[GoLUG] The upcoming Wayland power grab (and you thought the systemd war was bad...)

[Kyle Terrien]

If you thought the systemd war was bad, just wait until the
powers-that-be force Wayland upon you.

https://gist.github.com/probonopd/9feb7c20257af5dd915e3a9f2d1f2277

Wayland does less, and it does less by design.  They don’t want you to
forward your Wayland session like you can with X11.  They don’t want
applications to share or record the screen unless they go through an
ever essentially changing GNOME API.

Wayland is like UEFI.  It specifies the minimum amount and leaves the
details up to implementation-specific extensions.  As a result, there
is no consistency between environments.

Also in Wayland, the display server and window manager is one big
monolith.  Your window manager is your compositor and display server.
It links against Wayland libraries and implements the entire display
server stack.  Want a lightweight desktop?  They must keep up to date
with the ever essentially changing GNOME and KDE extensions in order
to stay “relevant.”

Want global hotkeys?  Too bad.  Only the window manager (which is also
the display server) can globally bind keys.  If an application wants
to bind a global key, then it must negotiate with the window manager
(using a GNOME/KDE-specific API, of course).

Wayland refuses to define standards for window decorations, instead
leaving it up to each application to (1) figure out which desktop
environment he is running in and (2) to properly draw the correct
decoration in the correct style for the correct desktop.  Each
application must be aware of every desktop in order to draw “native”
decorations.  Of course, such a Sisyphean task is infeasible.

And what is the benefit?  Surely this is a lighter and faster
windowing system.  Nope!  Wayland is heavier and slower than Xorg.

https://distrowatch.com/dwres.php?resource=showheadline&story=20024

Meanwhile, (and this is the really crazy part) the people in charge of
proselytizing Wayland have quietly taken over Xorg development.
Conflict of interest much?  Yep!  The Wayland people are ignoring
bugfix patches and reverting patches that were already merged, slowly
choking the Xorg project.

Now each distribution effectively maintains its own fork of Xorg, out
of necessity.

The most prominent upstream Xorg fork is XLibre.  Slackware-current
recently added it to testing.  The BSDs are looking into adopting it.
Debian is taking a “wait and see” approach.

https://github.com/X11Libre/xserver/wiki/Are-We-XLibre-Yet%3F

XLibre is now 1 year old, and it has survived a tidal wave of
political attacks.  Red Hat employees banned the lead developer from
FreeDesktop.org infrastructure.

https://mail-index.netbsd.org/netbsd-users/2025/06/06/msg032747.html

> I don't know why, but it really looks they're quite scared by one guy
> that's just trying to actually bring X11 forward. Hard to find he right
> words for telling how honored I'm about that.

The developer points out the parallels with the Xorg/XFree86 situation
20 years ago, and how history is repeating itself:

> It's not the first time this happens in FOSS world, and it's not the
> first time it's happening in X: remember what Xfree86 board did to
> the honorable Keith Packard, back about two decades ago - what lead to
> the birth of Xorg and the death of Xfree86. Same is happening again.

----------------------------------------------------------------------

I have been saying for years in personal conversations that Wayland is
broken by design.  It’s time to say it publicly.  Wayland is broken by
design, and anyone with a brain can figure it out.

What they should have done (since backwards compatibility was
explicitly not a goal) was design and implement a Plan 9 like
windowing system.

https://doc.cat-v.org/plan_9/4th_edition/papers/812/

> For graphical applications, the virtual terminal model is extended
> somewhat to allow programs to perform graphical operations, access
> the mouse, and perform related functions by reading and writing
> files with suggestive names such as /dev/mouse and /dev/window
> multiplexed per-window much like /dev/cons.

Wow, what a concept!  An application could draw graphics by simply
writing to a character device file.  Output shows up in the same
window, because the whole windowing system is a (graphical) terminal
multiplexer.  (You could also run the windowing system inside a window
for a true “inception” experience.)

Imagine being able to do real graphics inside an SSH session (because
it’s just another stream).  But nope!  You can’t have these good
things because Wayland says so.

Linux has spent the past two decades implementing namespace features
into the kernel (features that Plan 9 originally innovated) because
Plan 9 was right.  Why can’t the desktop people figure out that Plan 9
was right?

The fact the desktop people never considered Plan 9 demonstrates their
ignorance of history.  Those who don’t understand history are doomed
to repeat it.

The fact the desktop people constantly essentially rework their
designs shows just how addled and short-term their thinking is.  Those
who are prideful have the worst falls.

It’s hard to predict what will happen next.  Fortunately, X11 has too
much debt leaning on it, so the rise of a fork like XLibre is
inevitable.  Unfortunately, a continuation of X11 development means we
will need to wait another 20 years before we can see a Plan 9 like
windowing system take flight.  In the meantime, I hope we aren’t stuck
with Wayland.

Wayland---Unofficial motto: “You get less because we know better!”

--Kyle


-- 
[*] Kyle Terrien
    Terrenus => from the Earth, to the Cloud
    https://terren.us/

Dilexisti justitiam, et odisti iniquitatem.  -- Psalmus 44:8