Arch on desktop/laptop because I’m very comfortable with it, and I can set it up the way I like.

Debian on servers because it’s stable and nearly everything has a package available, or at least instructions for building.

Same as OP, but I’m not likely to change them out. I’ve tried a lot of distros over the years and this is what works best for me.

You could get started with Qt, specifically the legacy widgets. There are bindings for Python available (pyside or pyqt) if you don’t want to learn C++ or another language right away. You can also port your GUI definitions to other languages at a later date.

This is basically the holy grail for finding how to do things or troubleshooting once you have the basics down: https://wiki.archlinux.org/title/Main_page

A large amount of information is transferrable to other distros, particularly if they use systemd.

I would also spend some time getting comfortable with the command line. There are a million tutorials, guides, and free courses on this topic. Find a shell and terminal emulator you like. I’m using Fish and Kitty these days.

Going against the grain, maybe consider EndeavourOS for a distro? https://endeavouros.com/

In this scenario, the thumbnails are going to be generated when you browse the directory. Probably what network filesystem you’re using. Alternately, maybe there is a maximum file size on previews? I know dolphin has that option.

I think that error is related to a missing dbus session but don’t quote me on that.

You will probably find it easier to use a system level service, but run it as your unprivileged user with User= and Group= directives. Once you get that working, there are various other parameters you can add to harden the service if you like.

This is a good reference for hardening: https://docs.arbitrary.ch/security/systemd.html

The arch wiki has a good general reference for all things systemd: https://wiki.archlinux.org/title/Systemd

I agree with the general idea of what you’re saying, but it’s a slippery slope.

Most people I know personally would never take the effort to learn anything past the point of “Ask ChatGPT” when they have a problem. What happens when the model is wrong, or simply cannot solve the problem? Or maybe they have no network connection and cannot run something suitable locally?

At that level of coddling, then they might not even have the ability to find and open a man page, or edit a config file without a GUI. And that’s a problem. It’s not even Linux-specific. I went to school with “smart” computer science students who don’t even understand file extensions or what a shortcut conceptually is.

What I’m getting at, is there needs to be some kind of balance, or people will just gradually become more useless.

My basic check is: Are there investors / vc people involved? If so, then it will inevitably enshittify. If not, then requires further investigation. OSI-approved open source is a big plus

Even when choosing what seems like good software, I think it’s important to consider switching costs. How easily can you move to another solution, say the second pick, if things go south?

I’ve been using Arch off and on for a long time, since it was horrible to install and updates did often break stuff. This is not the case now 🖖, and the Arch wiki is your friend.

1. Consider using btrfs with automated snapshots using yabsnap. It includes a configurable pacman hook in case something goes awry. Also just nice to have snapshots in case you accidentally delete a file or something.

2. Use paru, an AUR helper. Good for random things which may not be officially packaged. Expect to run into failures, and learn to diagnose them. Sometimes it’s just a new dependency the packager missed. For both paru and pacman, clean the cache once in a while or automatically, or things will get out of hand.

3. Do the “manual” setup, at least the first time, so you have an idea what’s going on. Don’t forget to install essential stuff like iwd (if needed) when you do pacstrap, or else you might have to boot from live again to fix it. Once you’re done, take care to follow the important post install steps, like setting up a user with sudo, a firewall, sshd, etc.

As for general setup, I’ve recently embraced systemd-networkd and systemd-resolved. Might be worth giving it a shot, since there is no default network manager like application. You can even convert all your wireguard client configs into networkd interfaces.

Best practice: Keep a personal log of various tweaks and things you’ve configured, and set up automated backups (more of general guidance).

Have fun!

Some interesting discussions there, looks like they won’t be hard up finding sponsors or places to move. Nice to see.

I would use the native version. For something like this, it makes sense that it should have less restricted/sandboxed access to the underlying system.