[firebaze]

As "just a user" in terms of using Linux as a DE, I personally had no gains, just losses.

For example, I wanted to create a tool to use my old smartphone as a trackpad. After a short time I realized my approach won't work on wayland. I started delving in the available options, /dev/uinput, libevdev etc., and realized it will be to much pain. No (generic) accessibility¹ support means no generic way to do that, and just for my setup it is too much work.

So I dropped that. Then I have a set of tools I use for quite some time, amongst which are AnyDesk, TeamViewer, Blue Recorder, Steam. These either don't work, work only with wlroots, or work via the X11 shim.

¹ for me just a nuisance, for all the people who really need accessibility support, well ...

[dijit]

> I personally had no gains, just losses.

Depends.

Screen tearing gets worse at higher refresh rates and resolutions under Xorg.

Additionally fractional scaling and mixed DPI's are something Xorg was having great difficulty supporting.

It also gets quite slow (high CPU and especially memory transfers) at high refresh rates and resolutions; I mean, you benefit from this without noticing.

You'll notice a loss in functionality (like screen recording) much more than you'll see a benefit like "this will actually work at 8k 120Hz"; but it's not true that there's no gains.

Wayland is fine, it fixes a lot of legacy that holds the ecosystem back, people don't like it overwhelmingly because changing esoteric and hacky software that is core to basically everything will be difficult and take time.

It might shock you to know though that GNOME and Fedora users are already using Wayland on the popular distros, the transition was completely seamless for those users.

[firebaze]

You're right and I don't debate any of your points. And no it doesn't shock me that big distros use Wayland already.

On the other hand, my points still stand I think. And regarding esoteric and hacky software: I don't think accessibility as a base functionality should require hacky and esoteric stuff, it should be a first class citizen in a 2023 window manager (?).

And I also want to remind how Linux (and, ironically, Microsoft) grew so strong: they are really, really careful with deprecating old stuff. As hacky that stuff may be. Up for debate if this is net good, sure.

[kaba0]

> it should be a first class citizen in a 2023 window manager

It should be, but linux is built by hobbyists in their free time — it is hard enough to get to a working state, let alone adding accessibility, even if it is the first on the list.

[shrubble]

Screen tearing is a function of the device driver, it's not inherent to X11.

[shrimp_emoji]

> Screen tearing gets worse at higher refresh rates and resolutions under Xorg.

No, it gets better, the same as aliasing does with higher resolution. That, and VRR hardware in the monitor, are the correct way to fix tearing in this day and age. V-sync is kludgey false dharma.

> It might shock you to know though that GNOME and Fedora users are already using Wayland on the popular distros, the transition was completely seamless for those users.

Sure, but it might shock you how much smoother your desktop can be if you disable compositing (on X, anyway). Many never have, and you can't in GNOME (of course). ;D I'm not sure if Wayland supports that in itself, however.

[PlutoIsAPlanet]

You'll probably going to ned to look at using libei

https://gitlab.freedesktop.org/libinput/libei

[firebaze]

Thanks, looks interesting!

[WhyNotHugo]

If you want to create a tool to use your smartphone as a trackpad, this is the protocol that you're looking for: https://wayland.app/protocols/wlr-virtual-pointer-unstable-v...

It's only a wlroots protocol for now, and I won't deny that this is definitely one of the fields that still need some more refinement/stabilisation.