← experimental
Contents
  1. What This Is
  2. How to Use It
  3. How it Works

Hi, I'm John. I'm an engineering manager. I used to be a software engineer, favouring frontend work. I've been reading a lot about software architecture lately, partly because it seemed like useful knowledge to have at this stage of my career, and partly because it feels more important than ever now that AI can write code faster than most people can review it. A solid understanding of architecture means you can actually evaluate what it's producing, catch the mistakes, and point it in the right direction.

So I've been reading a lot of books. Books are great, I love books. (No, they're not as effective as "learning by doing", but they occupy a niche that I will never fill with coding: winding down for half an hour before bed, not staring at a screen.) The more I read about software architecture, the more I noticed that a lot of books cover the same ground. But some go into greater depth, some are more opinionated, and occasionally they just disagree with each other.

Rather than just agree with the last thing I'd read, while previous books fade from memory, I wanted a way to combine all that knowledge in a single place for easy reference.

What This Is

This is a wiki full of software architecture concepts, split into rough sections that seem to make some sense. All the material comes from relevant books, referenced throughout.

I used AI to read ebooks, split them into chapters and process them in batches. For each chapter, the AI pulls out interesting concepts and either creates new pages or updates existing ones, combining perspectives from different sources rather than just summarising each one in isolation.

Whether that makes it better than the usual AI-generated content is for you to judge. It works for me, anyway. Hopefully you find it useful too.

How to Use It

You can click around the website, hosted on GitHub Pages. Or you can clone the underlying repo and point something like Obsidian at the wiki directory. You can also spawn an AI agent within the project to ask it questions about the content.

The website has a simple search feature you can access via the button in the header.

How it Works

Check out the README in the repo for more details.

I originally wanted to produce a directory of markdown files, which, when opened in Obsidian, provided a nice UX.

Then I created a build process to turn all of those files into static HTML pages.

If you have any suggestions, or notice any problems, please raise an issue on the GitHub repo.