With WSL you can use “Linux the good parts” (command line tools, efficient-enough paradigms for fork() servers) and completely avoid X Windows, the Wayland death spiral, 100 revisions of Gnome and KDE that not so much reinvent the wheel but instead show us why the wheel is not square or triangular…

It's all opinion of course, but IMO Windows is the most clumsy and unintuitive desktop experience out there. We're all just used to the jank upon jank that we think it's intuitive.

KDE is much more cohesive, stable, and has significantly more features.

>the Wayland death spiral

That sounds like Wayland getting worse, but it's actually been slowly improving and it's pretty good now. Only took a decade+ to get there.

Mir was good from year one.

Judging from what happened to X11, that means wayland will be deprecated very soon. /s

/s indeed because there are actually no plans at all to replace Wayland!

I think the infamous cascade of attention-deficit teenagers (CADT) has slowed down quite a bit in the desktop space because... well, most developers there are over 30 now.

Not unlike Win10 vs 11.

It blows my mind that people can complain about the direction KDE is going when trying to paint a picture about how it's so much nicer to use Windows. I know the boiling frog experiment is fake, but just checking: are you sure the water isn't getting a little uncomfortably warm in the Windows pool right now?

I know you're saying you don't have to use it, but for any that didn't know, WSL2 does ship with it's own Wayland. And it does have some weird bugs.

After having used i3 and Sway, Windows is surprisingly bad at handling windows for an OS called Windows.

It requires a bit of work to setup to your liking of course, but hey, at least you have an option to set it up to your liking

Agreed. I used tiling WMs for a long while (ion3, XMonad) and it was such a productivity boost.

Then I was forced to use a Mac for work, so I was using a floating WM again. On my personal machine, ion3 went away and I never fully got around to migrate to i3.

By the time I got enough free time to really work on my personal setup, it had accumulated two huge monitors and was a different machine. I found I was pretty happy just scattering windows around everywhere. Especially with a trackball's cursor throw. This was pretty surprising to me at first.

Anyway this is just my little personal anecdote. If I go back to a Linux install I'll definitely have to check out i3 again. Thanks for reminding me :)

Compiling and testing cross-platform software for Linux lately (Ubuntu and similar)... You can't even launch an application or script without CLI. Bad UX, IMO. For these decisions, There are always reasons, a justification, something about security. I don't buy it.

> You can't even launch an application or script without CLI.

Care to elaborate? I'm not sure I understand what you're saying here.

I compile my program using WSL, or Linux native. It won't launch; not an executable. So, into the CLI: chmod +x. Ok. It's a compiled binary program, so semantically I don't see the purpose of this. Probably another use case bleeding into this. (I think there's a GUI way too). Still can't double click it. Nothing to launch from the right-click menu. After doing some research, it appears you used to be able to do it (Ubuntu/Gnome[?]), but it was removed at some point. Can launch from CLI.

I make a .desktop file and shell script to move it to the right place. Double click the shell file. It opens a text editor. Search the right click menu; still no way. To the CLI we go; chmod +x, and launch if from the CLI. Then after adding the Desktop icon, I can launch it.

On windows, you just double click the identified-through-file-extension executable file. This, like most things in Linux, implies the UX is designed for workflows I don't use as a PC user. Likely servers?

This sounds very weird to me. Any sane build toolchain should produce a runnable executable that already has +x. What did you use to compile it?

Removing double-click to run an executable binary certainly sounds like something either Gnome or Ubuntu would do, but thankfully that's not the only option in town. In KDE I believe the same exact Windows workflow would just work.

>Any sane build toolchain should produce a runnable executable that already has +x. What did you use to compile it?`

`cargo build --release`

Good to know KDE doesn't do that!

Even stranger then. Just to make sure I'm not missing something, I just tried this on my Mac:

  $ cargo --version
  cargo 1.86.0

  $ cargo new hello-rs
     Creating binary (application) `hello-rs` package

  $ cd hello-rs && cargo build --release
     Compiling hello-rs v0.1.0 (/Users/int19h/src/hello-rs)
      Finished `release` profile [optimized] target(s) in 0.73s

  $ ls -la target/release/hello-rs
  -rwxr-xr-x@ 1 int19h  staff  468608 May 20 20:16 target/release/hello-rs*

  $ ./target/release/hello-rs
  Hello, world!

Are you sure it's not because the package in question does some kind of weird custom build steps?

Might have got lost in translation when I moved it from WSL to a windows-made zip file. I think that workflow nukes permissions.

Yeah the typical way programs are run is by using a .desktop file that's installed. The reason nobody cares is because running random executable that have a GUI is a pretty rare use case for Linux desktops. We don't have wizards or .msi installers, we just install using the package manager. And then it shows up where it needs to.

If you're on KDE, you can right-click the start menu and add the application. Also, right-click menu should give you a run option.