Fair…

PS: Wait… that’s a hobby and they don’t get paid for lying? That’s even worse than I thought.

You could…

But then one is an open system where you can disable the UI put on top and have a working linux system, while the other is a closed blob destroying compatibility and trying hard to lock you out from accessing the underlying linux system.

It has lots of small issues that add up to a frustrating experience for mainstream users.

And 90%⁽¹⁾ those are out of Linux’ actual resposibility because they are caused by third parties screwing up… sometimes even intentional (from companies producing lackluster drivers only having a fix cobbled together for Windows specifically -looking at Realtek networking for example- to ones actually going out of their way to block Linux (MS FUD included…).

⁽¹⁾ The other 10% exist on Windows or Mac also, but people just accept them because they are used to not having a chance to change it. Seriously the amount of obscure regedits or third party tools usually surpass the number of linux issues fixed by editing an easy to read txt file.

Linux is Linux.

We should send all those people, pages and guides suggesting distros to hell.

And then instead we suggest update-schemes (fixed, rolling, slow-roll), package managers and Desktop environments. People with enough brain cells to start a computer are then absolutely able to chose a distro fitting them based on that. Everything else coming with a distro is just themeing/branding anyway…

(and just for the use statistic: Archlinux, Opensuse (Leap and Kalpa), Debian here…)

I’ve been using Arch and Manjaro for couple years each and in my experience they both break regularly. But, for some weird reason, Arch Linux is praised, when Manjaro is shamed upon.

No, there is not some weird reason but actual very good ones.

Things can break on a bleeding edge update scheme. That’s to be expected from time to time. But the questions are “why did it break” and “what is done to fix it”.

If something breaks on Archlinux it’s because of some new package with a issue that escaped testing. Then the fix come out as fast as possible (often within minutes even, but let’s assume hours as those things need to move through mirrors first…).

If something breaks on Manjaro it’s either because of the exact same reason as above, but 2 weeks later. Because Manjaro keeps back updates for two weeks “for stability reasons”, yet doesn’t do anything in those 2 weeks. So they just add the same problem later, completely defeating the argumant about stability. Oh, and fixes are of course kept back for 2 weeks, too, because… reasons.

Or it breaks because they fucked up their internal QA. For example by letting their certificates expire again and again and again and again… of by screwing up their very own pacman-wrapper and then ddos’ing the AUR for all users, not only Manjaro ones.

Or -speaking about the AUR- it breaks because they give their users full access to the Arch User Repository (without any warnings about user content being less reliable and used at your own risk) pre-installed. Also they do it on a system generally out-of-date because it lags 2 weeks behind. Which is not what AUR packages are build for (they assume up-to-date systems) and is a straight path to dependency hell and breakings… not because something went wrong but because the whole concept of an out-of-date system not running their own also 2-weeks behind version onf the AUR is idiotic. On the “plus” side they have an easy fix: blame the user, because he should obviously know that an pre-installed part of Manjaro is conceptionally flawed and shouldn’t be trusted.

Right decision but for the wrong reason.

There is a difference here.

Unlocking home later in the boot process is not a problem, so the you can indeed have a keyfile on your root and get your home unlocked and mounted after root is done.

Swap however needs to be available early, at least if you want to use it for hibernation.

This would -at least as far as I understand it- limit your swap’s functionality for hibernation etc. Because there your swap needs to be available early. You can still do it in theory, but the key file then would need to be included in you initrams, which kind of defeats the purpose.

There is however a much more easier option: either use LVM on luks (so the volume is decrypted with the password and then contains both, root and swap) or just use the same password for root and swap while switching over to the systemd hooks (as those encryption hooks try unlokcing everything with the first provided password by default, and only ask for additional password if this fails).

EDIT: Seeing that you crossposted this from an archlinux-specific community: You can find the guide here. It’s for using a fully enrcypted system with grub as bootloader, but the details (in 8.3 and 8.4) are true for all boot methods. Replace the busybox hooks with their systemd equivalents (in minitcpio.conf for archlinux but again this isn’t limited to that init system), then add “rd.luks.name=<your swap’s uuid=swap” to your kernel parameters and also replace the “cryptdevice=UUID=<your root’s uuid>:root” that should already be there for an encrypted system (that’s the syntax for the busybox hook) with “rd.luks.name=<your root’s uuid>=root”. On startup you will be asked for your password as usual, but then both root and swap will be decrypted with it (PS: the sd-encrypt hook only tries this once… so if you screw up and misstype your password on the first try, you will then have to type it again two times, once for root, once for swap…)

And just lke with the war on drugs, countries will realize it’s a lost cause. And will then instead try to coopt the system to spread their own desinformation. If you can’t win, exploit it for your own gains…

Welcome to our wonderful post-factual age.

Those usage stats are a fantasy build by nicely asking your browser about your pc’s details. But the answer is complete fiction. And one people often intentionally set to display Windows because idiotic corporate-created webpages will refuse to work properly otherwise.

(I haven’t touched Windows in many years and still I would end up in those stats as a Windows user (and Chrome which is also wrong)…)

It’s basically all just marketing bullshit.

They are written but don’t replace something in the read-only OS. They are just overlayed, so once removed the original is still there. How they do it differs. There are actual overlay filesystems for the job, or some use btrfs where all subvolumes behave mostly like virtual partitions (and copies of a subvolume only take space for changes of the original).

From my personal testing experience I would say the concept is solid but the existing distros are not there yet, with some missing features, minimal documentation and several rough edges in their containerisation approach (as in: some features and things not working because the container wasn’t well adapted to the immutable OS yet).

An immutable OS is fixed and mounted non-writable. Every update you get, every program you install is handled on top of it via containers or filesystem overlays so the underlying OS is untouched. Basically the same concept you know from smartphones or other devices with a “reset to factory settings” function. No matter how hard you screw up your system, you can always reset to the base OS, either by granulary deactivating things installed on top, or by a reset to the working base OS.

The software is the problem if it’s produced with a corporate mentality of “ship first, fix later”.

Yes, time consuming. But it’s still working hardware, not a now useless, now unrepairable paper weight…

Which is the definiton of “bricked” although people nowadays start to use the term inflationary.

Good thing is the kind of people making decisions based on buzzword-bongo filled PR campaigns like Crowdstrike’s are already forcing their IT to use Windows anyway.

“One of the primary arguments for manually installing and configuring your system is that it will teach you a lot about linux and operating systems in general. While there is a small grain of truth to this, it’s incredibly misleading. […] A significant drawback of the “manual means advanced” slogan is that it creates a narrative that is precisely the opposite of the truth. Computing is about automation, and advanced users are those who write software or use existing software to make their computing lives more painless.”

Sure… and how exactly is one aquiring the “expertise to automate” when they never actually did it manually to know what to automate?