I haven't seen many desktop apps emerging in the last decade, so whom is that "legacy is holding back the Linux desktop" argument addressed at? X may not be the most modern architecture, but it's still the API almost all F/OSS desktop apps are ultimately written and tested against. For many apps, wayland doesn't bring a new perspective, but end of life with no new apps taking their place. For other apps, wayland means additional testing and porting effort. And wayland, like gnome and almost every other "innovation" by RedHat, brings fragmentation into the free desktop world also for BSD and Mac OS ports. Desktop app development, being starved off developers anyway, doesn't need a "Python 2" moment, that is, a decade+ long phase of transition in the name of dubious progress, or other impediments getting thrown in its way. The end result is simply that less usable software will exist.
It would be funny that at the point where it seems that Linux desktop hardware support is at its peak, the actual use of Linux desktop would be at its lowest. I don't think that's the case, though.
GNU/Linux gets used alright, on IoT, servers and VMs, where the desktop is irrelevant.
While Linux community can pat themselves on the back for having Linux kernel as part of ChromeOS and Android, they tend to forget it is hardly exposed to userspace frameworks used by app developers.
Crostini only works in some models, and follows the same approach as WSL, running GNU/Linux in a virtualized Linux kernel, completely unrelated to ChromeOS Linux kernel.
ChromeOS vNext could be based on Fuchsia and still offer Crostini.
This would be more believable if there was a surplus of new Windows/Chrome OS desktop apps. Even macOS which is doing a little healthier basically gets a couple of indie commercial apps that people care about them.
The rest of the new desktop apps are electron which doesn't care about X vs Wayland
You are overestimating the difficulties of porting old stuff to wayland. XWayland doesn't have a prespecified EOL, it's just going to sit there indefinitely to support all the legacy apps. This means that there is no need to hurry with migration.
And the issues in X that wayland aims to fix are much bigger than the issues that python2 had when it got replaced.
“You are overestimating the difficulties of porting old stuff to wayland”
The same is true for moving away from win32 GUI API. How many times has Microsoft tried?
I have experimented with wayland and considered submitting patches for various 30-bit color issues, but the community is so catty. Everything is met with snark. It reminds me so much of trying to help a damaged person who doesn’t trust me, which is so weird and off-putting, in this context.
> The same is true for moving away from win32 GUI API. How many times has Microsoft tried?
I'm not very familiar with this example, but experience from proprietary systems rarely applies to FOSS because the way free software is developed is so different. Furthermore, Qt and GTK apps work out of the box with both wayland and X. The same applies to many other toolkits and libraries.
> I have experimented with wayland and considered submitting patches for various 30-bit color issues, but the community is so catty. Everything is met with snark.
What part of the ecosystem did you attempt to contribute to? I got a couple of patches merged into Sway and some of its utilities and the community always felt rather welcoming.
