Recently, many people have expressed concern over Ubuntu's desire to migrate from the X Window System to the Mir display server, which the Ubuntu team will manage themselves. The bulk of the concern seems to be confusion over why Ubuntu developers wouldn't use Wayland instead.
In this article, I'll explore the official reasoning for the decision, while also exploring some additional considerations that most people aren't talking about.
Mobile Is Ubuntu's Focus.
Hate it or love it, Ubuntu has a clear vision regarding their switch to Mir for a display server. And it's fairly clear that many people who have expressed frustration over Ubuntu's desire to switch to Mir haven't bothered to read through the official MirSpec page. It's on this page that the Ubuntu team explains why they're going with their own solution and not with an alternative like Wayland. Remember, despite Ubuntu's gift of a great desktop environment, their growth focus will be on the mobile/tablet front. In the eyes of the Ubuntu developers, ignoring mobile isn't an option any longer.
I don't think that anyone can argue that sticking with X is a valid option these days. With Ubuntu's push into the mobile and desktop environments, X isn't very attractive any longer. It uses thirty-year-old code that has seen a lot of patchwork along the way. It's served us well, but I am ready to look to alternatives.
The Community Isn't Happy.
Being immersed within the Linux community both in work and in play, I've noticed that many in the community are concerned about what might happen to Wayland if Ubuntu—and potentially others—make Mir their exclusive focus. I think this is a frighteningly limited and backwards mindset. Why in the world would we be rooting for the underdog, when it may not necessarily be the best option available?
I honestly don't care who comes up with the next X server replacement so long as it meets these criteria:
- It's open source in nature. This means that if another project wants to tweak it or build onto the next-gen display server, they can without any hassles.
- The desktop remains a primary focus for the next-gen display server. As things stand now, Wayland is being designed with the desktop Linux distro in mind. Mir on the other hand, is trying to be the universal solution for mobile, desktop and tablet development. And if they drop the desktop, I'm sure that other projects will step up to take its place.
Wayland vs. Mir
Some people within the Linux community have made this rivalry among display servers into a sort of angry "zero-sum game." But I don't see any losers coming out of the match-up competition between Wayland and Mir. If Mir ends up becoming the go-to display server for Ubuntu and eventually other distros, then Wayland has a choice—step up their game and catch up, or give up.
I should note that this is all hypothetical at this point, but it does give us something to consider. While Wayland exists and is likely to begin appearing in new distributions any time now, its development isn't moving fast enough.
Now, here's where things really become interesting. Wayland has been in development since 2008. And, to be honest, it's still trying to iron out basic functionality such as minimizing and maximizing windows. Mir, on the other hand, was announced in 2013 and is already showing up in rough pre-release video through XMir, since the Unity desktop is waiting to be re-written for Mir specifically. Now step back for a second and really think about this. Wayland has had years to get its code together so it can actually replace X. Mir, which is brand new, looks to be closing in on Wayland and doing so in vastly less time overall. Yet, we're still cheering Wayland as our savior to X server? Give me a break!
At this point, I think two things are going to happen. The first is that Mir will beat Wayland to the desktop in a truly usable state. Despite Wayland having 1.x already available, it's hardly usable. The second point is that it's entirely possible that Wayland could overtake Mir in the long run for desktop users. And to be honest, I'd love to see Wayland do this. After all, both projects hold tremendous promise, and both projects will free us from X!
Wayland Is Still Ahead of Mir.
Despite Wayland's slow pace, it's still vastly ahead of Mir at this point. I acknowledge this, and I will even say this presents some interesting challenges to those developing Mir. It's also important to note that the upcoming release(s) of Wayland will likely bring forth the missing functionality, such as minimizing and maximizing windows. And if this happens before Mir is released, this will put Wayland much farther ahead of Mir.
What's my response to this possibility? Awesome! Because this would give us two great display server options to choose from as time passes along.
Remember what I said previously, this isn't a zero-sum game. Competition between the two projects only serves to benefit Linux users collectively. So if Wayland comes out the winner for the desktop, with Mir being useful only to Ubuntu, I'm completely fine with that. What I find so perplexing, is all of the complaining about something that isn't a threat to Wayland within the Linux community. It makes little sense to me and offers no value to anyone.
The Canonical-Is-Too-Powerful Argument
In my effort to better understand the paranoia about Canonical "seizing control" over the hearts and minds of Linux users, I've come to the following realization: The Linux community includes people full of passion and opinions. This is especially true when it comes to developers. Speaking as a non-developer, I have the advantage of looking at this with indifferent eyes.
Anytime I get into a debate over Mir vs. Wayland I find the conversation quickly goes from facts to opinion. For example, one recent conversation started off talking about the head start Wayland has had over Mir and how quickly Mir is catching up despite the obvious challenges in creating a brand new display server. It started out as a healthy debate, but quickly the conversation went into speculation. And the ultimate statement was that Canonical is somehow not to be trusted. The argument was that "Canonical was becoming too powerful." When I pushed further to better understand the rationale, I was told that Canonical is becoming the next Microsoft. When I pointed out that Canonical is releasing Mir as an open source concept, I was then told that Canonical might do something evil with the project such as closing the code later on. You simply can't make this stuff up, folks!
Another individual brought up a fantastic point that I wanted to share. They mentioned that despite the code being open source, Canonical might force video card vendors to stick with Mir standards only. After all, their rationale is that video card vendors aren't too likely to support X, Wayland and Mir. Now this was a solid question worth exploring. My view on this is as follows: whichever project wins in the long term will be in a position to work with those building video cards. If this happens to be Mir, other distributions will be free to follow along or not.
As I bring this to a close, I want to acknowledge that there are a lot of heated debates and some strong opinions floating about Mir vs. Wayland. I maintain my position that either the best option or perhaps both options, will come out ahead in the long term. My view is that Wayland has had years to get a working product (and, no, it's not there yet in my opinion).
Currently, Mir is very much still in the concept stages; however, I've seen Canonical on a development warpath before. And no matter what you may think of them, they do development quickly. Based on a wide variety of factors, I think we'll see Wayland win early on with their display server offering, while the Mir team catches up later.