Content is files, not a database

Every lesson, article, pattern and field note on this site is a markdown file in a git repository. No CMS, no database, no admin panel.

The advantages are unglamorous and large: content review happens in pull requests, history is git history, and the whole site is greppable. Finding every page that mentions a dataset is one command.

grep -rl "ride-hail-trips" content/

The disadvantage is that non-technical contributors need a text editor and a branch. For this project that is the whole contributor base, so the trade is easy — but it is a real constraint and the wrong choice for a site where a marketing team publishes weekly.

There is a second-order benefit worth naming. Because content is text in git, the whole archive can be regenerated, reformatted, or bulk-edited with a script, and the diff is reviewable before anything is committed. Doing that against a CMS means an API, a migration plan, and no undo.

Pick the storage that matches who edits it.