Cathedral vs Bazaar model

After reading The Cathedral and the Bazaar, I better understood why open-source projects can grow so fast even without a highly centralized team. The “bazaar” model feels very different from the traditional idea that large software must be carefully built by only a small group of experts. Instead, the reading shows how open collaboration, frequent releases, and contributions from many people can make software stronger over time. I especially liked the idea that a project does not need to be perfect before being shared, because real improvement often comes from many different people looking at the same problem from different perspectives.

And I do not think AI coding tools fundamentally change this idea. AI can definitely help with repetitive work, such as generating boilerplate code, summarizing files, or helping contributors understand unfamiliar syntax faster. These tools can make contribution easier, especially for beginners. However, I still think the “community” is the most important part of the bazaar model. AI cannot replace maintainers’ decisions, user feedback, design discussions, or the shared understanding that comes from many contributors working together. In that sense, AI may speed up the process, but the strength of open source still comes from the people behind the project.

Group project progress

During the past few days, I submitted my first PR for AFFiNE, which was really exciting for me. The issue was a language switching bug in the settings panel. The hardest part was actually finding the exact file related to the bug because the project is pretty large and I was still learning the file structure. Even though my PR was later replaced by a maintainer’s version, it still felt rewarding because it showed that my debugging direction was correct and my work helped point toward the final fix.

For the next few weeks, I hope my teammates and I can keep finding more manageable issues and maybe explore backend or desktop/mobile related changes instead of only web fixes. Reading other groups’ reports also made me realize many of us share similar blockers, especially environment setup and understanding large codebases. Overall, this week made me feel much more confident in reading unfamiliar code and contributing to real open-source projects.