The clock is turning back: we used to call these things "workstations," a name that stood for a powerful but small computer sitting on a desk somewhere, definitely not something that everyone had or needed. We should call them that again: most of us won't need "workstations," but some us do.

The word "PC" has run its course. Tablets and phones are far more "personal" than a big clunky desktop would ever be. So, yeah, I would say that conceptually the PC has died, or rather has become a workstation again.

By the way, I'm one of those people who will always need a workstation... :) But it doesn't mean I begrudge or don't understand the changes in the industry. My mom sure as heck doesn't need a workstation for her email and web browsing.

Just want to point out that "fallback" does not exist anymore as a separate mode. Clutter (the window manager) will simply use the 2D LLVM pipe if 3D graphics are not available.

"GNOME Classic" is a "mode," meaning a pre-configured template for a desktop session. You can also create your own modes according to your tastes.

Same here. Xfce is mature, and everything just works. I don't have time to beta-test in my everyday work.

It's still impossible to do very essential things in GNOME Classic, such as moving the panels around or fully configuring them. So, in many ways GNOME 3 hasn't reached feature parity with GNOME 2 or Xfce.

That said, it's really nice to see GNOME listening to users. An especially important part of GNOME 3.8, in my view, is that more options were added to the settings rather than removing them. This shows that the team really is trying to stabilize the core before adding more features, which is really the right way to go about things. I think in a few years GNOME 3 will be a great desktop, suitable for various work styles.

Canonical used to pay a salary to a sole employee who did work on KDE/Kubuntu integration, but decided they could not afford it anymore and need to focus their energies on Unity. Blue Systems stepped in to offer the funding instead.

In any case, it didn't mean "ownership," it was merely offering some direct finanical support in addition to the resource support. Canonical will continue to offer resource support to all the ?ubuntus, but I don't blame them for not funneling money into areas that are not their core priorities.

Canonical doesn't "own" the ?ubuntu flavors. Instead, it provides resources: build machines with integrated QA, bandwidth and recognition.

This means that it's up to the volunteer teams who make these flavors (or the upstream desktop environments) to do the integration.

It's actually not clear that Wayland/Mir support will happen for them. It will surely happen in GTK+3 and Qt, but both Xfce and LXDE currently use GTK+2. There are talks about migrating Xfce to GTK+3, and perhaps the addition of modern graphical backends will be the tipping point for the decision.

Of course, once/if it does happen, these flavors will enjoy the same bugfixes and updates to Wayland/Mir that the basic Ubuntu distribution will have.

Same here! I do my audio recording with Ardour and an Edirol UA-25. It's immediately recognized in Ubuntu, out of the box.

Ardour may not be as good as Logic (though Ardour 3.0 just came out a few days ago, I'm eager to try it), but the JACK system on Ubuntu is astoundingly flexible. Digital plumbing!

My problem with these "just work" descriptions is that people have very different needs.

I use both Ubuntu and Mac OS regularly. The things that I need to "just work" are a lot of programming tools (gcc, python), databases, and servery stuff like databases, web servers, etc. Getting those to work on Mac OS is unpleasant. MacPorts and Homebrew are both terrible in comparison to the APT world. "apt-get install apache2" is very much "just works" in my book. On the Mac, I'm fine as long as I use Xcode and other Apple-specific tools, but anything else ends up being frustrating.

People complain a lot about desktop choices for Linux, but I never found any of them any worse than Mac, and some are better tailored for certain workstyles than others. All the major ones (GNOME with Shell or Unity, KDE, XFCE) are mature enough now for everyday work, even if they weren't so a few years ago. I really find all of them easier to work with than Mac's desktop. I don't like Mac's bubblegum dock, and I find the Finder to be perhaps the worst file manager ever made.

Another aspect of "just work" is installation. Installing a free OS can be painful on some hardware (and trivial on others). Since you can't (easily) install Mac OS on non-Apple hardware, this problem doesn't exist there, so it indeed "just works" in this respect. If you want a "just works" experience with a free OS, just buy a machine from System 76, a truly wonderful company that has yet to disappoint me. Comes with Ubuntu and everything working, great hardware and great support. And for me, all the things that I want to "just work" indeed do.

I use Windows 7, too, and it's fine, but I really need the Unix stuff to do my work.

Can we retire the "just works" phrase, or at least find better ways to qualify it?

I'm sad to agree with you.

