Critical.
Authoritative.
Strategic.
Subscribe to CIO Magazine »

Ubuntu's risky leap: Unity on Wayland

Ubuntu Linux will adopt Unity on Wayland in favour of GNOME on X.org.

Future versions of Ubuntu will ship with the Unity interface

Future versions of Ubuntu will ship with the Unity interface

Today Canonical and Ubuntu founder Mark Shuttleworth announced on his blog that the Ubuntu distribution will move away from the traditional X.org display environment to Wayland a more modern alternative.

Read Shuttleworth’s blog here.

The news comes only days after Ubuntu announced it would change its default user interface from the GNOME environment to its own Unity, which is better suited to netbooks.

There are a number of possible side effects as a result of moving away from GNOME and X.

Despite all its posturing about how “Ubuntu is not dropping GNOME” and “GNOME is still available” the decision to move away from the mainstream GNOME desktop to its own Unity could be seen as a slap in the face to the rest of the GNOME community developers.

There’s now little reason for these GNOME developers to recommend Ubuntu as an operating system.

By focusing on Unity (on Wayland or X) for Ubuntu, Canonical has essentially forked its own Linux distribution.

I understand Wayland is “compatible” with existing X apps, but the porting work involved may lead to bugs and other inconsistencies Ubuntu can do without.

Even if the mainline Ubuntu distribution offers users the choice of installing the GNOME desktop, it’s only a matter of time before a GNOME-only distribution emerges much like Kubuntu is for KDE.

This GNOME-only Ubuntu distribution could be spearheaded by Canonical or it could be community-driven.

My gut feeling for Canonical moving so aggressively in this direction is the rise of ultra-mobile computing.

Shuttleworth wants a piece of the netbook and tablet PC action and the standard PC interface (GNOME, KDE, Windows, etc) just won’t cut it.

Microsoft has tried to transplant its PC desktop experience onto small, more mobile devices with only mixed results.

The big winners in this race are the ones who dared to redesign the PC interface from the ground up. Apple’s iOS and Google’s Android are leading the revolution.

Why might I suggest this? Well if you examine Shuttleworth’s “reasons” for wanting to switch from X to Wayland it revolves around user-experience.

He speaks of “super-smooth graphics and effects” and considering the Android environment.

These are things that modern desktops are less likely to suffer from due to improvements in processing power and display technology. So the impetus lies with portable computing.

Moreover, Shuttleworth doesn’t really give compelling reasons to move away from X, only to say “it’s extremely hard” to get good graphics and effects with it.

Given the mobile phone I’m carrying runs X and its performance is indistinguishable from an iPhone or Android device there isn’t a lot weight behind Shuttleworth’s reasoning.

X might not be as responsive and scalable (down) as it could be, but that doesn’t mean it’s performance is so bad it needs to be turfed out.

I’ve even asked Linus Torvalds the million dollar question of whether Linux needs another type of display environment and he said something along the lines of “it would be silly to replace X”.

I also remember once during a presentation by Rasterman (Aussie open source graphics guru) he said the overhead of using X is not much to worry about either.

Another problem for Ubuntu will arise if Wayland can’t provide the feature set of X people will be forced to look elsewhere.

Ubuntu is definitely navigating uncharted waters here and it’s a one-two punch to its steady ascension to fame.

Sure Ubuntu is being innovative and it deserves credit for that, but I question whether it needed to radically change its desktop and interface for its mainstream distribution.

Just reading over Mark’s blog again I’m waiting for him to write something like “this alternative to the mainstream Ubuntu will be called …buntu”.

Instead he’s jumping over the frying pan and into the fire. He’s prepared to break something that only needed a minimal amount of fixing to be a credible alternative to today’s mainstream desktop operating systems – with the immeasurable advantage of being free – and start again.

I wish Ubuntu every success in its new quest to conquer mobile computers. But as far as the desktop goes it’s a very risky move that could alienate developers and users alike.

In years to come it might look back and wonder why it didn’t focus on making its good distribution even better instead of radically overhauling it.

Rodney Gedda is Editor of TechWorld Australia. Follow Rodney on Twitter at @rodneygedda. Rodney's e-mail address is rodney_gedda@idg.com.au. Follow TechWorld Australia on Twitter at @Techworld_AU.

