what google summer of code taught me about reading code

i spent two summers as a google summer of code student under kde — one on arianna, one on kdenlive. i went in expecting to write a lot of code. i came out having read far more than i wrote, and convinced that was the actual point.

a mature project is a foreign country. it has its own customs, its own history, its own reasons for doing things the strange way it does them. before you can change anything, you have to learn to read it.

read for intent, not syntax

the first week i tried to understand every line. that does not scale — the codebases were hundreds of thousands of lines deep. what worked was reading for intent: following one feature from the ui all the way down, ignoring everything that was not on that path, and building a mental map of just that slice.

the commit history is a diary

git blame stopped being a way to assign fault and became a way to ask questions. why is this check here? who added it, and what did their message say they were afraid of? half the time the answer to "can i delete this" was written years ago by someone explaining exactly why you cannot.

code tells you what. the history tells you why. you need both, and only one of them is in the file.

ask the smallest possible question

my mentors were generous, but their time was not free. i learned to do the reading first, form a specific hypothesis, and then ask a question that could be answered yes or no. "how does this work" wastes an afternoon. "i think the snapping logic lives in the timeline model, not the view — am i looking in the right place?" gets you unblocked in a sentence.

i write more code now than i did then. but the habit that stuck was not a way of writing. it was a way of reading — slowly, with intent, and with the history open in the next tab.