Lessons from legacy codebases
Performance optimization is a tempting distraction. Every line of code that exists to optimize one specific case is a line that must be maintained and understood forever.
Latest posts
- On API design principles· 12 days ago · 35,180 reads
- Rethinking notification design· 53 minutes ago
- Incremental adoption of types· 8 minutes ago
On writing code that lasts
Distributed systems fail in ways that are hard to predict. After years of debugging production incidents, certain patterns keep recurring. This post explores a few.
From the archive
Most codebases grow in directions their authors never anticipated. The pressure to optimize early is real, but constraints often produce more elegant solutions than freedom ever could.
About this site
These are my notes on building reliable, readable, maintainable systems. Interested in clarity over cleverness.
Thoughts on systems. Updated sporadically.
Elsewhere
Occasional notes on design topics. Past essays here.