You can big bang!
summary
Argues that full rewrites succeed for manageable, data‑driven services when backed by concrete plans: replace Scheme with JavaScript (V8) and Python with Go, validate by diffing outputs into a database, and add redesign scope so users and business gain value. Shows a production run over 10 million games, runtime cost drops of 98% reduction in runtime cost, and build time for a new integration falling from week to 3 minutes, with shipping on deadline and measurable user performance wins.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:33Why the internet says never rewrite
- 4:19Context behind rewrite warnings
- 6:46When rewrites make sense
- 7:07Understand before rewriting
- 9:33Design versus business needs
- 12:24Avoiding second system syndrome
- 14:32Bulletproofing big bang rewrite
- 17:01Executing the rewrite
- 18:27Reinvesting success dividends
- 20:12Real-world rewrite examples
- 26:30Closing remarks and Q&A
related talks
53mThe Unbearable Weight of Massive JavaScript
45mDid We(b Development) Lose the Right Direction?
31mMaking the Jump: How Desktop-Era Frameworks Can Thrive on Mobile
11mRevolutionizing JavaScript Performance
28mLa nouvelle vague
23mFast by default: algorithmic performance optimization in practice
25mSpeed at Scale: Optimizing The Largest CX Platform Out There
56mBetting on Performance: A note on Hypothesis-driven Performance Testing