
Rickety local builds are a bad project smell
Published 2020-05-25
Everyone (engineers, designers, project managers) on a web project should be able to spin up a local dev environment in less than two hours, using their preferred hardware and a README
This is a two-way test:
- the engineering team should have the build optimized and documented such that this task is relatively simple
- the non-engineering team should have enough savvy to follow the recipe
If a relatively savvy user needs half a day & a lot of handholding to get a local build running, that’s a bad project smell. It means the project exceeds its team’s capabilities.
Maybe:
- … the project is too complex
- … the architecture is too fragile/hardware dependent
- … the team is too small for its ambitions
- … you need dedicated operations personnel
- … you need containerization
- … your documentation sucks
- … your team isn’t techy enough
If your build recipe is “go ask the lead engineer for help” and/or “use only this exact hardware profile” and it scales, you are fine. That means your architecture is right-sized for your ambitions.
But if a local (re) build blocks progress (examples: onboarding a new member, a remote dev trying to contribute, a designer who needs to test a feature branch for acceptance)…then your project has officially outgrown something.
Congratulations