I’m a massive tinkerer, I love breaking things to see why it broke and how to fix them. I prefer breaking software because if it goes too far, I can just reinstall everything. It was a great learning experience, and I’m glad that I had the time and opportunity to do it. As I grow up, messing with software is no longer just for fun and learning, it’s had become a way for me to optimize my workflow, to be more in tune with technology.

Even before I picked up programming, I customized my computer so that I can do whatever I’m supposed to do (probably Word and PowerPoint IIRC) as efficient as possible. Back when I was using Windows, I added custom keybindings with autohotkey, installing custom programs such as rainmeter to make my computer looks and performs differently. I was having a blast, but Windows has issues, which makes my experience wasn’t as good as it could be.

Problems with Windows

The biggest problem is performance. Windows is a tightly packaged product with many interconnecting parts. You can’t really modify Windows further than what its developers intended for you to do. Some programs can hack into Windows component and change them, but most of the time the best you can do is to add stuffs on top of Windows, which is not great in terms of performance. Windows being closed-source also means that these modifications, especially ones that poke deep inside, can easily break after an update.

Other than that, I mentioned above that whenever I brick my system, I can just reinstall everything. Windows wasn’t really designed to be reinstalled every other month. Just setting up all the regular applications (browser, text editor, …) was annoying, let alone customizations. Because of this, I thought that maybe it’s time for me to grow out of my hobby of customization and just use my computer the way it was “designed” to be used.

Switching to Linux

That was until I was introduced to Linux and especially tiling window manager. The lightweight and open-source nature of Linux means that you can add, remove, and modify whatever you want, and it stays performant. The options are also much wider, there are plenty of desktop environments to choose from. When I first started, I used KDE Plasma, and that alone gave me more customization options than I ever had with Windows. The performance just blows Windows’ out of the water, and I have more faith that it stays working because customization is a feature of KDE, not a hack.

The problem of constant reinstallation was also fixed, thanks to the faster installation process and package manager. It usually took me a few hours to install Windows, while for Linux (Kubuntu), it only took about 30 minutes, sometimes even faster. The package manager makes setting up applications much faster and easier than on Windows, given that the applications that I use are available on the package registry. And this is just the tip of the iceberg.

Using tiling window managers

Linux desktop environments are cool and all, but I needed more customization. Essentially, I want to create my own desktop environments. Naturally, I was gravitated towards windows manager, especially tiling ones. The idea of non-overlapping, non-minimizing windows organized into “tiles” and workspaces was so eye-opening for me. And the fact that most of these window managers are bare bone, and you are forced to customize them just makes them more appealing to me.

Eventually, I settled to the setup that I used to this day. It was configured from the ground up, gives me lots of control and feels lightweight and fast. At this point, I have learned a bit about programming and was comfortable with text-based configuration. This type of configuration makes reinstalling or setting up somewhere else even easier. After reinstalling, I just have to install all required apps with the package manager, then just copy the configuration files to the right directory.

Satisfied with the setup was an understatement to say the least. It looks the way I wanted it to be, it functions the way I wanted it to be, and I can somewhat rely on it being how it is. Although I’m open to alternatives (river, niri), I’m extremely happy with my current setup.

What about my Linux distribution

Note that I only mentioned the installation speed of Kubuntu, not that I used it for my current setup. This is because Linux distributions kind of don’t matter that much. People often associate distributions with their looks, and I used to be like that, too. But what really matter in a Linux distribution is the package manager. Everything else, including the bootloader, the init system, and the desktop environment are fully configurable.

The package manager determine how many packages are available to be downloaded, and how fresh these packages are. When I started making my Sway configuration, up until recently, I have been using Fedora and DNF because they strike a good balance between bleeding edge and stability. However, there are some problems with Fedora.

While setting up my dotfiles is as simple as a cp -r, installing required packages isn’t. I tried to write a shell script to automate that process multiple time, but every time Fedora decided to rename or remove a package — which happened more than I thought —, my script broke. This is because the package installation process is nondeterministic, as there are no equivalents of lockfiles for DNF. Also migrating to a newer Fedora release, while miles better than Windows, was still a pain in the back.

Entering the Nix rabbit hole

NixOS is something that I frequently heard about, but hesitant to give a try. I considered a “weird distribution” like SteamOS or Alpine because they do things differently and that difference requires special caution to work with. For example, this is what happens when you do something perfectly normal in other distributions: Downloading an executable and running it.

The package failed to run because it depends on a system DLL. That DLL may not be installed or have mismatched versions, which can affect the behavior of the executable. Such executables are called impure, and NixOS doesn’t like that. So to get around this, we have to build the application from source. In fact, the easiest way to get an application to work on NixOS is to build it in a special way to make Nix packages. This, and the fact that the process of building and publishing packages are so streamlined that the Nix package manager has the highest number of packages (as of the time of writing).

NixOS makes me thinks a lot about software distribution, which is a topic for another article. Let’s first get back into the less serious world of software customization. If you haven’t guessed it, I migrated my setup to NixOS, and there were some mayor improvements:

• Home manager allowed me to configure everything, even my web browser, in a single declarative format.
• Flakes allows me to version-pin my packages in lockfiles, so if there are breaking changes, I can still install the old version of my packages.

This addressed all the problems that I have with Fedora, even improve on things that I didn’t know I needed. Let home manager manage everything seems a bit dangerous, so I still maintain both my normal dotfiles and my home-manager configuration. Getting applications to work is a bit more difficult, but if something doesn’t work, I can just use a flatpak or even Windows executable and wine.

Conclusion

My current setup now uses NixOS and home-manager. After a few months of usage, I can say that NixOS is amazing. It’s absolutely not for the average user, heck, you need to learn a functional programming language just to configure your system. But NixOS had a solution for almost all of my customization problem since using Windows about reliability and stability. I have confidence that I can easily replicate my setup on any machine with NixOS or the Nix package manager. And I’m sure that new updates won’t break my configuration.

So that’s my journey of customizing my desktop, how I went from Windows to generic Linux to NixOS. Desktop customization never failed to give me new things to learn, Nix even deepens my understanding of functional, declarative programming. This is why customization is so integral to how I learn and use technologies. Customization is often considered “unproductive” by the developer community, but I think that finding what works for me, optimizing my workflow, learning and generally having fun is worth the effort put into it.