What I learned in Alibaba

I joined the DingTalk division of Alibaba Group as an intern at the beginning of August. I “formally” started learning computer science in December last year (how I learned CS will hopefully be updated in a later post) and I didn’t have any internship on my resume, so I honestly didn’t expect they would accept me. But anyway, they did. I’d say I performed adequately in the interview; I got most of the technical questions right (fortunately they weren’t too harsh on LeetCode questions). I remember I “introduced” Raft to a P9 during my second interview and said I was going to read Designing Data-Intensive Applications. It turned out that I didn’t read a page of it after I started my internship, and here’s why.

On the second day I was thrown into piles of production code. My mentor said they expect interns to start working immediately. I replied: “Okay. But how do I set up Maven? “

So I rushed the entire development stack in my first month, including the basics of Java (I used Thinking in Java), Spring(Boot), Most of the Alibaba middlewares, Design Patterns and Refactoring (you know which two books i am talking about), and of course, Maven (it was actually the easiest of them).

But these are not that important. I can read book and learn all these without doing an internship at all. Here are some of the things I want to highlight during my days at Alibaba:

Understanding Code

I probably read the largest amount of code in my life during the past two months, and I realized “reading” code has nothing in common with reading a book.

• Don’t try to understand everything. Actually, you can’t. Just read what you have to read, e.g. these related to your current work. Your new functionality might involve 3 other applications. Only read the code that’s relevant.
• But make sure you understand the details (of the code you have to read).
  • When following a function call chain, you will run out of brain cache. Take notes if necessary.
  • A particular good method is to log down (in your code) the key data objects and see how these functions changed them. A key data object is typically an input or an output of a key function.
  • Use your IDE well. I particularly like Find usages and Hierarchy(control+oprion+H) in IntelliJ.
• Write code that is easy to understand. We have coding rules that specifies how to name a variable and such. I really appreciated it that most of us followed them.

DevOps

CI/CD is good. Just feels so nice to have a CI/CD platform. Release stages are clear and everything is under control.

Test and test. Unit test. Functional test. Code review. Integration test. Regression test. That’s what I call risk management.

Log is God. I can’t remember how many times I resorted to logs. Use your grep and tail -f well. Probably you also want tail -f | grep --line-buffered.

Communication

I had been self-studying for months and I always thought I could solve everything by myself (in fact, by the Internet). Therefore, during my very first Alibaba days, I tried to read code without asking any questions (partly it was because everyone seemed so busy). BUT I JUST COULDN’T UNDERSTAND.

In retrospect, it was because there was so much business logic and data objects that I couldn’t possibly understand without asking my collegues.

So gradually I started asking. I would say things got much better.