I work alone on tiny programming projects, and my Git history is a sea of `fix`, `update` and `stuff`. When I return three months later those messages tell me nothing. I'm trying a habit of writing what changed and why: ‘Handle missing CSV headers’ instead of ‘fix import.’ How much detail is useful for a solo developer? I'd like a history I can scan without turning every commit into a miniature essay.
Writing useful commit messages when working alone
- programming-help
- writing
- useful
- commit
150
5 replies
For solo work, I like commit messages that answer two questions: what changed and why. You don’t need a novel, just enough that future-you c...
For solo work, I like commit messages that answer two questions: what changed and why. You don’t need a novel, just enough that future-you can tell one fix from another. “Handle missing CSV headers” is way better than “fix,” and that’s already a good level of detail.
I’d keep the message short and put the deeper explanation in the code or a note if needed. If every commit turns into a mini changelog, you stop reading it. The title should be enough to orient you when skimming history.
I’m in the “one sentence, maybe two” camp. If the reason is not obvious from the diff, I add it. If the change is very mechanical, the what...
I used to think I needed full sentences every time, but that got tiring fast. Now I aim for useful nouns and verbs, not polished prose. It’s a log, not a diary entry.