qpsLCV5
So, a lot of people have already mentioned that the arch wiki contains great info. What’s missing, IMO is this: Installing Arch as described on https://wiki.archlinux.org/title/Installation_guide will leave you with an EXTREMELY basic system - you just have a bare command line, with none of the tools you’d use daily for actually using your pc.
This is where the learning comes in - choosing the software you need on your system, and learning how it all interacts with each other. IMO, you can be an experienced sysadmin, and never really have to deal with the details of what’s going on during installation - it’s the applications on top that actually do the work, and that you need to configure and run. Sure, you’ll need to learn systemd and other components, but that all comes with use of the software you need, not necessarily the base system.
This is also why I strongly recommend having a second, working machine with a browser while installing Arch for the first time. A plain arch install does not come with the tools you’re used to to connect to wifi, or even wired networks. and without a working browser, it can be hard to figure out how to connect to the internet. First things i had to do when setting it up were searching for the proper network tools and then choosing between desktop environments and window managers. For learning I recommend a WM, as a full blown desktop environment like Gnome comes with a whole host of tools already, but with a WM you need to set things up yourself so you learn more. (I went with Sway, but if you have an nvidia GPU i cannot recommend it - it works but with many little issues.)
i’m thinking long term - sure, right now google knowing everything about me isn’t dangerous. but if a massive political slide to the right happens in countries that host services, suddenly all the saved data from many years ago can be used against me. and don’t fall for the “end to end encrypted” bullshit either - all these services can flip a switch and have your encryption keys instantly. (or, if its an open source app that ACTUALLY keeps keys on the device only, which is extremely rare, it’s one update away from happening, and you better read the whole diff every update and compile the app yourself.)
that’s why i choose to self host everything. yes there’s a risk of being hacked, or installing something malicious because i don’t read every diff on every update. but i feel more confortable with it being my own responsibility, and my services are also all on seperate virtual machines to hopefully isolate any breaches.
personally, i’d have pretty big benefits for my homelab if i could use my own ipv6 range for everything. having only a singe public IP is just very limiting.
sadly, my ISP does give out ipv6 for home networks, but i cannot connect to any of them from my mobile phone with the same carrier. so that’s fun. they talked about rolling out ipv6 on mobile networks years ago, but i guess it’ll take a few more…
my phone won’t even do “force stop” anymore… fairphone 5 running whatever os fairphone ships, and all force stop does is put the app in the background or whatever, if it has an issue the issue will still be there when opening it again.
i’d suggest starting by finding out what package in your distro actually decides where audio goes - mostly it is pulseaudio (older) or pipewire (newer).
depending on the details of how your distro and the dongle work, it could either be a simple “pactl set-default-sink <headset-name>”, or a more complicated set of udev rules or pipewire/wireplumber scripts.
note that distros using pipewire still often support a lot of pactl commands, so it may be worth looking at the simple option even when not using pulseaudio.
i think flatpak has done a lot to make this easier, but at the same time… i’ll admit i’m not a fan of it (mostly due to random issues).
the way i see it, more distros need something like arch linux’ AUR. if an application is reasonably easy to build, it really does not take much to get it into the AUR, from where there’s also a path towards inclusion in the official repos.
i don’t know too much about other distros, but arch really makes it amazingly easy to package software and publish everything needed for others to use it. i feel like linux needs more of this, not less - there’s a great writeup that puts why linux maintainers are important way better than i ever could:
https://web.archive.org/web/20230525163337/https://kmkeen.com/maintainers-matter/
if it’s free, you’re the product.
signal seems really good right now, with open source clients, but they already show that they’d like to keep the ecosystem locked down by not allowing 3rd party clients. at some point they will need a way to pay for their datacenters, and even if they claim the foundation or whatever is doing well, i can see the pestering for donations getting much worse in the future.
that said, threema is far from optimal too, im still waiting for matrix servers to become solid options. last time i wanted to set up synapse, the only captcha they supported was fucking Google captcha :|