Lightning Talk: What Can We Learn from the Edge Projects?
summary
Argues Ruby is viable for enterprise workloads with concrete fixes not slogans: rewrote C database drivers in a week to remove objections, measures a Merb request at ~500 microseconds and finds controller cookie setup costs ~70 microseconds so trimming tens of microseconds yields hundreds of extra requests per second, notes alias_method_chain produces noisy stack traces, recommends designing clear module APIs instead of gluing code together, pushing more integration tests, and turning quick hacks into real fixes over time.
AI-written summary · about
chapters (7)
- 0:00Intro
- 0:27Overview
- 1:03Build Ruby for the Enterprise
- 1:47Tradeoffs
- 4:52Microseconds
- 5:57Rubys not APIs
- 6:47Examples
