I think Ubuntu may be trying to move too fast. Moving fast is great if you can pull it off, but it's not worth breaking the basic functions of the OS to get a more flashy UI. If Ubuntu does want to copy Apple, there's one major thing they need to learn: Apple releases features when they're done.
I know well and good that Ubuntu was conceived because Shuttleworth felt that Linux was ready for the big time and that Red Hat etc had too much penchant to devote resources into low-level bickering that ultimately had relatively little effect on your average end user instead of focusing on improving user experience, but now that Ubuntu has moved the user experience so far forward, they should reconsider that mission. The places where Linux is most lacking is low-level compatibility for things like fast 3D acceleration and power management.
Canonical should use some of its funds (aka "Mark Shuttleworth's money") to buy the 100 best selling laptop models each year, set rigorous testing standards, spend six months developing a new release and then take however long is necessary to make sure that everything passes on the last three years' best selling laptops. In this process, they should not be shy about contribution to X, kernel, etc., and should distribute patched versions of these if necessary to get compatibility.
That, combined with Ubuntu's user experience work, is what will really make Linux a completely viable desktop computing platform. Far too often things break between releases and/or upgrades.
I think you made my point more succinctly than I did.
I think the UX was just fine in last year's Ubuntu. You could argue about whether Gnome 2.x was as slick as Windows 7 or OS X (I think it was at least better than Windows), but I don't think there's much doubt the system was usable by non-geeks. The first thing to break for me was suspend/resume, with one of 10.10's kernel updates. I reverted to an older kernel and hoped 11.04 would fix the problem. It didn't, and it precluded running the older, working kernel. Performance also got worse, and I can't think of any noteworthy improvements as I didn't consider Unity ready for prime time.
So then 11.10 came out. Reviews said Unity was great now and everything ran smoothly, so I pulled the trigger on the upgrade. Unity did, in fact mostly work, though it was slow and glitchy. Oh well, back to Gnome Classic. Of course, it's Gnome 3 now and I can't even move the clock. That won't do, but I've heard the new Gnome 3 gnome-shell is awesome, and I have a video card that can handle it. It loads, slowly, but UI components sometimes vanish when I try to interact with them. Eventually, X crashes. Oh well, that gives me an opportunity to see what progress KDE 4 has made. I can report that the error messages for Plasma crashing look like they've had some attention from a designer since the last time I saw them. Good work.
I'm running Linux Mint Debian Edition with Xfce now. Still no suspend/resume, but everything else works. The Linux desktop experience is almost back to where it was two years ago. Yay leadership!
This has to be one of the worst UI decisions I've ever encountered...
> that's completely obvious right?
I recently tried to upgrade my older 9.10 install to 10.4 LTS. I got a bunch of errors about x.org upon upgrade (I think maybe because I downloaded and installed Nvidia's Linux driver a while ago), and yep you guessed it - hosed system upon reboot.
These sort of showstopper problems should not be occurring - not in the year 2011. Totally inexcusable as far as I'm concerned.
As an example, if Ubuntu had instead teamed up with Adobe, Mozilla (for Firefox), and Google (for Chrome), and shipped a fix for the Flash problem that's been around for years (referring to Flash's instability on Firefox, Chrome, etc.), I think they would've pleased far more users w/a much more subtle change than the massive UI revamp that is Unity.
Or they make an unholy alliance with one or two manufacturers to ensure (NOT excluding other mfgrs, mind you) top notch compatibility on a high end, mid range and netbook class unit.
That's it. They would do so much better like that.
Releasing early and often is good. Pushing that software on people who didn't choose it is not. I feel like there should have been a "Try Unity" button at the Ubuntu installer so people like me could just wait until it's ready.
Maybe I'll be happy with Unity in two years. But in that time I'll have to suffer stuttering, lag and the whims of armchair UI theorists who think Alt-Tab should jump workspaces by default.
I actually bought a Mac on Friday. I need a development environment that works. I don't like spending time yak-shaving when I could be solving real problems.
What's wrong with that? Just provide a standard install location and let those that want to utilize it do so.
As an avid Mac OS X/Linux/Ubuntu user, I really truly don't get the direction Canonical is heading with Unity.
I almost want Ubuntu to either abandon the traditional Debian subsystem or have a "Power User" configuration where you can tell it to use /etc files over the GUI config.
Debian testing/unstable works just great; I keep meaning to move toward it instead. Though I really do feel that Ubuntu packages come better configured and ready to go. None of this has bit me in the ass yet.