the illusion of best practices

Stub i haven't finished writing this yet. i publish drafts early as part of WFD 17.

i keep hearing people use the phrase "best practices" as a conversation-ender. like it's some kind of universal law that applies to every codebase, regardless of what we're actually building.

this is the trap we've built for ourselves. we take things that worked for google at google's scale and try to apply them everywhere else. i've done it too. i spent weeks on a kubernetes setup for a site that gets 100 visitors a day because "that's how you do devops." what a waste of time.

most "best practices" are just cargo cults. we see a big company do something, we assume they're smarter than us, and we copy the ritual without understanding the context.

it's the same reason people go on certification warpaths. they want a piece of paper that says they know the "right" way to do things. but the right way depends on the constraints. and your constraints aren't theirs.

if you're following a rule and it's making your work harder or slower, the rule is probably wrong. or at least, it's wrong for right now.

arguments against following best practices blindly:

• they ignore your specific scale and constraints
• they often lead to premature abstraction and over-engineering
• they turn engineering into a checklist instead of a thought process
• they're usually just opinions that got popular on hacker news

arguments for questioning them:

• you ship faster when you only build what you actually need
• simple code is easier to delete when you realize you were wrong
• you learn the why instead of just the what
• it forces you to actually understand your own system