Web Performance Deep Dive
summary
Claims measurable thresholds — 250ms competitive edge and Amazon/Google numbers (100ms ⇒ 1% sales, 500ms ⇒ ~20% traffic drop) — and concrete fixes: memoization reduces a naive Fibonacci(40) from ~1407ms to ~2ms, replacing forEach with short-circuiting iteration avoids scanning thousands of items, and avoiding eval/with/catch reduces dynamic-scope overhead. Network specifics: initial TCP window and RTT make small transfers costly (128KB can incur ~160ms), so upgrade stacks, enable window scaling and TCP Fast Open, use gzip/CDN/HTTP2, inline critical CSS, async/defer scripts, and minimize reflows with fragments and cached DOM lookups.
AI-written summary · about
chapters (19)
- 0:00Intro
- 1:19Challenge
- 7:48Javascript Tips
- 10:09Intentional Speed Delay
- 13:3214 Rules of Web Performance
- 18:52Optimizing TCP
- 20:39Fetching using 3G/4G
- 23:28How to Save Battery
- 24:10HTTP 2
- 26:29Critical Rendering Path
- 32:20Older Browsers
- 34:38Gmail Mobile
- 36:52Performance (60fps)
- 39:10Lazy vs Synchronous Update
- 40:02Layout To Paint
- 46:59Reducing Reflow Impact
- 48:34Thought Leaders to Follow
- 51:49Reduce Inspection Time
- 53:30Avoid keeping alive references