Mark calls anybody who disagrees with his vision (which was never entirely presented to the public, but is instead rolled out in "surprises" like Ubuntu phone and then tablet) a "whiner." Anyone who criticizes a decision is being "selfish" and "childish." (These are all words he uses in interviews, blogs and bug comments on Launchpad.) But you are very right that he is setting up a straw man. Sure, there are a few childish whiners, but a large group of people criticizing the directions of Ubuntu are people heavily invested in free software, and they have an interest in seeing Ubuntu succeed. These critiques are not "selfish," they are all about trying to make Ubuntu succeed.

But Mark avoids any kind of discussion by pretending that he's dealing with "dumb" people.

Mark, it's not a black-and-white world.

I'm surprised that the biggest deal about Ubuntu phone isn't mentioned!

You'll be able to plug this phone into a dock (or otherwise connect it to a big monitor, keyboard and mouse) and use it essentially like Ubuntu desktop. There, you'll be able to run all your usual desktop applications as well as your phone applications, on a big screen with full resolution. (The do need to be built for ARM, but already most of the software in the Ubuntu Software Center has ARM versions.)

Nobody does this yet. There are dockable Android phones, but Android is not a desktop OS, and the experience on a desktop is quite miserable, both in terms of UX (mouse support is awkward) and in terms of available applications.

Phones are powerful computers! It's silly that we carry all that power around with us and yet can't apply it towards the usual desktop experience. I see the Ubuntu phone as finally being able to bridge this gap.

Even more: I can imagine desktop applications that make use phone features. GPS is not something we usually have in laptops, but phones have it, and there can be cool desktop apps that make use of it. And there's tilt-control: I can imagine big desktop games making use of tilt: the phone will become something like a game "controller" (even though the entire computer is inside, too). And, of course, you have cellular internet built in. In a way, phones, as hardware, offer more features than desktops, and app developers will surely take advantage of it!

I'm very excited about this feature, and hope to see it fronted more as one of the big advantages of Ubuntu phone!

I think the naming scheme is clear and brilliant: a number according to size, not "version" or "bandwidth speed" or any of that crap that most non-techie consumers could care less about. No more of that "S" or "Galaxy" stuff. No more internal numbering. "Nexus" is the brand family, and that's it.

Consumers are starting to think of this new network-computer world of "phones," "tablets" and "mini-tablets" exactly in terms of comfort vs. portability. How big is it? That's the only major quesiton they care about in terms of usability. Sure, there are other factors important to some users, but that should be the starting point.

I hope for their sake they keep the Nexus family numbering scheme consistent from now on. The can release new versions of these (like Apple does) with a simple numbering scheme (say, Nexus 7 version 2), but the name itself being tied to the size is perfect.

A desktop shell is a very small part of the experience. Try sending an email using Thunderbird via a tablet, and you'll quickly see the problems... What Ubuntu needs in order to be a good tablet OS is an application ecosystem, at the very least the basic stuff.

Thing is, having a relatively cheap reference platform (mature distro on an ARM tablet) will allow a lot of devs an opportunity to make touch-friendly version of their apps. And, a few years from now, we might have a smoother experience between desktops/tablets/phones.

The advantages of having a single platform for all are mostly for the devs; but when devs are happy, the benefits trickle down to users, too, who have a much more vibrant ecosystem. This is exactly what MS is doing with Win 8/RT.

Would also be great to see Ubuntu support running Android apps! Lets have the best of both worlds!

Let me attempt clarify this further:

Wayland does not entirely replace X11. You can definitely continue running an X11 server that renders to your local Wayland (it is a client of Wayland). The challenge is that this server is not ready yet. it needs to be programmed essentially from scratch, and include all the weird parts of X. But, the xwayland project is already running pretty nicely for the main parts of X11 that most toolkits require. Check out the demos!

Toolkits such as GTK+ and Qt already support pluggable rendering backends (GTK+ even has a cool web backend!), so they would just need to create a new one to support Wayland. (It would be MUCH simpler than their X11 backend.) But, even if they don't support Wayland, they would be able to continue working through their X11 backend and the xwayland client.

Bottom line is that Wayland (once xwayland is finished) will have only advantages. The toolkits already support X11, so all apps written for GTK+ and Qt will continue to support network transparency. But, when running locally, they will have the option of using a Wayland backend for much improved throughput.

Thing is, what if people write applications without GTK+ or Qt, that use Wayland directly? These, of course, will not be able to enjoy network transparency. But, writing applications like that is highly specific and likely the coders will be doing it for a good reason.