Re: [dev] Wayland compositors

From: Tobias Bengfort <>
Date: Fri, 17 Sep 2021 06:33:39 +0000


I think much of the hate for wayland is misdirected. Don't get me wrong,
the state of wayland is bad. But IMHO that is not entirely the wayland
people's fault.

What they did is saying: "Hey guys, we are tired of maintaining X. We
will start a new project with a tighter focus. The wider linux community
has to step in for all the parts that we no longer support. Just use
dbus or something."

This move was not great and also not very well communicated. But hey,
it's open source and I cannot force anyone do work on X if they don't
want to.

I think what they intended was that some people mess around with some
ideas and after a while everyone agrees on a single set of protocols (or
maybe even a single compositor implementation).

Imagine a scenario where we had a single compositor that had an API for
window management. That is entirely possible. (I built a window manager
that uses the sway IPC mechanism a while ago.) Wayland does not prevent
the old server/window manager separation. But it also does not enforce
that separation.

Unfortunately the wider community did not come together on this.
Especially gnome created a bunch of specific solutions that will never
work outside of the the gnome ecosystem. The coordination that is
happening at is
negligible. The best thing we have in that direction is wlroots, but
most major desktops refuse to use it.

Most people in this thread seem to agree that the wayland protocol has
some improvements, but the lack of features is a dealbreaker. We could
absolutely have a full replacement for the X server based on wayland.
The wayland people won't build it for us though. Neither will the gnome
people. You can decide for yourself whether you want to keep whining or
actually do something about that.

Received on Fri Sep 17 2021 - 08:33:39 CEST

This archive was generated by hypermail 2.3.0 : Fri Sep 17 2021 - 08:36:07 CEST