You don't have time for a rewrite
1 min read
Premise: Rewrites feel like progress, but they are often procrastination disguised as engineering. Most teams need a thin slice and a migration plan, not a reset. Links (optional)
- Any post you’ve written on “shipping as a habit” or “tool-agnostic pragmatism”
Pointers
- The rewrite impulse: cognitive relief + fantasy of a clean slate.
- A rewrite is a scope increase, not a refactor.
- “Thin vertical slice” framing (tie back to Ryu 2.0 acceptance tests).
- How to decide: 3 signals you actually need a rewrite (and 3 you don’t).
- A pragmatic alternative: strangler pattern, wrappers, and guardrails.
- How AI coding tools make rewrites more tempting (and more dangerous).
- The hidden cost: re-learning, re-debugging, re-building trust.
Avoid / traps
- Don’t be anti-rewrite absolutist.
- Don’t use architecture jargon as a crutch. Keep it concrete.