A bit about the history, where my interest in own website publishing software comes from and what brought me here.
It all started with hand-crafted HTML, as it did for many from that time. And of course, the hand-built HTML was much better than those new editors that created HTML. Of course, we knew much better how to do it right. Somehow it was even correct because back then browsers were quite weak and computers had much less power, and complex pages with complex HTML took longer for the browsers to display.
At some point during all this, I got to know Radio Userland. It was a super exciting tool for the time, a database, a built-in programming language (Frontier - I never did much with it myself), that lived directly in the system and within this core of database and language then a web server that runs on your own computer and basically only has two tasks: to display the website and to create the website as HTML. To this, an editor for texts and scripts in outlines from which the entire website was generated.
I was fascinated from the beginning, but rather deterred by the programming language. But the concept was definitely one that interested me. So the next step for me was to build it all myself - but with a programming language that I was more enthusiastic about: Python. That was the birth of the Python Desktop Server - pyDS.
I was also a bit inspired by Niklas Luhmann's Zettelkasten. The idea of collecting and storing knowledge over a long period of time, constantly expanding, linking, and organizing it has always fascinated me. That's also the reason why my blog tools always had not only a simple blog, but also collections of links and references, images, larger pages with links in wiki form and random stuff that just came up as knowledge.
There was even a whole range of users of this software, which ran on a rather unusual stack at the time, because I built the entire server in the app itself. And the database was also rather unusual for normal projects (e.g. column-oriented storage). My first blog ran on this software for a while.
But eventually, the end came, it just didn't run smoothly anymore. The software was based on an old Python version and some of the libraries were no longer supported - it was time for something new. At that time, I was once again flirting with Wordpress and had it in use for a while, but I wasn't really happy with it. So there was the next rewrite, this time with Django. Also an exciting development, much of it was based on concepts that were interesting to me at the time (e.g. no static types on content, but generally posts, images and collections - the latter were saved searches that spat out a list of posts). That kept me going until 2010.
Why only 2010? Bitrot is real. The software becomes outdated if it is not regularly modernized and I was once again at the point where I could no longer realistically keep the old software running without a greater time investment. Which I simply no longer wanted to make - so I switched to Wordpress. I simply saved the old blog as static HTML with the idea "someday you'll transfer that".
Of course, that didn't happen. I stayed on Wordpress and it's doing quite well, but it's a server-side solution with PHP and MySQL. None of that is what excites me. But I also didn't have the motivation to change it, Wordpress was "good enough" for me. Blogging also became less and less frequent over time. The blog went to sleep.
What changed in 2026? AI Coding Assistants. I can bring my quirky ideas to code much faster. I have real access to huge stocks of ready-made libraries, because I no longer have to understand them myself - I just have to understand the concepts and have ideas about how I can bring them in.
So I sat down - after various other projects with Vibe Coding - and designed the basic structure of bDS, the basic vision. Much of the Python Desktop Server has survived, only now the core is Electron and TypeScript. But also much of what I learned from the past has survived: simple formats, files and directories first, database only for caching, reproducibility of the database, versioning, synchronization via version control - many of these are things I also wanted before, but were simply too complex to tackle.
So here we are, February 2026, bDS is far enough that I can write and publish with it again and the result is what I had before: static HTML that is trivial to host, and a lot of files on the disk where the content lies.
Let's see how long this will last. 8 years of my self-made stuff, 12+ years of Wordpress, let's see if bDS has what it takes to run longer.
In the long run, I want to collect articles here that explain parts of bDS. Let's see how much I can put together.