Call me cynical, but I bet even if the Ubuntu desktop becomes measurably more "usable" than OS X or Windows, people will still use OS X or Windows because they have already learned how to work around the unusability of those OSes. At some point, people don't know what's usable, they just know which buttons to press to get what they want.
However, if we make Ubuntu easy for new users to pick up, and we make sure new users (read: kids) have Ubuntu, then we'd be on to something.
(BTW, I call it "the Ubuntu desktop" instead of "the Linux desktop" because the latter term is meaningless. My definition of the Linux desktop is xmonad, emacs, urxvt, firefox, and amarok. I doubt many other people have the same taste, although I advocate it to anyone that will listen.)
I think as time goes on computer manufacturers will make slow movements towards it to avoid paying the 'Microsoft tax' particularly on low-end computers where Linux thrives. If computer manufacturers want to use it hardware vendors will be forced to support it. Once there's a sufficient market-share and the environment is better set-up to support it commercial software development will start inching towards it.
This has been a big struggle so far as there's long been a catch-22 that not many people use Linux and it's difficult to use Linux because no-one supports it as not many people use it. The cycle is slowly but surely changing - and already has happened to a degree with ultraportables - but expect it to take 5 to 10 years for it to get a decent amount of market-share as it continues to improve. I'm not saying it will ever 'take over' Windows, but it only needs 5-10% market share to be taken seriously - much like Mac OS X - and it's already got a highly dedicated user-base that's more likely to grow over time especially as there's as thanks to ultraportables there is now product that can be used to expose people to it.
What didn't help was the highly fractured community, something which has significantly improved as of late largely thanks to Ubuntu; over-idealistic "free software" evangelists who didn't like seeing the bigger picture; poor user experience, something which has changed significantly with text files being the exception rather than the rule; traditionally poor software support that's slowly but surely changing thanks to web-based apps, a wider selection of native applications, improvements to Wine and Microsoft's move towards .NET, and hopefully in the long-run greater, better and easier commercial applications; the lack of support by OEMs but this has largely changed with Dell, the biggest computer manufacturer in the world, slowly but surely moving towards Ubuntu; and the "Linux is hard" meme that perpetuated thanks to overenthusiastic highly technical users a decade ago which made Linux a dirty word.
It's very easy to forget that many of the variables that made alternative OSes getting a reasonable market-share impossible even 10 years ago have changed beyond recognition.
I contend that Linux's big market share opportunity will come not from conventional PCs no longer being sold with Windows, but from the erosion of the conventional PC markets. Nobody wants a big box of wires in their office, and I contend that even laptops are too big. I believe that the future involves smartphone class devices and accessories to make them more useful for present-day laptop type duties.
Lots of professionals are already issued BlackBerries and their ilk, and current/near future phones are shipping with 1G of system RAM, and 16-32G of flash. Give me better productivity software and remote desktop over bluetooth, and I'd kiss my ThinkPad goodbye.
This is interesting - For me what will be the fascinating to see is how they tackle this from technical, community and business perspectives.
At the very least will give some insights into the challenges on increasing usability in Linux (I find it hard to find two people who agree on how "usable" Linux is, let alone on the challenges) - and at the best introduce some significant improvements.
