Disclaimer: These are my personal views and do not represent any organization or professional advice.

Fri, 16 Jan 2026 10:18:32 +0200

KISS Linux Retrospective

KISS is a Linux distribution I created in 2019 focused on simplicity, minimalism and sustainability. It was designed to be maintainable by a single person without any required infrastructure. The installation contained everything necessary to make this possible (full source code plus documentation).

The idea was to put all of the power into the user's hands. They would decide the makeup of their system, what software it runs, when and how it is updated and so on. Thus removing the possibility of the rug being pulled out from under them (which is too often the case with how fast software moves nowadays).

Of course, this was only possible if the system was minimal, simple (in design not use) and made up of few moving parts. The user would need a certain level of expertise as well as discipline to make it work. Dumb system, smart user. Smart system...

This prevented KISS from gaining wide adoption and garnering a large number of users but that was fine. The users it attracted were smart, capable and dedicated enough to maintain their systems and create a nice community around KISS.

Due to the high barrier of entry, every user in turn became a contributor and although at its core the distribution was small, its users expanded it, adding support for other software and architectures. Many created forks in order to explore different core ideas.

The repository system being decentralized meant that every user could independently share their repositories, packages and extensions without needing access to infrastructure or permission from me. The user was free to at any time disconnect from my repositories, use someone else's or even maintain their own.

I used to say that so long as there was one person using KISS somewhere the distribution would live on. And... live on it has. Despite my public abandonment of the project nearly five years ago, the distribution is still active with many old and new users. In other words, though not perfect, KISS was a success.

You can find KISS in the following places:

My old website: https://kisslinux.github.io
KISS Community: https://kisscommunity.org/
(where public development now takes place.
If you want to use KISS start here.)

Thank you to everyone supporting it and keeping it alive.

—Dylan Araps

Addendum

Privately I still maintain my own version of KISS and tinker with its core every now and then trying out ideas. At the moment, my system looks as follows.

I am using mimalloc as my system malloc.
I am using muon in place of meson.
I am using llvm+clang and have removed gcc, binutils, etc.
I am using blake3 for all package management checksums.
I am using zlib-ng in place of zlib.
My init and service management is based on s6 (basically what I was working on prior to leaving.).
I am using niri as my wayland compositor after trying it on a whim 6 months ago.

If I could go back and change one thing about the distribution it would be its release model. Instead of rolling release where maintainers are forced to work within the constant churn of new software versions and ever moving targets, I believe it would've been better to make a focused and polished release every three months.

This is what I've come to like about the BSDs. They have polish and cohesion. Something sorely lacking in Linux distributions which feel and act as though they are disjointed and haphazardly constructed (because they are).

Focus should not only be on each piece of software making up the distribution but also on what the resulting compilation of them looks like. Software shouldn't be installed, it should be integrated.

Sure, this would necessitate larger changes to the source code of packages. From changing default configurations, writing missing documentation and most invasively, patching and making larger code modifications... So be it. With the extra time gained from a slow release schedule this work could be done.

Look at it this way: The software in the distribution is for the distribution and not the distribution for the software. But, most importantly, the distribution is for its users. If it's not serving its users what is it for?

These are just some ideas. I may put them to action someday, I may not.