Rebuilding the Site Around the Work

The site was one page that had accumulated everything: a live control surface, a set of service descriptions, some writing, and a lot of pages that existed because I’d built them at some point.

It had no shape. You couldn’t tell what was live, what was working, and what was an idea.

Four questions, four sections

Projects — things that exist, with an honest status on each one. Prototype, working, in development, private.

Lab — the technologies I’m actively touching. Deliberately high level. It’s a snapshot of interests, not a catalogue of products.

Build log — the working notes. What was learned, what broke, what I’d do differently. Not case studies.

Live demos — what’s actually deployed and running, separated from what’s experimental, with the distinction stated plainly on the page.

The part I didn’t expect: the classification problem

Every project needed a visibility level, and that forced a decision I hadn’t made before: what can I show publicly, what can I show as a teaser, and what stays private.

The teaser category is the interesting one. A page can show the problem, the approach and the outcome, without giving away the implementation. It shows capability without handing over the thing that makes it hard. That’s a harder discipline than just writing a good description, because the temptation is always to include the interesting bit.

Writing the status badges also made me honest. A project that isn’t working should say so. “Concept” and “prototype” are legitimate answers. “Nearly finished” is not.

Content as files

The writing lives in markdown files with front matter, and the pages render them. Which means a new build log entry is a text file, not a deploy. That’s mostly a convenience, but it also means the content outlives any particular page implementation — if I rewrite the site next year, the notes are still just notes.

What I’d tell myself at the start

The split I got wrong the first time was by section of the site rather than by stage of the work. People don’t come to a portfolio asking “show me your services” — they ask “what have you built, and does it work”. Stage first. Everything else falls out of that.