Tags mark shuttleworthX.orgcanonicalgnomeWaylandubuntu

More about AppleGoogleKDEKDELinuxMicrosoftUbuntu

Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

9 Comments

Luke Leighton

1

For ARM9 and even ARM11 cores, and certainly for MIPS CPUs below about 600mhz, the lack of a good 1st level cache means that context-switching KILLS performance.

X11, which is a separate server, requires a context switch [Look closely at Unity and you'll probably find that it's a library, not a server]

Thus, on ARM9 CPUs running at 400mhz for example, and even on ARM11 CPUs at 667mhz running large 1280x768 LCDs, you simply cannot get a responsive OS when using X11, because each time there is a switch between OS processes and the X11 server the CPU's 1st level cache is entirely thrown away!

When people say that Microsoft + Intel have held the advancement of computing back by 20 years, this is what you get happening, and it's only now that ARM CPUs are beginning to break the 1ghz barrier that they're being taken seriously enough, and the limitations of the "underclass" architectures is being "discovered".

Anders Skare

2

Ubuntu caters for the large audiences, not the specialized ones.
1. Wayland will be able to run X11 *apps* rootless, so the X application investment will *not* be thrown out the window. We *will* be able to continue leverage our previous investment..
2. Most people already use alternatives to remote X-protocol, specifically VNC. The loss of a network based display protocol is a non-issue for the larger audience since we don't rely on it today.

So in light if this, I really don't see anything negative with the direction Ubuntu is going. Keep in mind that they're a distribution for the large unwashed masses (the average Jane and Joe). The few people with special interests have a myriad of other distributions to choose from, or they could continue adding X to their Ubuntu distribution if they need it, same as today.

It's not like Ubuntu will be changing their policy regarding how a user may customize his/her system. It'll still be a derivative of the wonderful Debian package tree and you'll still be able to install whatever you like on it.

If you're specialized today, then you probably don't run the default Ubuntu install out of the box / as is, but have already made a number of adjustments, so how will this change to the "default" package tree be any different for users with specialized needs compared to how it already is today?

I'd say to Mark: a very good decision which is likely to spur some interesting new avenues for innovation or at least incremental progress. Behind you 100% on this one.

tesla

3

So, does this mean there will be a "Gubuntu" in the works somewhere?

bjorn

4

Anyone that actually DID X programming has to know it is VERY resource inefficient, architecturally inellegant and promotes buggy code. X had been designed primarily with a client/server (where the server runs the display) model in mind, but on a normal desktop/notebook/pda/phone you dont want that overhead.
Linus Torvald is not the only one that has a degree in Computer Science, you know. He is just getting old.

Zac

5

Another good and gutsy move. I 100% support it.

Now please push one Linux install file type with innovative features and do away with deb's and rpm's.

John

6

bjorn: I've done a lot of X programming and I find it a wonderful system. Actually the X server has very minimal overhead, what really slows down stuff is bloated toolkits like GTK and Qt. I've run bare X with Xt/Motif clients on very limited ARM9 devices and it runs without a hitch.

Also architecturally X is unmatched to this day. The clean separation between policy and mechanism, and client and server, and the fact that you can run programs over the wire by just changin an environment variable are wonderful features even if you personally don't happen to use them.

Moving away from X because of misconceptions of inefficiency is a terrible mistake.

Bill

7

@John,

X is fine if all you want to do is run a few xterms. That bare X on ARM9 you talk about won't do the sort of compositing, animations, 3D effects, buffer object management, and tear free frame updates that are expected in a modern display system.

John

8

So this will make it similar to the way OS X handles X, just another application/server running to servegraphics to the client apps that need it?
Not really a problem as long as the X integration is seamless, and can still leverage all the loo and feel of the native apps (this is not really the case in OS X, but that may be more that the apps were never designed with OS X in mind)
But I would like to see better resolution switching and detection, make it smooth, professional, fade in, out etc. Many users just don't want to see the text mode console when switching users or logging out. Ubuntu has gotten better, but I do find it painful sometimes undocking my laptop form my screen to find I no longer have the local LCD working, and I have to restart X to get the laptop display re detected etc.

senorVince

9

Isn't Ubuntu already a "credible alternative to today’s mainstream desktop operating systems"?

Comments are now closed

Computerworld
ARN
Techworld
CMO