumsy oaf” of a husband, which meticulously recorded the daily routines. If you don’t have children yet, you might want to refrain from reading it; it’s easily discouraging.

An ancient saying goes, “Take the high, achieve the middle; take the middle, achieve the low.” Seeing the checklist my wife prepared, which represents the “high” level, I knew that with my understanding of our daughter’s temperament and my rough personality, I would certainly execute it at a “low” level.

The reason is simple: the best candidate for this checklist should be a full-time stay-at-home dad, and I am clearly not one. Therefore, in terms of execution, compromises are inevitable. I could only follow the “feed the child, keep them warm, keep them calm, prevent them from getting upset, and prevent them from getting sick” principle, and as much as possible, refer to my wife’s given best practices to find a “roundabout solution.”

For example, in order to let the child sleep soundly and for me to work on my drafts earlier, I had to come up with the innovative method of “carrying the child to sleep.” I would carry her around the room, swaying and strolling, and finally place her carefully on the bed.

To be honest, ensuring that the child doesn’t wake up due to changes in context (perceived temperature, light changes, sound vibrations, etc.) during the process of putting her to sleep is a more delicate task than disarming a bomb. Children are natural multi-functional sensors, capable of capturing various external signals, and how they respond to these signals (whether by simply turning over or by crying out) depends on their nature.

Putting a child to sleep is a technical and physical activity. After that, I am basically all sore back, aching waist, and cramping legs, almost collapsing on the ground. But the problem is, time is occupied by work and taking care of the child. When can I find time to write the drafts?

Staying up late seems viable, sacrificing sleep time to catch up on the drafts. However, I dare not do that. It’s not that I am unwilling to endure the hardship of staying up late, but I want to ensure the quality of the content production, and I am also concerned that I might make mistakes the next day due to tiredness while sending my child to school.

So when multiple important and urgent matters press down on me, I pay even more attention to the quality of sleep. Only by maintaining high energy levels can I “multitask.”

In situations where my regular time is fully occupied, I can only squeeze into the time rift — using fragmented time to complete sections. For me, the fragmentary time I can utilize is the subway commute to and from work. It takes about two hours round trip from home to the office every day, minus the time spent on transfers. At best, I have a little over an hour to thumb-type on my phone. Despite the noisy and crowded environment on the subway, I found that people are actually more focused under pressure. During that time, I was able to write about 800 words every day on the subway.

In order to help you learn Spark more easily, I will actively think about life metaphors or analogies. Many examples in the column, such as construction site brick moving, Spark Group stories, and a glass of milk, are inspired during my daily commute. In the evenings or weekends, I systematically organize, illustrate, add code and comments to the fragments accumulated throughout the week, and finally edit them into a complete manuscript.

The pace of work and life for modern people is fast, and our time is fragmented. People often use fragmented time as an excuse to refuse to learn: “The daily commute is only 2 hours, with several transfers in between. My thoughts are constantly interrupted, making it impossible to focus on learning. It’s better to watch videos!”

However, in reality, a systematic knowledge system and fragmented content consumption are not in conflict. Building a knowledge system does require a large chunk of concentrated time, but once established, each knowledge point in the system can be learned in 20-30 minute fragments - the Pomodoro Technique with its 25-minute units is scientifically proven.

Taking Spark MLlib as an example, after studying that module, you may feel overwhelmed by the number of feature processing functions and model algorithms supported by Spark MLlib, to the point where you might want to give up. However, with a bit of sorting, we can find that different feature processing functions and model algorithms can be classified into certain categories. The differences and connections between different categories have also been systematically explained in the course.

Therefore, to master Spark MLlib, we don’t need to deliberately set aside large chunks of time to study every day. Not to boast, the Spark MLlib module in the column is already sufficiently systematic, from categorization to applicable scenarios, from basic classification to typical case analysis.

Through this “classification guide,” we have already grasped the main framework of Spark MLlib. What we need to do next is to utilize fragmented time, delve into time cracks, and learn about each specific function or model algorithm to enhance our existing knowledge system.

The same principle applies to learning Spark as a whole. The key is whether, after a day of work, feeling as tired as a sloth, you are willing to dive into time cracks and use all available spare time to continuously improve your knowledge system with perseverance, like water gradually wearing away the stone.

In fact, time cracks like commuting are not the only passive ones. At work, we can actively divide time into one crack after another according to our needs. In each crack, we only focus on one task and do not accept any distractions.

For example, during the 8 hours of work each day, we can carve out multiple discontinuous coding time periods. During these periods, we pay no attention to any instant messages and concentrate solely on writing code. Of course, from these 8 hours, we can also carve out multiple meeting cracks, during which we temporarily ignore any bugs that need fixing and focus solely on discussion, communication, and persuading others.

In summary, the essence of time cracks is focus—100% focus on one thing. This is somewhat similar to how a CPU works. The clock cycle of a CPU is fixed, and in each cycle, it can only process one task. The serial working method may seem silly, but at the end of the day, you will find that the CPU has actually accomplished many things.

Speaking for myself, in the past four months, time cracks have helped me finish reading a book called “The Art of Thinking Clearly” and complete a Geek Time course called “36 Lectures on Technical Management.” While busy with output, being able to maintain consistent input makes me feel grounded and happy.

Being a determined producer of (technical content) is a long-term goal I have set for myself. And to continue producing high-quality content, continuous learning is essential. Water is soft yet strong, adaptable to all shapes, and never surrenders to difficulties. Bruce Lee once said, “Be water, my friend.”

Let’s seize every opportunity for growth, enter time cracks, continue learning, and encourage each other.

Finally, I have prepared a graduation survey for you. It is not long, and you can complete it in about two minutes. I look forward to hearing your thoughts. Thank you.