Taming the Beast: Managing a really ambitious codebase
summary
Argues that combining Ember Engines with Yarn workspaces yields a mono-repo layout that is fast for users and manageable for developers, replacing a 300,000‑line CoffeeScript monolith with lazy‑loaded engines, a shared add‑on, and per‑feature packages. Reports builds and deploys become quick, regressions drop, teams stop duplicating work, and CSS falls by ~90% after adopting Tailwind while large refactors get safer with TypeScript; notes past pain points like 13,000 comments and being stuck on Ember 1.11 are resolved.
AI-written summary · about
chapters (13)
- 0:00Start
- 1:27The Beast That Needs Taming Is Complexity
- 4:23Strategies for Dealing with Complexity
- 6:43Zone Creation Tool
- 8:35Product Label
- 8:54Re-Entry Interval
- 13:05Using Yarn Workspaces with Ember Engines
- 14:35Ember Engines
- 15:56Clear Communication
- 21:00Mount the Engine
- 21:59Tips for Working with Engines
- 22:42Sharing Services
- 28:40Use Typescript
related talks
36mPanel: Ember at Scale
28mModernizing a Massive Monolith
33mScaling Frontend: Supercharging Productivity with Dedicated Infrastructure Teams
24mManage Your Huge Codebase like a Boss
01h 02mStability Without Stagnation: Lessons Learned Shipping Ember
24mHow to prevent a rewrite
34mMass Frontend Refactors: Less Stress; More Success
21mYehuda Katz and Tom Dale (Tilde Inc.) interviewed at Fluent 2014