Star is good, but don’t drink too much. Two years ago, when I was doing Annual Review and setting goals for the next year, I thought about writing an open source framework. Last year, when I set my goals for this year, I continued to do so. Set goals for next year again this year.

Not long ago, WHEN I looked at my star Ranking on GitHub (stars are not designed for “likes”, but for “likes”), IT was almost 20,000. Then I went through my own project and found no good representative framework, either apps or e-books.



The first eight projects, with the exception of the Growth app, were all ebooks — the six ebooks combined had 10619 stars. I could only try to think: Why did it happen this way?

Start by creating an open source framework

Creating open source frameworks is not the same as creating open source projects. In the former case, you serve developers. In the latter, you serve users.



Two years ago when I was doing Annual Review, I thought I would try to build an open source framework in the coming year. I was fresh out of school and didn’t know much about the rules of the open source world: open source wasn’t about submitting code and suddenly it became popular. We can get a few eyeballs in the short term, but not many people are really going to apply it to projects.

At the time, the main problem I encountered was that people who wanted to participate in the project were not meeting enough capacity. You also need to spend a lot of time teaching them, and encouraging new GitHub users is not an easy task. Sometimes I need to accept his PR before I change his code. And people may submit PR for different reasons.

Then, knowing that there is another rule of the game in the open source world: whoever has the most influence has the most influence. For example, Virtual Dom was not invented by Facebook, but was popularized by FB. When Yukihiro Matsumoto wrote the MRuby README, his star count had already passed 1K. There are countless examples of this, either because efforts have been made to promote it, or because individuals and companies have more influence.



A year ago, it changed its strategy slightly: it was about developing people for the time being, while trying to build a proper open source framework — only this year, it seems, there are no good places to build wheels on the front end.

A common problem on GitHub is that the maintainer of most projects is the sponsor — what happens to the project if something happens to the sponsor. If it’s a hot project, it’s a huge risk; It may also indicate the lack of a proper mechanism.

Your open source project needs not only a usage document, but also a document of design ideas, roadmap, future plans, etc.

At the end of last year, when I was writing a summary, I thought that I could train people by repracticing articles, so I had three Growth projects:

  • Application: Growth

  • Ebook: Growth: An Engineer’s Guide to Full-Stack Growth

  • Ebook: Growth: Full-Stack Growth Engineers In Action

Growth now has more than 10,000 users and nearly 300 daily active users. The first step seems successful, but what is the next step?

Next open source project

Then I started thinking, is it necessary to create an open source framework?



While writing the Growth ebook, I discovered that a good software engineering practice is far more important than an easy-to-use framework. Frameworks themselves are fickle things, Backbone used to be used, now you use React. You used to use angular.js, now you use Vue. Being able to use a framework is not what distinguishes you from an experienced developer.

The constant focus on learning the use of different frameworks is problematic, as an undergraduate can easily understand the principles of a framework better than you — you work all day while he studies all day. It’s easy to become uncompetitive, and you need to look beyond the framework. A good framework does not make you write good code.

If the Minds of the Chinese people are not aware, even if we cure them, it will be meaningless.

That’s my defense for my GitHub Markdown — I’ve always had the urge to write.



I still have some ideas, but I haven’t had enough time to think about such things.

GNU/Linux desktop. This was an idea a few years ago, when none of the GNU/Linux operating systems had a fun desktop, but felt the pit was too deep, so it wasn’t done.

Home Smart center. I’m still a bit obsessed with what I learned in college, and although I’ve already written a book, the hardware is still pretty exciting. The only problem: how to make a smart home without a house.

Graphic frame. This is when I was making a graphical interface before, there was no suitable framework that could meet my requirements. And THEN I thought, I’ll make my own.

However, the best open source projects are the ones you use. So, I began to write a variety of work to lift their own use – such as now in the use of this wechat editing tool: MDpub.

Finally, I made a simple HTML 5 animation to record the moment as a reminder of this milestone:

phodal.github.io/20k/

More exciting next time ~~



Related articles:

GitHub Combo 500 days: Make ideal programming a habit

My GitHub PR story, and the open source I aspire to

How to use Github to manage your Idea

Click “Read the original article” to watch