The version control habit I learned from patient records

Before I ever used git, I'd been taught the philosophy behind it — in the dispensing lab, by the rules of patient records. Medicine has maintained audit trails for centuries: who administered what, when, and why. Nothing is erased; corrections are appended. Git is that discipline, implemented in hashes.

What records discipline teaches

History is not optional. In health records, deleting history isn't just bad practice — it destroys the evidence future decisions need. Software learned this late: immutable logs, event sourcing, append-only audits. The hospital got there first.

Corrections are new entries, not rewrites. A wrong dose recorded stays recorded, with a correcting entry attached and its reason. In code terms: revert commits, postmortems, blame history. The alternative — silent overwriting — is how systems (and patients) get hurt twice by the same mistake.

Attribution is part of the data. Every record knows who made it. Not for blame — for context: the night-shift note reads differently when you know it was night shift. Code review, commit authorship, deploy logs — same principle, same value.

How it changed my repos

I stopped deleting things. Deprecate, mark, redirect, explain — but keep the trail. Squash thoughtfully, rewrite never, force-push as a last resort on shared history. A repository is a patient chart. The next person treating it needs the full story, including the mistakes. Especially the mistakes.