This sample post gives the new publishing pipeline a second article to work with. It is not a report of a finished deployment.
Reading should be simple
An article does not need a database query every time someone opens it. Markdown becomes HTML during the build. The same build can produce search data, a sitemap, and RSS feeds.
Put the moving parts behind a boundary
Requests for conversations, comments, and newsletter verification need validation and persistent state. Those belong behind focused API routes, separate from static reading.
The intended split makes it easier to keep the site readable even when email delivery is paused.
Diagram: two different paths. Reading uses prebuilt HTML. Submission processing will go through a Worker to persistent state; that path is still being implemented.
Keep a record of the choices
I want this site to be useful as a notebook, including the decisions I change my mind about. Each deployment will show its Git revision in the footer. Meaningful page changes will have their own RSS feed; a typo correction does not need an announcement.
Around the table.
Comments will open when moderation is ready. In the meantime, you can keep reading through RSS.