We de-risked our editor upgrade
10 hours ago
- #risk-management
- #software-rewrite
- #collaborative-editing
- Full rewrites of software are generally considered risky and often abandoned after significant resource investment.
- The team faced challenges with their Slate-based text editor due to an upstream rewrite that deprecated their version, leading to maintenance of a public fork.
- Issues included browser updates breaking the editor and the Slate-related code being difficult to debug, limiting the number of engineers who could work on it.
- The decision to upgrade was driven by the need for better stability, a more approachable codebase, and to allow more engineers to contribute.
- The team involved the entire engineering team early, sharing progress and architectural decisions to dispel the editor's 'mystery' and gather input.
- A proof of concept focused on collaboration, tables, and annotations to de-risk the upgrade project.
- The upgrade project was successful, with contributions from over 60% of the engineering team, and the new editor was rolled out to customers after thorough testing.
- Key takeaways include defining the necessity of the rewrite, de-risking critical parts first, constant demos and discussions, and building team excitement.