Under the hood
The guest list maintains itself.
How: Trevor adds one line to his show notes, "Guest: Name (Band)," and the site reads it, even mid-paragraph. It handles multiple guests, bands, and accented names, and leaves the host off his own list.
Why: a new guest appears on the site the day the episode drops, with no one touching a page.
The copy keeps its own numbers current.
How: sentences on the About page read live counts at load. "Four seasons and forty-one episodes" becomes "five seasons and forty-two" on its own, with written fallbacks if the data is unreachable.
Why: no stale facts, and no one has to remember to update them.
Search answers in under half a second, on 3,000+ videos.
How: indexing the full archive takes 13 to 20 seconds, too slow for a visitor to wait on. The site rebuilds the index in the background and serves a stored copy. Three layers of stored copies, each added after a slow load was measured, not guessed.
Why: the archive can keep growing without the site getting slower.
A daily video that curates itself.
How: the homepage walks a fixed shuffle of the entire archive, one video a day. Trevor can pin, schedule, or skip from his control panel, and a skip sends that video half the archive down the list so it never steals tomorrow's.
Why: fresh every day, zero effort, and the owner still has the final say.
The calendar runs the festival page.
How: on fixed dates the planner and a banner on every page change themselves. Before the weekend, an invitation. After, past tense. Then the banner disappears. No one publishes anything.
Why: an event page that goes stale the Monday after looks abandoned. This one never does.
Mistakes can't reach the public.
How: an automated check runs before every release. A guest link pointing at a page that doesn't exist, a guest who drops off the list, a stale search index, bad festival data, a Watch button with no confirmed video, or an image carrying hidden file data stops the release. The live site stays as it was.
Why: Trevor isn't technical and shouldn't have to be. No one checks by hand, and nothing broken ships.
Private data stays private, automatically.
How: images can carry hidden data, like the device, the location, or the account that made them. The site's own logo files turned out to carry a design account's identity. Now every image is checked on its way out, and anything carrying hidden data fails the release.
Why: one leak is too many on a site that features real people.
