Valve Engineer Mike Blumenkrantz Hoping To Accelerate Wayland Protocol Development

[phoronix]

Phoronix: Valve Engineer Mike Blumenkrantz Hoping To Accelerate Wayland Protocol Development

Valve open-source graphics software engineer Mike Blumenkrantz is well known in the Linux community for his work on the Zink OpenGL-on-Vulkan driver code, various Mesa driver optimizations, and creative writing on his blog. He's also taken up a new task: further accelerating Wayland protocol development...

https://www.phoronix.com/news/Blumenkrantz-Faster-Wayland

[skeevy420]

On today of all days

[hf_139]

I think that is great.

Finally someone does something to get rid of those ever-blocking GNOME "developers" and traitors.

[avis]

Maybe just maybe it could occur to some people that there's no pressing need to implement 100+ Wayland extensions over and over and instead have a single implementation that splits the whole shebang into a server that renders everything, manages input, clipboard and shortcuts (and a dozen other core features) and window managers that paint window decorations and manage windows' placement and interaction. You know how it's been tried and tested, and actually worked beautifully sans a few features that were well outside the protocol scope and now cannot be implemented properly because it was envisaged over 40 years ago and people couldn't have forseen where and how display tech would develop.

Or we can have what we have now: over two dozen half-incomplete implementations, all with various bugs and present or missing features, where you never know what you're getting and you just cannot depend on anything and use automation how it was done with Xorg. You wanna turn off the monitor using a bash script in Wayland? Good luck writing a script for that that works across all the implementations. And there are dozens of such scenarios.

Why is it that multibillion corporations can only afford having a single display server but extremely underfunded and understaffed Linux wants to have several dozen? It will never work, period, nor there's any need for that. The display server is not something that needs to be reimplemented over and over again. 99.9999% of people in the world don't even know there's such a thing under the Sun.

And while Linux fans continue to press with "multiple implementations are not a bad thing", how many of them have donated a single cent to their favourite Wayland server? Less than a hundred in the entire world with donations below $1000? Who are you kidding when decent software engineers e.g. in the US earn on average over $150K a year or $12.5 a month? Your donations are worth a few full days of work. Do you really think implementing the Wayland spec is all so simple anyone can do that in spare time? Then why of all the Wayland implementations we have just one that is sort of complete, Kwin, then comes Mutter that's lacking some serious new features, and then all the others are just laughably incomplete? I tried Wayfire that is supposed to be the best next independent thing a month ago and it crashed a few times in a row in just one day. All under God bestowed the amdgpu driver mind you and while running Linux 6.10.10.

16 years in and Linux doesn't even have a crucial most important thing implemented for everyone, not for the selected who chose KDE or Gnome as their precious thing. Many of use don't want to touch either. There's nothing in the Wayland spec that says that you must use KDE or Gnome either.

It's just sad.

[Phoronos]

Maybe just maybe it could occur to some people that there's no pressing need to implement 100+ Wayland extensions over and over and instead have a single implementation that splits the whole shebang into a server that renders everything, manages input, clipboard and shortcuts (and a dozen other core features) and window managers that paint window decorations and manage windows' placement and interaction. You know how it's been tried and tested, and actually worked beautifully sans a few features that were well outside the protocol scope and now cannot be implemented properly because it was envisaged over 40 years ago and people couldn't have forseen where and how display tech would develop.

Or we can have what we have now: over two dozen half-incomplete implementations, all with various bugs and present or missing features, where you never know what you're getting and you just cannot depend on anything and use automation how it was done with Xorg. You wanna turn off the monitor using a bash script in Wayland? Good luck writing a script for that that works across all the implementations. And there are dozens of such scenarios.

Why is it that multibillion corporations can only afford having a single display server but extremely underfunded and understaffed Linux wants to have several dozen? It will never work, period, nor there's any need for that. The display server is not something that needs to be reimplemented over and over again. 99.9999% of people in the world don't even know there's such a thing under the Sun.

And while Linux fans continue to press with "multiple implementations are not a bad thing", how many of them have donated a single cent to their favourite Wayland server? Less than a hundred in the entire world with donations below $1000? Who are you kidding when decent software engineers e.g. in the US earn on average over $150K a year or $12.5 a month? Your donations are worth a few full days of work. Do you really think implementing the Wayland spec is all so simple anyone can do that in spare time? Then why of all the Wayland implementations we have just one that is sort of complete, Kwin, then comes Mutter that's lacking some serious new features, and then all the others are just laughably incomplete? I tried Wayfire that is supposed to be the best next independent thing a month ago and it crashed a few times in a row in just one day. All under God bestowed the amdgpu driver mind you and while running Linux 6.10.10.

16 years in and Linux doesn't even have a crucial most important thing implemented for everyone, not for the selected who chose KDE or Gnome as their precious thing. Many of use don't want to touch either. There's nothing in the Wayland spec that says that you must use KDE or Gnome either.

It's just sad.

blablabla...
Do the coding and show us code.

[Quackdoc]

I feel like make has consistently proven that he is the only sane developer working on linux stuff. That's not to say I always agree with him, the stuff he says and does and what not. But here especially seems to be a quite well thought out methodology.

[smitty3268]

how many of them have donated a single cent to their favourite Wayland server?

How many cents have you donated to change things the way you'd like? The developers want it to work the way it does, that much is clear. If you're the one who's unhappy, maybe you're the one who needs to contribute to get it to change.

[mirmirmir]

They got dawg in them

[bkdwt]

Gnome developers need to go to hell!

[fitzie]

this is like shoveling the driveway in a blizzard. trying to fix one little broken thing with wayland when the cabal has the fundamentally wrong attitude. the corporate interests out there still funding this linux graphics need to step up and put an end to this madness. at this point the only way forward is to tear down freedesktop.org and rebuild a sane institution that organizes around replacing x11, cross compositor compatibility, and start catching up to the capabilities of other modern desktops, like with a roadmap and everything. gnome should probably be banned from leadership roles until they add CSD and publicly apologize for causing so much frustration for app developers and users for so long after it was clear their vision was faulty.