This post is part two of a multipart exploration of “open source in China”. If this is the first time you’ve ever heard of or thought about China’s open source ecosystem, I recommend reading Part I of this series, “Open Source in China: The Players”, so you have a sense of who are the major companies and organizations that move and shake that ecosystem. (See Part III "Open Source in China: The Trends", if you want to jump ahead.)
This post focuses on how open source factors into the overall market incentives and competitive dynamics that these players play in, thus “the game”. In a future post, I will discuss emerging trends and macro geopolitical implications.
What the Market Wants
For any large tech company, building a sticky ecosystem and platform is necessary to dominate the market. That calculus is no different for the Chinese tech giants. Of the ones with a decent open source presence that I highlighted in my previous post -- Alibaba, Baidu, Bytedance, Didi Chuxing, Huawei, Meituan Dianping, Tencent, Xiaomi -- they all have their own core domain but all encroach upon the others’ core domains as well. The competition is fierce and cutthroat with a lot of overlap. To become a sticky platform, you need to build an ecosystem. And strategically speaking, open sourcing technologies can effectively drive ecosystem formation.
One of Tencent’s open source projects, wepy (a framework to build mini-programs inside WeChat) is a good example. For WeChat to become the dominant platform, the app OS within the mobile OS, it needs developers to build more mini-programs to make WeChat something users can’t lie out. The best way to attract developers to build is to open source and maintain a framework like wepy. The strategic consideration is similar to why Apple, a company not known for open sourcing much of anything, open sourced Swift to incentivize developers to build more apps to make iOS a more valuable and sticky platform.
A quick scan of which GitHub project repositories are pinned can give you a quick sense of how these companies are prioritizing open source to support their core business. And if there’s no pinned project or active maintenance, which is the case for quite a few of these big tech companies, it’s an indicator that they don’t quite know what to do with open source and how to harness its power to build and benefit their ecosystem.
There’s one big tension here: platform lock-in. Any sort of lock-in, be it platform, vendor, or something else, is directly in conflict with the core value proposition and community expectation of open source. And developers eventually vote with their feet (or fingers) on whether a technology is true open source, or dependent on a specific platform, thus fake open source. To continue with the Swift example, its success is very much connected to the fact that it works on both Apple’s platforms and Linux, the widely-adopted open source operating system. How Tencent’s wepy and other big tech sponsored open source projects fare in its true open-source-ness remains to be seen.
Unlike big tech companies, where open source is done in service of their core business, for commercial open source startups, open source is in their DNA. It isn’t one strategic initiative out of many; it is who they are. On top of that, these startups need to make money from the open source projects they drive, unlike big tech, where the connection between open source and revenue generation is several steps removed.
Because most of the Chinese commercial open source startups are less than five years old, it’s a stretch to say their open source activities are building a platform just yet. Thus, there's also less concern for any platform lock-in from users, which helps with technology adoption.
Open source, as a software development strategy, tends to yield the strongest and most robust technology over time. So by and large, these startups are racing in the open to make their open source technology more mature, which will naturally help with commercialization down the road. Some of the most active projects, like TiDB (by PingCAP) and Apache Kylin (by Kyligence), have improvements and changes made every week if not every day. That level of activity is rarely present in big tech open source projects, where most of them are untouched for months.
Talent Acquisition & Retention
Open source is a great way to recruit and retain technical talent, and all the Chinese tech players who’ve embraced open source on some level are trying to maximize that benefit. A commercial open source startup routinely hires engineers who’ve made contributions to its open source project first. Big tech companies do the same. Every company’s recruiters scan through active and well-respected open source projects for engineers they can poach. Once successfully hired, companies with a decent portfolio of open source projects can better retain the talent by allowing them to contribute to open source as at least a portion of their work. Developers generally like to both work with open source technologies and contribute back to them. The process also helps them build a public collection of their work, a following among other engineers, and pave the way for better job prospects in the future. The psychology and incentive are not unlike an artist or musician.
It’s worth noting that talent acquisition via open source is not only happening domestically in China but abroad as well. By default, all open source projects are public and global, and so are the people who work on them. Open source projects are where developers from everywhere come together, which makes open source arguably the strongest technical talent pool in the world.
What the Government Wants
The Chinese central government plays an outsized role in the country’s technological development. That shouldn’t surprise anyone. However, there are complex nuances to how the government’s vision and plan translate into tangible implementations on the ground. The Chinese governmental apparatus is often viewed from the outside as a monolith. It’s not.
The most relevant central government department is the Ministry of Industry and Information Technology (MIIT). MIIT has embraced open source. A quick search of the term “open source” on the MIIT’s website will yield many news releases of officials supporting open source in the context of cloud computing, big data, AI, and other areas. Those news releases may be terse and bland to read, but their importance lies in their very existence as signals of approval.
Why the central government would embrace open source is rather straightforward: it prefers to favor flexible technologies that aren’t tied to certain vendors, companies, or countries, so it can control and shape them at will. The thinking here is not that different from the rationale behind any large enterprise’s adoption of open source, in or outside China. “Self-reliance” as a national theme and technological imperative will be front and center for China for many years to come. This shift has been happening for some time, but accelerated by the U.S.-China trade war and Huawei sanctions. This acceleration has pushed strategic industries, e.g. banking, insurance, telecom, to more quickly adopt either domestic technology or open source technology, but most preferably domestic open source technology. All these shifts are leading to a boom in usage and business for both the big tech players and small commercial open source startups that I’ve mentioned before.
One important governmental player that is usually overlooked is the provincial government. While big, bold plans from the central government -- One Belt One Road, Made in China 2025, China Standards 2035 -- grab the most headlines, how these plans get carried out domestically depend on the competency and track record of provincial governments. Provincial governments compete for the resources that come out of these plans. They also have a decent amount of freedom when it comes to implementation details. And by doing a good job, the provinces get wealthier and the officials who got the credit tend to rise up the ranks.
In terms of tech, the two provinces that have stood out so far are Zhejiang, where Alibaba and Netease are headquartered, and Guangzhou, where Huawei and Tencent are headquartered. Beijing and, to a lesser extent, Shanghai are the other two tech hubs, but they are cities directly governed by the central government with no provincial layer in between. Given how important open source is to the overall technology vision of the country, any plans or initiatives that come from the top will likely filter down to companies based in these places first.
There are obviously many more nuances to how various levels and departments of the Chinese government impact tech and open source than I can capture in a few paragraphs. But I hope this overview provides a good starting point. For an analysis of the macro trends and geopolitics, please see Part III of this series: "Open Source in China: The Trends". If you missed Part I, go here.
Chinese Version Below
对科技最有关的中央政府部门是工业和信息化部（MIIT）。MIIT已经接受了开源。在工信部的网站上搜索一下 “开源” 就会有看到在云计算、大数据、人工智能等领域支持开源的新闻发布。这些新闻稿读起来可能简明扼要，没什么意思，但它们的重要性在于它们的存在，作为认可的信号。