Wayland is getting closer

[dreamer]

Someone PMed me about Wayland. I wrote this and might as well post it here. It's opinions but I tried to stick to facts to the best of my knowledge. I see Wayland as another nail in the coffin for the Linux desktop, but I'm sure there are more positive forum members that might want to comment. The main problem is that X.org users will find themselves in the same group of "legacy users" that SysV users already belong to.

I think the Wayland idea is bad (but that's just a personal opinion). However some of the things that have been pointed out by other people include:

1. Wayland is not an evolution of X.org. It's a completely new paradigm so a lot of software has to be rewritten. It's not evolution, it's revolution.
2. It's not possible to run GUI apps as root. You won't be able to open Thunar as root for example.
3. Apps can't communicate with each other by default. Remote desktop and screen sharing, screen recording and even screenshots, color pickers and night mode apps all need to access custom APIs so old tools won't work anymore.
4. Wayland is just a protocol, not a display server. It's up to each desktop environment (toolkit and compositor) to do a lot of the work that was previously done by X.org. Gnome and KDE run on Wayland at this point. Smaller desktops don't seem to have the manpower to adapt to Wayland.
5. Theming possibilities will be more limited and Gnome expects each app to theme itself including the window border (no universal standard for client side decorations - CSD - and the concept of a window manager is alien to Wayland).
6. It will be harder to create cross desktop applications. Wayland is a core protocol - a lot is undefined and left to toolkits and compositors to implement. Fragmentation? Maybe easier to just use XWayland instead?
7. Video drivers need to be rewritten and if I understand correctly software has to target Wayland or use XWayland compability layer. Basically point number one again. There is a saying "If it ain't broke, don't fix it" and also this story about Netscape comes to my mind:
https://www.joelonsoftware.com/2000/04/ ... do-part-i/


You should take this with a grain of salt, but I think it's fairly accurate. Running Xfce on the BSD kernel is a smaller change (from a GUI user perspective) than running Xfce on Wayland (which isn't possible at this time). There are positive things about Wayland like support for different DPI settings on different displays and fractional HiDPI support for GTK.

Wayland is more "modern" and has a focus on security, but there is a huge price to pay "when you turn everything upside down". I would have preferred an evolution of X.org. There is a project called MicroXwin that tries to modernize X.org, but it's just a hobby project without big funding. I would have preferred that compared to a complete paradigm shift like Wayland. It seems MicroXwin is also capable of running Android apps.

https://volkspc.org/MicroXwin/

My point is "don't throw out the baby with the bathwater". Maybe look at Windows and macOS and use/evolve what you already have because evolution is better than revolution 9 times out of 10. Just my point of view as a casual user.

Of course it's too late for that. Wayland is (already) here and is going to pull Linux desktops/distros/applications in that direction.

[Jan K.]

As long as it's not "hard-coded" in the kernel, isn't it a bit like the "to use systemd - or not" debate? meaning you can always build a Linux to taste.

And even if "hard-coded", the kernel can always be compiled to taste as well thanks to it's modular structure?

https://wayland.freedesktop.org/architecture.html

Be aware, I'm only a Linux student atm.!

[Artim]

"Tried, tested, and proven" is the way to go - until it's not. I wonder if the "revolution" isn't a little bit of some people trying to make a name for themselves by coming up with something totally new. But like the article in the OP said, it's not as if code rusts and decays and has to be replaced just because it's "old."

[manyroads]

Without taking the OP up point by point, I'd say many of the disadvantages noted in the OP might in 'fact' be design constraints for security and stability purposes. An additional group of points seem to indicate a typical issues around legacy maintenance, which my nearly 50 years of software experience say is not only a pain, but also largely undesirable. Wayland is still a fair distance in the future but many issues are known to need 'help'. Given Wayland is Open Source, I think it will roll out like many such projects, that is to say slowly. After all it is already over 11 years old. I think given Wayland's lengthy history, we can assume serious problems will be addressed and the roll-out will be incremental. I personally see little reason for serious concern/ worry.

Wikipedia has a pretty interesting overview: https://www.wikiwand.com/en/Wayland_(di ... _protocol)