Why wayland anyways?

X11 is obsolete and wayland is the only other game in town.

I started off in a purely cli computing environment. The earliest gui on unix for me was probably solaris 6 and openwindows which was x11 based, it also had display postscript capabilities. Though I first ran linux in '91 it wasn’t until slackware with x11 came along that I used it as my primary computer at home. I am no expert on x11 deficiencies but I have observed that it has not evolved in a very long time which is not good.

When I set up three computers to run plan 9 I was intregued by the adherence to the fundamental unix concept that device use was reading and writing files even for the 8 1/2 windowing system.

Later the berlin windowing system came along which sounded interesting though as I recall did not attempt to support any x11 compatibility.

Nowadays having very low latency with large pixel counts, fast refresh rates and 3d rendering is desirable. Security is paramount. Muli-user on a single gpu (sriov per vm, container, …) is also a want. Remote display cannot be ignored. Abandoning x11 compatibility is okay with me.

Wayland, especially with weston and purely wayland/weston apps, is lightweight and quick. I have yet to try out wlroots stuff. I know of no other linux gui project that has come as far as wayland.

What else is in wayland’s favor?

1 Like

It is available in Fedora, which leads systems my colleagues use (RHEL and Ubuntu) and has SElinux in default installs. My use case is making sure some scientific batch processing codes will continue working on linux: replacing libraries that are no longer maintained, porting to current R/Python versions, etc.

As you summarised wayland’s features are very desirable in a modern graphical environment that X11 could not provide for technically.

There are many developers willing to work on wayland.
There is next to none that wish to work on X11.

Here is a project that obviates an advantage towards wayland:
introducing graphics offload

Not quite yet, but going that way

1 Like

What else is in wayland’s favor?

It’s also been designed for modern modularity to suit future needs. Because X.org has a dated design that doesn’t take that into account, extending its functionality is much like adding duct tape on top of duct tape, rather than fixing the stuff under the duct tape. In many ways, Wayland is a necessary replacement for X.org.

2 Likes

I personally do tend to prefer Xorg over Wayland, but obviously I won’t use it anymore if Xorg is deprecated and Wayland is all that’s left. I’m not going to become the Linux equivalent of those people trying to keep Windows XP alive, I know when to quit.

However, I will share some things I think need to be hammered out in the transition to Wayland. The main one is that, in my opinion, the Wayland protocol is a bit too opinionated and puts its ideals of how things should work over how people actually expect them to work. So getting rid of all the legacy X11 cruft that people don’t use at all? That’s fair. But refusing to permit X11-like behavior that people actually ARE using, solely because it doesn’t fit into some kind of vision for what Wayland should be according to idealistic developers? That’s a big problem.

In some cases, this behavior is allowed by compositors on Windows and Mac, but Wayland is putting the binders on, being restrictive, and requiring a totally different approach to thinking about things.

Do I get why they are wanting to put their foot down and not allow any of this stuff into their shiny new protocol? Yes, absolutely. But if they want Wayland to work for everyone and replace Xorg completely, they need to start backing off on their ideals a little and make a few concessions to the kind of people who are still using Xorg because Wayland doesn’t meet their needs.

I don’t expect Wayland to be a drop-in replacement for Xorg, but I do expect Wayland to support at least the same kind of features as the compositors on Windows and Mac, and if they refuse to allow for features other platforms have that some developers need because of their values and opinions, then I think there’s a problem that needs to be acknowledged.

I’m talking about things like application windows knowing absolutely nothing about the displays and position of nearby windows. Yes, that might be a security risk, but it’s also needed for interoperability in some environments. To make my point a bit better, allowing users to flash their own BIOS via software rather than have it be a mask ROM that has to be physically replaced by the user to be updated is also a security risk, but at some point you have to make concessions for the sake of convenience and productivity that you might not want to make if security were the only consideration.

I think Wayland simply doesn’t seem to get that yet, because X11 was always there as a fallback, but if they really expect the X11 holdouts to adopt Wayland, they are going to have to address the kind of concerns those people have that they have, up to now, not wanted to take into consideration, and could afford not to take into consideration because Xorg was the legacy option. Well, now Xorg isn’t going to be there as a legacy option, and those users still need something Wayland isn’t offering. So something has to give, IMO.

It’s not reasonable for the Wayland developers to assume every single application on every single environment is going to be able to constrain itself to work within the limitations of the Wayland protocol. That is to say, there needs to be a way for people to choose between the default behavior of Wayland that seems to be high security, focused on process isolation, and geared towards mobile devices and touchscreen kiosks, and something a bit more relaxed that is likely to be fine for desktop environments.

As it stands, something that is well-tuned for people with keyboards, mice and multiple monitors is being replaced with something that is primarily for laptop users and people using touchscreens, that kind of puts the needs of desktop users in the backseat. I really hope that changes before Xorg is completely abandoned.

I think this kind of rant may be well intentioned but is a bit over the top here.

Possibly take that to a format (forum, mailing list, etc. ) that is specifically wayland focused.

Yes, X11 will eventually fade away.
No, Fedora does not have a timeline available for when that might happen and wayland is continually being improved to better meet users expectations.

We, on this forum, are primarily focused on fedora itself. What is happening on wayland is managed by the upstream developers so specific user needs, wants, & complaints should be directed to the developers, not the downstream users.

1 Like

Sorry for the earful if you guys are not Wayland developers, I just may have been… a little misinformed about just how deep Fedora’s involvement with the Wayland project is, and also about the timeline for Xorg being discontinued (like I thought it was going away at the end of this year). Let’s just say I probably need to learn not to believe everything I read.