How this blog works
On this page (6 sections)
Until now, every page on this site was a finished file. To change a word, I changed the source, built the whole site again, and uploaded it. That is a good way to run a library of books. It is a poor way to run a blog, where the point is to write something this afternoon and have it appear tonight. So we added a blog that I can write from a web page, and this post describes how it works.
Where the posts live
Posts are kept in a small database at Cloudflare, the company that already serves this site, and pictures are kept in Cloudflare’s file storage. When you open a blog page, the page is drawn from the database at that moment, using the same header, footer, fonts and colors as every other page. If I ever change the look of the site, the blog follows it at the next update. The rest of the site is unchanged: the books, the reader, the search and the Denominational Mapping Instrument are still plain files.
How I write a post
There is one private page for it. I can type into it with a small toolbar (headings, bold, italic, links, lists, quotations, pictures). I can paste from a Google Doc, and the paste is cleaned: headings, emphasis, lists, links and quotations survive, while the fonts, colors and spacing that Google adds to every word are dropped. Or I can give it the link of a Google Doc, and it copies that document in once. A later change to the Google Doc does not reach the post, so nothing on the blog can change by accident.
A post starts as a draft that only I can see. I can preview it exactly as readers will see it, publish it, take it back to a draft, delete it, and restore it. A deleted post shows readers a plain “removed” page instead of an error.
Pictures
A picture is shrunk to a sensible size in my browser before it is sent, which also strips the hidden camera data that photographs carry, including where they were taken. Every picture must have a description before a post can be published, so that people who cannot see it are not left out. The first picture in a post is the one that appears when someone shares the post on Facebook.
Who can get in
The private page sits behind Cloudflare’s sign-in, and only my e-mail address is allowed through. The blog’s own code then checks the sign-in again on every request, so if the first lock were ever misconfigured, the second would still refuse a stranger.
For readers
The blog lists posts newest first, twenty to a page. There is an RSS feed for anyone who uses a feed reader, and the blog is in the site map for search engines. Each post ends with a “Reply to the author” button. It opens the contact form with the post already named, so the message reaches me with the right context.
What is not here yet
There are no comments and no reader accounts. Replies come by e-mail, through the contact form. Accounts for readers may come later, and if they do, I will write about that here too.
The blog was designed and built together with Claude, an AI assistant made by Anthropic, and tested before it went live: the sign-in checks, the cleaning of pasted text, the pictures, the publishing and deleting, and the editor itself in a real browser.