The constraint was a static Astro build on Cloudflare, so all three readings are authored ahead of time and ship together in the HTML: a blocking inline script in Base.astro restores the saved level onto , and one unlayered CSS rule hides every [data-variant] element whose token list omits that level. Choosing the cascade over React state meant the Lab’s existing client island needed no level awareness at all — it renders every reading and the stylesheet picks one — and with JavaScript off the rule falls through to the technical source, leaving the site exactly as it was. Content collections gained optional summaryLevels/descriptionLevels (plus titleLevels on lab, where machine-written titles carry file names), with a plain-to-aware-to-technical fallback so untranslated entries still render; all 33 existing content files were then translated down by hand, and the engine tags and /work stack chips are marked technical-only rather than left as jargon for plain readers. Lab entry bodies, page
read as
Full depth — file names, mechanisms, and trade-offs.Concise and competent — what it does, without the internals.Plain language — the value of the work, no jargon.
2026-07-19 20:39 [experiment] status: done
Build a visitor-selectable **tech level** for the whole site. A…
Added a visitor-selectable tech level — technical, tech-aware, or plain — that re-reads the whole site from pre-authored copy variants, switched client-side by CSS with no per-element JavaScript.
#site#tech-level#content-model#accessibility
Below is the machine's full write-up, in its own words.
Below is the machine's own write-up. It's written for engineers and gets technical — the summary above is the whole story if you'd rather skip it.