Back

Home

The biased thing about Wayland

As much as I try to avoid any of the popular dramas within the Linux spheres, it pretty much is impossible to avoid the tragedy that is the discourse about X vs. Wayland, with more and more vocal (terminally) online users advocating for the obsolescence of X/X11/X.org.

The second part of my obligatory introduction paragraph already highlights the first issue I noticed after (involuntarily) reading some of those "debates": Both Wayland advocates and X defenders are almost entirely unable to grasp what each piece of software is doing BEYOND their subjective usecases. Vocal Wayland advocates demand the widespread abolishment of X11 and/or X.org, despite X11 being the reference implementation of the X window system. X11 is thus on par with Weston, wlroots and whatever KDE Plasma and GNOME directly integrated into their respective window managers.

Defenders of X fall into the very same pattern by arguing against the Wayland protocol itself, rather than the various references and practical implementations that are responsible for the vast majority of incompatibilities and bugs. They also largely argue in favour of the reference implementation X11, not the X protocol. This results in debates almost entirely being a petty fight over reference implementations that vastly differ from each other and thus are... well, entirely pointless from the start.

The sheer subjectivity (i.e. highly individual usecases and configs) both sides rely on extends to features that are questionable from a more objective standpoint but this matter is being exaggerated by the fact that both only focus on what "works" for them individually and thus each try to shove their purely individual preferences down everyone's throat. X users try to counter Wayland propaganda by pointing towards the network capabilities of X; Wayland users attempt to point towards the lack of root access of Wayland implementations. When those arguments fail, Wayland users point out that "Wayland doesn't cause screen tearing", which X users relativize by pointing to X's feature that mitigates screen tearing. Eventually, it all boils down to "Wayland is modern, X/X11/X.org is not and even former X/X11/X.org devs agree with that", whereas the other side settles with "but Wayland is nearly two decades old and still hasn't seen a mature adoption, thus it's bad".

An example case against the current state of Wayland: mandatory compositing

Due to all discussions being convoluted to extreme degrees, one argument I still have to come across is the forced compositing of all Wayland implementations. I still am trying to figure out which part of the Wayland ecosystem is explicitly demanding compositing; the Wayland protocol itself or the implementations (Weston, wlroots, KWin, Mutter) solely for the sake of understanding the inner workings and how to possibly circumvent this **if** X should truly become abandonware. All of my machines currently run a stacking window manager (mainly Openbox but I recently installed and configured Fluxbox as each system's backup session in case I either mess my Openbox or my Polybar configs up). Stacking window managers will never transition to Wayland due to stacking window managers being designed to **not** depend on compositing, which is de facto mandatory when implementing Wayland compatibility (whether it's technically mandatory or sheer peer pressure doesn't matter in this case).

While two of my machines can handle and do rely on compositing via picom, both of my laptops are far too weak and respond poorly to compositing. My (nearly ancient) Acer Aspire, which already was prone to overheating during its Windows 7 days, is particularly sensitive to compositing and this issue becomes especially apparent during web browsing and "larger" software upgrades such as kernel updates and software needing to recompile itself (in this case it largely affects Lagrange due to being the largest of the non-binary – no pun intended, this is strictly technical – packages I use). To keep this still-functional laptop at its optimal efficency, any OS I want to run on this machine must at least grant me the option to turn compositing off. This already disqualifies any distribution shipping with GNOME and/or KDE Plasma by default due to being the ones that currently are the closest to ending support for X. Popular distributions such as thoe from the Red Hat and Ubuntu families are out of the equation, as well. This leaves me with very niche choices such as BunsenLabs, ArchBang, Archcraft and antiX, to name a few that are (more or less) tailored towards old and low-spec hardware – or some DIY distribution like Arch (which I already am using) and Gentoo.

But what about XWayland?

Personally, I interpret the very existence of XWayland part of the very misunderstandings and resulting petty fights. It can be seen as devs and users alike attempting to push Wayland adoption for the sake of "modernity", thus countering their own arguments addressing the supposed improved security and the desire to abandon X. With XWayland nothing actually changes besides perhaps a reduction of screen tearing cases and... maybe better support for larger and more monitors. Those issues, however, are of pure cosmetic nature and thus don't justify a compatibility layer for an abstraction aiming to fully replace another abstraction layer.

Personally, I don't want to rely on three abstract layers – because a lot of X programs just run like ass on XWayland – just to run a bloody stacking window manager because one abstraction layer is opinionated against it. And I already can tell how the Wayland crowd would respond to this, namely by advising me to buy a "modern laptop" or to just stick with "X11/X.org because no one's forcing me to migrate but I'll have to deal with the consequences of running abandonware".

This appears to be the main selling point of Wayland; nothing but hot air under the banner of "modernity", satisfy some "developer experience", whatever. And they exclusively focus on Linux, not any of the BSD's.

Actually, this makes *BSD a serious consideration, in case Wayland becomes mandatory across the entire Linux ecosystem. I guess there's still hope for my still-working "ancient" hardware.