How this site is put together
Table of Contents
This site used to be a static site generator with a dependency tree. Now it is a directory of files. Every page you can read here is an HTML file somebody wrote on purpose, styled by one stylesheet, with one small script for the theme toggle, the menu, and search.
The rules
The design is pinned, not generated. All styling lives in a single site.css, every page carries the same navigation and footer byte for byte, and a check script refuses to ship the site if any page drifts from the shared chrome, links something that does not exist, or loses agreement with its markdown source.
Writing starts as markdown. A publish script stamps the date and reading time, moves the piece into place, and updates the index, the feed, and the sitemap in one motion. The markdown copy stays behind as the canonical text, because this site has a second face.
The second face
If you have a terminal nearby, try ssh arv.sh. The same markdown that becomes these pages is rendered as a terminal interface by a small Go program. Two renditions, one source, and a test that fails if they ever disagree.
There is nothing clever underneath: no framework, no build, no bundle. Files go in a directory, and a server hands them out. That is the whole trick.