Motherwell

BREAKING: Les Wexner’s attorney during his House Oversight deposition: “I’ll f*****g kill you if you answer another question with more than five words, ok?”

I think everyone's takes on this are like 10 years out of date xitter.cf/davidshor/status/19850…

Developers are full of clever ideas for internal tools… myself included. But we’ve learned the hard way. Most of them cost more to support than they’re worth. I rarely want to approve the time or budget unless there’s a clear path to value. Even then, I’d rather start smaller. So I mapped out an Internal Tooling Maturity Ladder for our team. Here’s the high-level version: Level 0 – One-off manual script (local, disposable, no repo) Level 1 – Shared manual script (still local, slightly more formal) Level 2 – Scheduled automation (pushes updates to Slack, Email, Sheets, etc.) Level 3 – Lightweight internal service (minimal UI, hosted) Level 4 – Fully hosted internal product (users, deploys, tech debt) We’ve overbuilt plenty of things that should have stayed at Level 1 or 2. I wrote about how we’re trying to do better on my blog: robbyonrails.com/articles/20… Curious how other teams approach this. When do you say yes to internal tooling? And when do you stop it before it starts?

This is better