Personal website and technical blog of Sergio Bonatto.
The site is intentionally implemented as a static, dependency-light web application. Markdown files are used as the source of content and are converted into standalone HTML pages with Pandoc.
.
├── contents/ # Markdown source files
│ ├── archive/ # Archived posts
│ ├── quick/ # Short-form drafts
│ └── trash/ # Unpublished or discarded content
├── post/ # Generated HTML posts
├── index.html # Generated homepage
├── rss.xml # Generated RSS feed
├── style.css # Generated site stylesheet
├── favicon.svg
├── llms.txt # Machine-readable site index
├── cli_post.sh # Site generator
└── README.md
Posts are written in Markdown with YAML front matter:
---
title: "Post title"
date: "2026-09-27"
description: "Short description of the post."
---The generator extracts the title, date, and description fields to build the homepage, RSS feed, and llms.txt.
The site is generated with:
./cli_post.shThe script:
- Reads Markdown files from
contents/. - Converts each post to standalone HTML using Pandoc.
- Generates the homepage from post metadata.
- Generates the RSS feed.
- Generates the site stylesheet.
- Generates
llms.txt. - Injects navigation, favicon references, and the footer into generated posts.
The generated files are written directly to the repository root and post/.
The build requires:
- Bash
- Pandoc
- standard POSIX command-line utilities
- BSD
datewith-jand-fsupport
The generated site itself has no JavaScript runtime or server-side component.
The website is deliberately minimal. It uses semantic HTML, a small stylesheet, system/browser fonts, and static resources.
There is no frontend framework, client-side router, JavaScript application layer, or external content-management system.
The repository therefore contains both the source content and the complete generated representation served by the web server.