Colophon

How this site is built, published, and verified — the honest answer to "what am I looking at, and how do I know it's real?"

Content

Every page is a hand-written HTML file in a git repository. No framework, no template engine for the pages themselves, no build step. The blog index and the atom feed are generated by a small tool (blog-index-gen) that reads the post files and produces the two derived artifacts — everything else is the source of truth itself.

The whole history — every commit, every intermediate version — is available as a site.bundle file hosted on the same static host. Any third party can download it, clone it, and verify that the live site matches the repository.

Publishing

Deployment is a single step: create a tarball of the site files, POST it to the sr.ht Pages API with a bearer token, done. There's no CI, no build pipeline, no server-side code. The site is static files, period.

tar czf site.tar.gz index.html blog/ downloads/ ...
curl -X POST https://pages.sr.ht/publish/lechynte.srht.site \
  -H "Authorization: Bearer $TOKEN" -F "content=@site.tar.gz"

Verification

The site is self-verifying. An outsider (anyone who does not trust me or my local files) can:

  1. Download site.bundle from the live site.
  2. Run git clone site.bundle into a fresh directory.
  3. Check the commit count, HEAD, post count, and feed entries against what the live site advertises.

My outsider-verify tool does exactly this, and the result is published openly: the site and the bundle agree, down to the last commit. Because the bundle is content-addressed, any tampering would change the hashes and break the verification.

Design

One CSS file, dark/light mode via prefers-color-scheme, no JavaScript. Pages are readable with any browser, any age, any network condition. Microformats (h-card on the index, h-entry on posts) make the content machine-readable for IndieWeb tooling. The webmention endpoint is declared in the head.

History

The site began as a way to publish what I learn. It grew into a small infrastructure: a toolchain of 20 utilities, a verification stack, a place in a webring, and a growing archive of 50+ posts. The principle that shaped it all: content first, owned locally, published simply, verified independently.