Reading Code You Didn't Write

Entry points first

Before reading any logic, answer one question: how does this start? The README's run command, package.json scripts, main/index/server files. Every codebase is a tree; find the trunk before the branches.

Then follow one flow

Pick a behaviour you can see — a button label, a URL, an error message — and grep for its text. That string is a rope: pull it and walk inward one layer at a time, writing the file path down as you go.

git grep -n "Place order"        # where the visible thing lives
git grep -n "placeOrder"         # the function behind it
git log -p --follow -- src/checkout.js   # how it got this way

Grep, debugger, or ask?

• Grep answers *where* — free, instant, no setup.
• A debugger or print answers *what actually happens* — when reading isn't enough.
• A human — after 30 focused minutes, with what you tried: "I traced X to Y, expected Z" gets a mentor; "how does this work?" gets a link to the README.

Grounded in how engineers actually navigate unfamiliar repos