When Windows Vista was introduced with DirectX 10 being Vista only and not down ported, with a new driver model, there was this rage that everyone would be going into Linux Desktop.
This keeps repeating every time a new Windows version comes out.
Instead, people pick Apple and Google devices, and even if ChromeOS and Android have a Linux kernel underneath, that is pretty much irrelevant for the ecosystem of JavaScript, Java, Kotlin and C++ applications.
My beef with people moving to Linux because they are pissed at Windows is that often, they don't want Linux. Or said differently, they want Linux to look and feel like Windows, and push for that.
I have seen many developers who are on Ubuntu apparently because it's free, but they behave like if it was Windows. They don't understand that Linux is not just Ubuntu and that they should not assume that everybody is running exactly their version of Ubuntu.
IMO if they want something that looks and feel like Windows, they should stay on Windows.
I have been using Wayland everyday for a year now. I don't get why people complain about it: it's great! I had some setup to do at the beginning (I am using Sway), and I haven't had any problem since then.
Edit: actually, to be very honest, it happened twice in the year that it somehow did not come back from suspend and I had to kill and restart it (but I can't even say if the bug was in Wayland, or maybe Sway, or something else).
> I don't get why people complain about it: it's great!
That's what's insidious about Wayland. If it gave everyone a bunch of little paper cuts, it wouldn't be nearly as polarizing. Instead, it's great for the common use cases but unusably bad for the uncommon ones, so people who don't have any of the uncommon ones end up thinking that the people who do are exaggerating the problems.
Any individual uncommon use case is uncommon, but there's a lot of uncommon use cases. I'd be very surprised if fewer than 20% of Linux users had at least one uncommon use case that X handles perfectly but that Wayland can't handle at all.
Here's my problem: I don't want tiling windows. I want title bars. I don't want to run KDE or GNOME. I want desktop computing like it was in 1994.
In the X11 universe, there's a galaxy of decent full featured window managers that fit my needs. Yes, some of them are archaic, but they work.
On Wayland, I don't have that option. There are:
* Experimental projects which went inactive several years ago. Since the compositor has so many more moving parts than an X11 Window Manager, I'm not sure how viable these are.
* Weston, which is extremely limited in features, so I'm not sure all Wayland tools are happy with it
* In-process wlroots projects like labwc and wayfire that sort of work, but often feel surprisingly incomplete. (labwc's theming is weak, Wayfire doesn't know what window shading is)
If there was something that felt complete and workable-- a FVWM or a WindowMaker, I wouldn't feel like I'm giving up quality-of-experience for the sake of Waylandness.
But then it's not fundamentally Wayland, right? Like there is still a chance that Wayland will end up having what you prefer (assuming that the people who like it on X11 port it to Wayland).
I am not an expert, but it feels like X11 has fundamental security issues that Wayland addresses. So to me it makes sense that Wayland exists. Then I would happily contribute to whatever I need in Wayland, but honestly it all exists already: I used i3wm, I'm now on Sway, and all the tools I need do exist.
It feels like the particular corner of the universe I reside on did not migrate to Wayland well. Perhaps it would less bother me if I didn't feel like I was being forced to accept KDE or GNOME or deal with a second-best experience.
Having just set up a new machine the last few days (Debian Sid on a new Thinkpad L13) and spending a little while in a KDE/Wayland session, I remembered what I don't like about the big desktops: since they don't completely own the distribution underneath, there's always going to be a control panel that doesn't work as expected, or some incompatible software in there. The illusion isn't quite strong enough.
My best theory is that the lack of compositor options is a timing issue: Wayland arrived at a time when most desktop Linux was moving towards the big desktops, and away from the classic standalone window managers. I suspect they were more or less a solved problem; FVWM in 2024 looks and works very much like FVWM in 2004, so there simply was less interest in them in general, let alone to bankroll a huge production number like making a Wayland-compatible rewrite. Tiling window management seems like a newish trend that acquired its own fanbase late enough that there was momentum to build things like sway.
I'm surprised there isn't a "reverse xwayland" -- a paper-thin compositor that exposes Wayland top-level windows in a manner that twm and friends know how to wrangle. I know I can launch a Wayland compositor with its own captured desktop but that defeats the purpose of wanting something else to own the frames, decoration, and movement.
> FVWM in 2024 looks and works very much like FVWM in 2004
Hopefully there will be a community effort to rewrite FVWM in Wayland at some point! It doesn't seem like sway (or i3wm for that matter) required years of development by a big team.
