Senior Devs on Code Comments: Do As I Say
Description
A two-part meme about developer hypocrisy regarding code comments. The top part contains a text-based dialogue: 'junior dev: why is it important to include comments in the codebase?', followed by 'me: because they supoorts future reading and modification', and the punchline from the junior dev, 'how come your code has none?'. The bottom part is the reaction, featuring a two-panel image of a goofy-looking orange bird with large, offset googly eyes. In the first frame, the bird stares blankly. In the second, a close-up shot is captioned with the unspoken thought, 'Listen here, you little shit'. The meme humorously captures the 'do as I say, not as I do' attitude some senior developers have, where they enforce best practices on juniors that they themselves neglect, often due to time pressure, overconfidence, or habit
Comments
7Comment deleted
Comments are for the weak. My code is 'self-documenting,' which is shorthand for 'if you can't understand it, that's a you problem.'
I do comment my code - the annotations are just sharded across 3,482 rebased commits. If you can’t reconstruct the architecture from git blame, are you even reading the documentation?
The real reason senior devs don't write comments? After 15 years, we've convinced ourselves that our variable names like 'x2_final_v3_PROD_DO_NOT_TOUCH' are self-documenting, and besides, if the code was hard to write, it should be hard to read - job security through obscurity
Every senior engineer has that one 10,000-line God class with zero comments that they wrote during a 'temporary' sprint three years ago, which is now the core of the entire system. We all know comments are crucial for maintainability - we just also know that 'future us' who has to read this code is a problem for future us to deal with. The real irony? That junior dev will be maintaining this uncommented mess in six months, at which point they'll understand why we get so defensive when called out on it
Comments rot faster than code; mine's self-documenting via the Jira epic titled 'WTF was I thinking?'
At scale, comments are the fastest stale microservice; I prefer tests, types, and ADRs - the only docs that fail CI when reality changes
My code isn’t under-commented; its intent is sharded across ADRs, git history, and two postmortems - eventual consistency for documentation