Office Hours – Rewrites, with Sunil Pai and Mark Erikson
Rewrites! Everyone wants to do one, but the software industry is littered with examples of failed rewrites. Our host Carl Vitullo is joined by Mark Erikson and Sunil Pai to talk shop about what they’ve looked like in our careers, and how to help make sure they’ll be successful.
Sunil Pai has loomed large in the React ecosystem for years, working on the core team, and he’s now building https://partykit.io. Mark Erikson needs no introduction! He’s maintained Redux and Redux Toolkit for years, is currently building http://replay.io, and has been a fixture of the Reactiflux community since the early days.
Join us and we’ll help you avoid the worst of the many footguns when rewriting software.
@threepointone / threepointone@m.webtoo.ls
@acemarke
@reactiflux
@vcarl_
Topics discussed:
[00:01:11] First rewrite
[00:02:52] Some web dev history
[00:03:54] Mark's refactor experience
[00:06:36] Zero cost feedback loops
[00:07:26] Sunil's two largest rewrites
[00:10:51] Mark's experience, government project timelines
[00:12:59] Replay, Mark's current work
[00:19:29] Rewrites on small teams
[00:20:11] Suspense rewrite at Facebook
[00:24:51] Tooling support to make rewrites work
[00:28:56] Zero cost feedback loops, revisited
[00:35:27] Automation as a way to enable communication
[00:36:32] How to evaluate a possible rewrite
[00:40:09] Smell tests for justifications
[00:44:45] Joel's Spolksy's cultural impact on rewrites
[00:46:23] Readability as a justification for rewriting
[00:50:15] Software lifespan
Rewrites! Everyone wants to do one, but the software industry is littered with examples of failed rewrites. Our host Carl Vitullo is joined by Mark Erikson and Sunil Pai to talk shop about what they’ve looked like in our careers, and how to help make sure they’ll be successful.
Sunil Pai has loomed large in the React ecosystem for years, working on the core team, and he’s now building https://partykit.io. Mark Erikson needs no introduction! He’s maintained Redux and Redux Toolkit for years, is currently building http://replay.io, and has been a fixture of the Reactiflux community since the early days.
Join us and we’ll help you avoid the worst of the many footguns when rewriting software.
Topics discussed:
- [00:01:11] First rewrite
- [00:02:52] Some web dev history
- [00:03:54] Mark's refactor experience
- [00:06:36] Zero cost feedback loops
- [00:07:26] Sunil's two largest rewrites
- [00:10:51] Mark's experience, government project timelines
- [00:12:59] Replay, Mark's current work
- [00:19:29] Rewrites on small teams
- [00:20:11] Suspense rewrite at Facebook
- [00:24:51] Tooling support to make rewrites work
- [00:28:56] Zero cost feedback loops, revisited
- [00:35:27] Automation as a way to enable communication
- [00:36:32] How to evaluate a possible rewrite
- [00:40:09] Smell tests for justifications
- [00:44:45] Joel's Spolksy's cultural impact on rewrites
- [00:46:23] Readability as a justification for rewriting
- [00:50:15] Software lifespan
- (00:00) - Introductions
- (01:11) - First rewrite
- (02:52) - Some web dev history
- (03:54) - Mark's refactor experience
- (06:36) - Zero cost feedback loops
- (07:26) - Sunil's two largest rewrites
- (10:51) - Mark's experience, government project timelines
- (12:59) - Replay, Mark's current work
- (19:29) - Rewrites on small teams
- (20:11) - Suspense rewrite at Facebook
- (24:51) - Tooling support to make rewrites work
- (28:56) - Zero cost feedback loops, revisited
- (35:27) - Automation as a way to enable communication
- (36:32) - How to evaluate a possible rewrite
- (40:09) - Smell tests for justifications
- (44:45) - Joel's Spolksy's cultural impact on rewrites
- (46:23) - Readability as a justification for rewriting
- (50:15) - Software lifespan
Creators and Guests

Producer
Carl Vitullo
Solopreneur just vibing, posts are probably bullshit. Community lead at Reactiflux, the largest chat community of React professionals.