I couldn't contribute to that effort because I am very much into tiling windows managers. I actually wish I could contribute to Sway, but it just seems to do everything I need. I thought I had found something I needed and implemented it only to realize that it already existed in a better way :-).
20 comments
[ 5.9 ms ] story [ 83.6 ms ] threadI'm glad mint is dragging it's feet on Wayland.
As usual, Microsoft is shooting itself in the foot as much as possible with the shenanigans going on in Windows 11.
Luckily for Microsoft, Weyland is here to basically block any sort of hope of a good desktop UI for at least another 15 years on Linux.
I understand what the goals of the project were, The practical effects of Wayland have been to paralyze progress on viable Linux desktop.
This keeps repeating every time a new Windows version comes out.
Instead, people pick Apple and Google devices, and even if ChromeOS and Android have a Linux kernel underneath, that is pretty much irrelevant for the ecosystem of JavaScript, Java, Kotlin and C++ applications.
Or they just stay on Windows no matter what.
I have seen many developers who are on Ubuntu apparently because it's free, but they behave like if it was Windows. They don't understand that Linux is not just Ubuntu and that they should not assume that everybody is running exactly their version of Ubuntu.
IMO if they want something that looks and feel like Windows, they should stay on Windows.
Edit: actually, to be very honest, it happened twice in the year that it somehow did not come back from suspend and I had to kill and restart it (but I can't even say if the bug was in Wayland, or maybe Sway, or something else).
That's what's insidious about Wayland. If it gave everyone a bunch of little paper cuts, it wouldn't be nearly as polarizing. Instead, it's great for the common use cases but unusably bad for the uncommon ones, so people who don't have any of the uncommon ones end up thinking that the people who do are exaggerating the problems.
In the X11 universe, there's a galaxy of decent full featured window managers that fit my needs. Yes, some of them are archaic, but they work.
On Wayland, I don't have that option. There are:
* Experimental projects which went inactive several years ago. Since the compositor has so many more moving parts than an X11 Window Manager, I'm not sure how viable these are.
* Weston, which is extremely limited in features, so I'm not sure all Wayland tools are happy with it
* In-process wlroots projects like labwc and wayfire that sort of work, but often feel surprisingly incomplete. (labwc's theming is weak, Wayfire doesn't know what window shading is)
If there was something that felt complete and workable-- a FVWM or a WindowMaker, I wouldn't feel like I'm giving up quality-of-experience for the sake of Waylandness.
Theming seemed fine? I picked up someone's stylesheet off GitHub and that was it.
I am not an expert, but it feels like X11 has fundamental security issues that Wayland addresses. So to me it makes sense that Wayland exists. Then I would happily contribute to whatever I need in Wayland, but honestly it all exists already: I used i3wm, I'm now on Sway, and all the tools I need do exist.
Having just set up a new machine the last few days (Debian Sid on a new Thinkpad L13) and spending a little while in a KDE/Wayland session, I remembered what I don't like about the big desktops: since they don't completely own the distribution underneath, there's always going to be a control panel that doesn't work as expected, or some incompatible software in there. The illusion isn't quite strong enough.
My best theory is that the lack of compositor options is a timing issue: Wayland arrived at a time when most desktop Linux was moving towards the big desktops, and away from the classic standalone window managers. I suspect they were more or less a solved problem; FVWM in 2024 looks and works very much like FVWM in 2004, so there simply was less interest in them in general, let alone to bankroll a huge production number like making a Wayland-compatible rewrite. Tiling window management seems like a newish trend that acquired its own fanbase late enough that there was momentum to build things like sway.
I'm surprised there isn't a "reverse xwayland" -- a paper-thin compositor that exposes Wayland top-level windows in a manner that twm and friends know how to wrangle. I know I can launch a Wayland compositor with its own captured desktop but that defeats the purpose of wanting something else to own the frames, decoration, and movement.
Hopefully there will be a community effort to rewrite FVWM in Wayland at some point! It doesn't seem like sway (or i3wm for that matter) required years of development by a big team.
I couldn't contribute to that effort because I am very much into tiling windows managers. I actually wish I could contribute to Sway, but it just seems to do everything I need. I thought I had found something I needed and implemented it only to realize that it already existed in a better way :-).
Also, ChromeOS is switching or has switched to Wayland. (They used to use their own homegrown solution.)