The open source boss left Google: stagnant on the Go language project, going to a smaller enterprise to seek change

Compilation｜Nuka-Cola, Yan Shan

On July 18, local time, Google Go language project leader Steve Francia announced his resignation. Steve Francia is well-known in the circle. He was not only the core leader of the well-known open source projects Docker and MongoDB, but also the author of open source projects such as spf13-vim, Hugo, Cobra, and Viper.

Steve Francia announced to join the Go team in September 2016. It has been 6 years since he decided to leave this time because he felt that he was “stagnant” in the Go project and had not learned for a long time, so he will focus on the next Data science and machine learning field, and chose to join the US hedge fund investment firm Two Sigma. Two Sigma is known for managing investment strategies using a variety of technological approaches including artificial intelligence, machine learning, distributed computing, and more.

From the statistics of the use of Go language by TIOBE Index, the development of Go language in recent years is relatively stable.

When Francia joined the team, the Go language was still a niche language. After several years of development, it has now become one of the mainstream enterprise languages. According to StackOverflow’s 2022 annual developer survey, Go ranks in the top 10 for both the most popular languages ​​and the high-paying languages.

Francia specializes in running open source projects, working at MongoDB and then Docker before joining the Go team. During that time, MongoDB grew from an obscure 30-person startup to one of the most popular databases in the world. After leaving MongoDB, Francia noticed Docker again, and joined in and was responsible for its strategy and operations, helping guide Docker’s participation in the OCI (Container Runtime Standard) and CNCF (Cloud Native Computing Foundation) ecosystem.

After leaving Go, Francia will continue to explore open source partnership opportunities in data science and machine learning. Francia emphasized that he is proud of the achievements of the Go team over the past 6 years, and is also looking forward to the future of the Go language. If you want to know more about why Francia left, what to do next, and what he misses most from his time at Google, take a look at his autobiography, excerpted from a blog post by Francia:

Go: My passion for the past decade

I first came across Go ten years ago and fell in love with the simple and elegant language right away. Go has brought me back to the enthusiasm of learning programming, so I used Go desperately, and slowly built up multiple tool libraries such as Hugo, Cobra, Viper, etc.

This passion also drives me to join the Go community. I have spoken at the first Gophercon, the first Gotham Go, and the first Gopherfest. I remember sitting in the room with Bill Kennedy on the eve of Gophercon’s inaugural convention, stacking all the bags at Erik and Brian’s feet. Before the second Gophercon, Brian and Erik invited me to host a roundtable with Mark Bates. We quickly became close friends and participated and hosted the next few Gotham Go and Gophercons together.

Through this experience, I have developed some friendships with several members of the Go team, and finally earned myself the opportunity to join it.

Go language enters mainstream enterprise applications

I joined the Go team six years ago with the idea of ​​taking Go from a niche language to a mainstream enterprise language. Looking back now, we should have accomplished or even surpassed this stated goal.

When you first joined the Go team, your main focus and roadmap was to address the needs of Google’s internal users. I believe that the current and future needs of the global user community should be addressed. After reaching a consensus, everyone began to focus on finding and solving the shortcomings of the Go language in terms of developer experience.

The first project I worked on reflected this change in focus. Under my leadership, the Go team decided to move forward with the biennial Go User Survey, one of the largest developer surveys in the world. The survey has given a huge boost to the Go roadmap over the past six years. Direct feedback from users led the team to decide to add modules and generics to the Go language, aiming to address two of the core barriers to Go in enterprise applications.

Based on feedback from business users, we’ve also made a major update to the Go brand. So Go.dev was officially released as a centralized platform for all Go web properties, covering 30+ case studies, key use cases, a learning center, a Go toolkit discovery portal, and more.

We also shipped VSCode Go and Gopls, which significantly improved the Go developer experience. We’ve improved documentation and language tutorials based on use cases identified by Go users, refined the Go installation process, and established Go workspaces to advance multiple module projects.

In order to go global, we began to actively embrace communities and contributors around the world. We tried our best to bring Go to as many countries as possible, including China, and held several developer workshops. The Go project has also changed from being dominated by Google employees to being dominated by community contributors. For continuous improvement, we’ve also made adjustments to the Go project’s code of conduct to ensure that the Go community remains warm and friendly. We have a vast network of Go developers sustained by nearly 200 conferences in over 50 countries.

In the past six years, the overall size of Go users has grown by about 10 times, and the frequency of user usage has also changed from occasional to daily use. Today, Go is widely regarded as a mainstream enterprise language. In addition, Go has also become a powerful cloud computing language, with more than three-quarters of CNCF projects written in Go.

I’ve been fortunate enough to play a little role in assisting the team and community, and have seen how, step by step, everyone has worked tirelessly to get Go to where it is today.

Why choose to leave

Since there are so many wonderful and wonderful memories, why did I choose to leave? In fact, with the release of the previous Go versions, the feeling of wanting to leave has become stronger. When I first joined, I was frantically digesting the knowledge feast brought by Google and Go. But now, I haven’t learned anything new in a long time. While I still love the team and the job at hand, this feeling of stagnation is hard to swallow.

With the release of the landmark Go 1.18 release, I feel my mission on the Go team is complete. I wanted to explore where to go in the next phase of my life, so I made a short list of what I hope my next choice will be:

• Continue to work with outstanding teammates

• Continue to work with good teammates

• Gain the opportunity to learn new skills

• Forging new industry paths

• Explore open source partnership opportunities in data science and machine learning

• Serving in a relatively small business where some changes can be made

After months of searching, I believe I have found a new position that meets all of the above requirements.

Answer: Two Sigma

Looking back on my entire career, I have had the privilege of being at the forefront of innovation across a wide range of industries. I led the team that designed the groundbreaking user experience for MongoDB, and I led the core engineering team at Docker that developed Hugo, the world’s leading static website generator, and Cobra, the popular framework that enabled the renaissance of CLI applications. After Go, I will focus on data science and machine learning.

Finally, I found my new home – Two Sigma. They have been at the forefront of innovation in data science, distributed computing, and machine learning for the past two decades, with deep investments in open source.

Through my interactions with Matt Greenwood, David Palaitis, Jason Bigler, and several other colleagues at Two Sigma, I sensed their potential to move machine learning and data science forward. Maybe once again I’ll be at the forefront of another paradigm shift that gave me the excitement and enthusiasm I had when I first got into MongoDB, Docker, and Go.

The text and pictures in this article are from InfoQ