>I don't think anyone seriously thinks X is the way forward do they?
I think "forward" is the wrong mindset here. X works. Plenty of tooling is just done, lots of people have their workflow just perfect, the whole ecosystem is mature, and at worst the Wayland devs decided to uproot the whole thing in the name of "progress".
Wayland could be nice, but its compositors require 10x the code that a WM would (which is why there aren't that many Wayland "WMs") and its supposed benefits aren't better enough to justify the ambitious rewrite. I could imagine a universe where Wayland did the minimum to replace X's jank and make the thing maintainable, and nobody would have objected to that IMO.
Those goddamn Wayland devs and their individual free will. How dare they not continue to maintain this miserable, crusty old spec and continue to make their marketable skills less valuable.
> Wayland could be nice, but its compositors require 10x the code that a WM would (which is why there aren't that many Wayland "WMs"
Nah, this is an ecosystem maturity problem that is improving rapidly.
There will always be a few hold outs for the old way, but people will and are moving on.
It’s silly to compare the number of anything with a project that had a 30 year head start.
> Wayland did the minimum to replace X's jank and make the thing maintainable
That’s called Xorg. It still exists and you are free to do whatever you wish, even invest in its development,
Wayland didn’t “uproot” anything.
One of the biggest pieces of “jank” replaced is the nonexistent security model. You can research why the proposed SECURITY extension was unworkable and smarter people decided it was better to start from scratch. This benefit alone is better enough for most people with a stake to justify a rewrite.
In pure economic terms there is certainly a price where Xorg can continue to remain viable - this isn’t some ancient artifact lost forever. In this lens it’s hard not to see these comments as anything but unproductive bitterness at not being able to provide or raise these funds. It doesn’t help that part of this price is a direct result of the qualities of the thing you’re trying to save.
>Those goddamn Wayland devs and their individual free will.
That's not the problem. The problem is all the heckling for everyone to switch to Wayland, and to make it default. And also pretending that reversing the "mechanism, not policy" wasn't a fundamental change of philosophy and was just "progress".
And to be clear, by "at worst" I meant "this is the least-charitable interpretation".
>Nah, this is an ecosystem maturity problem that is improving rapidly.
TinyWM is 50LOC and wlroots's TinyWL is 900LOC. This hasn't changed. Writing Wayland compositors is a pain in the ass compared to WMs.
Wayland started in 2008 IIRC, so here in 2024 Wayland is 16 years old. It doesn't have teething issues, it just has issues.
Plenty of WM makers have just straight-up said they won't ever port their WM to Wayland (so talking about time and how "X is older" is irrelevant here), because Wayland's opinionation breaks too many things.
>One of the biggest pieces of “jank” replaced is the nonexistent security model.
You mean like how X clients can read keyboard inputs? There's so much FUD around this. /dev/keyboard does that already, you need to sandbox every app anyway - at which point your sandbox should just interdict the X interface. Security is the worst argument. Wayland isn't necessary for security, and for the longest time it's been locking the door and leaving the window open. Portals.
> The problem is all the heckling for everyone to switch to Wayland
No one’s making you switch and Xorg is as open as it ever was. X is dying, slowly, that’s just a simple fact of the universe.
> Plenty of WM makers have just straight-up said they won't ever port their WM to Wayland
There are plenty of up and coming WM makers willing to produce suitable stacking and tiling WM/compositors. Active projects exist today.
> You mean like how X clients can read keyboard inputs?
No.. how every X client has global access to all of the X state of every client and how the only response to this are extensions that no one really uses for well established reasons.
> you need to sandbox every app anyway - at which point your sandbox should just interdict the X interface.
I think "forward" is the wrong mindset here. X works. Plenty of tooling is just done, lots of people have their workflow just perfect, the whole ecosystem is mature, and at worst the Wayland devs decided to uproot the whole thing in the name of "progress".
Wayland could be nice, but its compositors require 10x the code that a WM would (which is why there aren't that many Wayland "WMs") and its supposed benefits aren't better enough to justify the ambitious rewrite. I could imagine a universe where Wayland did the minimum to replace X's jank and make the thing maintainable, and nobody would have objected to that IMO.