3 ms·

> I find the one based on duplication of effort the hardest to swallow Having a display stack specific to one distribution makes things quite a lot more diffic

by bkor 10y ago

> I find the one based on duplication of effort the hardest to swallow Having a display stack specific to one distribution makes things quite a lot more difficult than just another desktop environment. That was heavily criticised. Mir supporters often mention that the toolkit will easily handle this; but that's not the case. The Wayland version of Firefox still has quite a few issues. It took a huge amount of effort to have X11 applications work nicely under Wayland using Xwayland (copy/pasting, dragging, etc). There's still various things left. Various things in Wayland are still a bit iffy. E.g. some Wayland solutions are specific to one desktop. Fixing all of this and getting an agreement across desktop environments is a huge amount of effort. Then this has to be doubled because hey 'Mir'. Obviously that's heavily criticized. Though various developers just said "No" (e.g. Kwin developer).

cookiecaper 10y ago
Yes, I'm usually not one to whine about "fragmented" project spaces and "wasted talent", but Mir is a special case because the Linux desktop is necessarily dependent on cooperation from hardware vendors. Historically, they've seen Linux as a source of confusion, and avoided wading too deep because they couldn't really make anyone happy, and because it was hard to tell if their investment would pay off or if they'd have to redo all the work when the "community" decided to shift again. With Wayland, we had the opportunity to present a unified, modern desktop environment to them and say "Here, support this and we'll be happy". Then Mir showed up with absolutely no substantiation or value proposition and retriggered the prospect of nightmarish half-support across a fragmented ecosystem, which tires everyone out, not just vendors. It was a great illustration that open-source and Linux haven't changed: they're still difficult to work with, and you still don't know if the engineering effort you're dedicating is going to be flushed away due to distro infighting. NVIDIA had already said they were playing a wait and see game before committing more resources to the next-gen Linux desktop, and Canonical only proved their decision right.

by bkor 10y ago

cookiecaper 10y ago

Yes, I'm usually not one to whine about "fragmented" project spaces and "wasted talent", but Mir is a special case because the Linux desktop is necessarily dependent on cooperation from hardware vendors. Historically, they've seen Linux as a source of confusion, and avoided wading too deep because they couldn't really make anyone happy, and because it was hard to tell if their investment would pay off or if they'd have to redo all the work when the "community" decided to shift again. With Wayland, we had the opportunity to present a unified, modern desktop environment to them and say "Here, support this and we'll be happy". Then Mir showed up with absolutely no substantiation or value proposition and retriggered the prospect of nightmarish half-support across a fragmented ecosystem, which tires everyone out, not just vendors. It was a great illustration that open-source and Linux haven't changed: they're still difficult to work with, and you still don't know if the engineering effort you're dedicating is going to be flushed away due to distro infighting. NVIDIA had already said they were playing a wait and see game before committing more resources to the next-gen Linux desktop, and Canonical only proved their decision right.