Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

OK here goes..

- FLAVOR: Ubuntu Desktop

- HEADLINE: Drop Mir & collaborate with Wayland

- DESCRIPTION: I know this is a touchy subject and I'm not looking to self-righteously re-re-re-ligitage everything but... between Intel walking away, licensing concerns, Ubuntu varients not jumping onboard, and various community concerns, would you re-consider abandoning mir and joining forces with Wayland? I understand you felt there were some technical shortcomings regarding how input devices were handled. Perhaps in today's climate those concerns can be better addressed by Wayland if you can provide the engineering leadership on those efforts?

- ROLE: Code Janitor



I was about to write this exact post. This is the main reason I use Arch Linux. Arch just follows upstream without worrying about holy wars.

There are dozens of ways Ubuntu can innovate/differentiate. Fighting a holy way of Mir vs Wayland (when you came second, and Wayland already had major support) is not a good use of your time. Same goes for Snap vs Flatpak. And what about Juju, or Bazaar (though that's dead now, isn't it?). You did well accepting systemd over Upstart. You could do the same with Wayland (over Mir), Flatpak (over Snap), and officially EOL Bazaar.

I'm not a Unity fan, but that seems like a genuine way for Ubuntu to differentiate and/or compete. Mir and Snap are not. They're just incompatible alternatives that divide the open source community. I know that alternatives can often be good, and inspire competition. But when it's clear you've lost, it's sometimes better to follow your Upstart/Systemd example and join the winning side.


> - HEADLINE: Drop Mir & collaborate with Wayland

That would more accurately be:

- HEADLINE: Port Unity 8 to be a Wayland compositor

This would be at least a multi-year effort. Consider how much time and effort went into porting Gnome Shell and KWin over to Wayland. It would be at least as much work do the same with Unity 8.


Finishing Mir is also a multi-year effort. So, where would the programmer-hours best be spent?


Exactly. Let's avoid the sunk-cost fallacy.


Sorry if I wan't clear, the point was that it wouldn't be possible to do in time for 17.10, which is what the original post was asking about.


> Drop [Canonical-specific] & collaborate with [leading variant]

That would be great in general. Linux Mint is known as "Ubuntu minus Canonical" for a reason.


Except they invent their own shit all the time also. Sometimes to the detriment of existing products.


Linux Mint is also "Ubuntu without Security". That's right -- Linux Mint does not install Ubuntu's security updates, nor supply their own.

Stated another way: 100% of Linux Mint machines in the world are either:

(a) already pwned

or

(b) vulnerable to multiple CVEs and will be pwned


+1 to explore using Wayland over Mir. I think it's intuitive that everyone (distros, developers, users) would benefit from unifying the Linux graphics stack to the greatest extent possible. The display server decision should not just be based on what is best for Unity, but what is best for the Ubuntu ecosystem as a whole. Having multiple display servers means there will be a duplication of effort for developers of application toolkits, window managers, and, of course, the display servers themselves. The proper question in my mind is whether all the additional downstream and lateral effort from maintaining multiple display servers is a bigger opportunity cost to the overall Ubuntu experience than the additional development necessary to get Unity running on Wayland. It's worth revisiting the display server question because this is something that may have major consequences for quite some time.


Look at the main request on this page: hiDPI. That is what we need a better graphics stack for.

The reason we need focus on Wayland is that NVIDIA and AMD need to focus on a specific graphics stack. That wasn't our choice. That was theirs. Thankfully, AMD is headed in the right direction with AMDGPU.

Unfortunately, NVIDIA shows no signs of sensibility here, so we must be wise, and focus their efforts away from Xorg. The best way to do this is for canonical to forget about Mir, and push its support behind Wayland.

We need your help, Canonical! Let us fight stagnation together, lest we find ourselves using Xorg for another 10 years.


There's no technical justification. Mir is a purely political decision. Canonical wants to exert more control over the platform. It's a totally rational position, but as you note, it seems to be failing.


To be fair, I could absolutely understand their decision. After watching the people behind Gnome3 and Wayland pretty much run amok with their ideas and their attitude towards other, dependent projects, I can't blame Canonical for not having that much faith in a possibly unstable foundation. Especially with their plans back then of building a mobile product on it. So I really can't see how Canonical are the bad guys here. It was not so much about more control but against losing it somewhere down the road.


Yeah, I didn't say they're bad guys. I just think it's important that people recognize Mir was never about technical considerations. It was a land grab, and I agree, we may very well have been better off with Canonical in control of the platform.


> - ROLE: Code Janitor

That sounds like a fun job. You just clean code all day?


Oh god yes. I forgot that one.


I'm sure Mir is a perfectly good replacement to crufty old x.org. But you know what? Upstart was a perfectly good replacement to crufty old sysvinit, and we all know how that one turned out. I predict this will turn out exactly the same; Ubuntu will go with Mir for a few years before eventually caving to community pressure and switching to Wayland, just like the ultimately gave up Upstart in favor of community-favored systemd.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: