Redundancy Driven Development - Navigating disruptive codebase changes.
summary
Shows how rewriting risky subsystems succeeds by adding reversible redundancy: extract interfaces, add a delegator that switches between legacy and new implementations, generate ordering with KSP-generated comparator and sort order annotations, and gate rollouts with a feature toggle 5% rollout while syncing checkpoints on app start/login/network/logout. Uses a two-way state machine to migrate ~10 million users and can revert users when tokens invalidate; example parallels Apollo AGC’s checkpointing and surviving five program alarms.
AI-written summary · about
chapters (6)
- 0:00Introduction to redundancy
- 4:47Redundancy in Apollo missions
- 13:29Rewriting app initializers
- 21:39Migrating authentication systems
- 31:23Adoption and cleanup
- 34:13Q&A and closing
related talks
44mReversible Decisions for Better Software Design
35mQuickly and Safely Refactoring Legacy Code
01h 29mMonolithic Delivery
43mPatterns of Legacy Displacement
39mRethinking the SDLC
43mHow to Innovate in a Legacy
01h 27mRedesigning Legacy Systems : Keys to Success
57mCtrl-Alt-Del: Learning to Love Legacy Code