I understand sudo needs a password,but all the other stuff I just want off.

Sudo doesn’t need a password, in fact I have it configured not to on the computers that don’t leave the house. To do this open /etc/sudoers file (or some file inside /etc/sudoers.d/) and add a line like:

nibodhika ALL=(ALL:ALL) NOPASSWD:ALL

You probably already have a similar one, either for your user or for a certain group (usually wheel), just need to add the NOPASSWD part.

As for the other parts you can configure the computer to not lock the screen (just turn it off) and for updates it depends on distro/DE but having passwordless sudo allows you to update via the terminal without password (although it should be possible to configure the GUI to work passwordless too)

I understand sudo needs a password

You can configure sudo to not need a password for certain commands. Unfortunately the syntax and documentation for that is not easily readable. Doas which can be installed and used along side sudo is easier.

For software updates you can go for unattended-upgrades though if you turn off your computer when it is upgrading software you may have to fix the broken pieces.

These are all valid reasons to request a password 🤔

• Wake computer from Screensaver? Password.

Check your screen saver settings. Dunno which desktop environment you’re using. KDE should allow you to not enter a password for this.

• Install something? Password.
• Updates (biggest one. Updates should in my opinion just work without, because being up to date is important for security reasons)? Password.

Installing stuff runs sudo in the background hence the password prompt. Updates = installing stuff. Look up “passwordless sudo”. At this point, when do you even want a password to be shown? If you don’t need a password, get rid of it entirely.

You can configure this behavior for CLI, and by proxy could run GUI programs that require elevation through the CLI:

https://wiki.archlinux.org/title/Sudo#Using_visudo

Defaults passwd_timeout=0(avoids long running process/updates to timeout waiting for sudo password)

Defaults timestamp_type=global (This makes password typing and it’s expiry valid for ALL terminals, so you don’t need to type sudo’s password for everything you open after)

Defaults timestamp_timeout=10(change to any amount of minutes you wish)

The last one may be the difference between having to type the password every 5 minutes versus 1-2 times a day. Make sure you take security implications into account.

Not sure what that is. Plesse explain more.

I like it as a concept, but it gets bothersome to maintain on the long run, sometimes you just want to install something not write configs.

I think Gentoo has a nice middle ground, where you can install packages as a one-off without adding them to the world file, which makes it very meat to maintain both your regular packages and some random things you’re trying out before settling in on adding them permanently.

That being said I’m currently looking into writing some ansible for kick-starting machines, so I’m very much moving in that direction. Why not use nix then? Few reasons:

• Using Nix means I’m forced to use Nix, whereas with Ansible I can use whichever distro I want, more than one even.
• I don’t want to have to define EVERYTHING, I want to be able to bootstrap systems quickly, but after the initialization I want to be able to mold each system to what I need without worrying about making it reproducible.
• Nix uses a language that’s only usable in Nix, in short I would need to study and learn something that’s only usable on one specific distro.

The rtl 8125 works

There’s a standard. /usr was often a different partition.

/bin - system binaries
/sbin - system binaries that need superuser privileges
/usr/bin - Normal binaries
/usr/sbin - normal binaries that require superuser privileges
/usr/local/bin - for executables that aren't 'packaged' - i.e., installed by you or some other program system-wide

I typically search the package name + fedora, it will probably tell me the alternative package that is in fedora.

Nowadays, I have moved to an atomic fedora distro, so I would severely limit the amount of package I install on my system for stability and security.

I think I only have two packages installed on my machine: fish, because it is the only popular shell that follows xdg dir; and a latex input method to use in slack.

Back in my slackware days I’d just convert other distros packages to the tgz format or compile the package and its requirements.

If the dependencies were really complex I’d draw a picture to help me understand them better.

Because it requires a very specific framework to be built from the ground up, and FDO doesn’t specify these. A lot of breakage would happen if were to shoehorn such changes into Linux suddenly. Android has many layers of security that they’re fundamentally different than that of the unix philosophy. That’s why Android, even if it’s based on Linux, it’s not really considered “a distro”.

Android apps are sandboxed by default while packages on Linux run with the users permission.

There is already something like this with Flatpak since it also sandboxes every installed program and only grants requested permissions.

It is technically doable, but that would require a unified method to call when an app needs camera, and that method will give show the prompt.

This would technically require developers to rewrite their apps on linux, which is not happening anytime soon.

Fortunately, pipwire is currently doing this work, like when you screen share on zoom using pipwire, a system prompt will pop up asking you for what app to share. Unlike on Windows, zoom cannot see your active windows when using this method, only the one that you choose to share.

There is a lot of work in improving security and usablity of linux sandbox, and it is already much better than Windows (maybe also better than macos?). I am confident, in 5 years, linux sandbox stack (flatpak, protal, pipewire) will be as secure and usable as on android and ios.

Doesn’t feel like that to me. I’ll need to see evidence that that is the main reason. It could be but I just don’t see it.

Linux users are often very passionate about the software they put on their computers, so they tend to argue about it. I think the customization and choices scares off a lot of beginners, I think the main reason is lack of compatibility with Windows software out of the box. People generally want to use software they are used to.

There’s no infighting. It just feels that way because you picked an inferior distribution.

Linux generally has a higher (perceived?) technical barrier to entry so people who opt to go that route often have strong opinions on exactly what they want from it. Not to mention that technical discussions in general are often centered around decided what the “right” way to do a thing is. That said regardless of how the opinions are stated, options aren’t a bad thing.

Because you don’t have an in person user group and only interact online where the same person calling all mandrake users fetal alcohol syndrome babies doesn’t turn around and help those exact people figure out their smb.conf or trade sopranos episodes with them at the lan party.

I use xorg on my desktop, because nvidia card don’t have good support for my applications on wayland. Specifically, typing in electron app will jitter. I will be switching to wayland if this problem is solved.

I use wayland on my laptop because it is more secure and supports one-to-one gesture, which is crucial for trackpad.