I'm conflicted about getting Linux to live on the mainstream desktop. Let's for the sake of simplicity assume that Ubuntu is the platform willing to become mainstream.
Here are some of the key issues I see with the current Ubuntu distro:
Unification: Linux at the moment fills a void for people who find value in it that they don't already get from either OS X or Windows. It's very technical and bars noone from tinkering with whatever he likes. Sadly, Joe Citizen breaks more than he fixes by tinkering, and usually wants to use his device to do some predefined task. Say they want to adapt Ubuntu to get more market penetration, wouldn't that risk moving it so close towards Windows and OS X that the original benefits disappear? (e.g. losing the techie crows by dumbing it down). But at least for Unity they have taken a page out of the Apple playbook. No more sudo unless you root your computer or you get to at least click "I know what I'm doing, leave me alone!".
Attention to detail: Application quality control, either it's 100% finished and polished or it shouldn't appear in the 'main ppa'. If people want to experiment, let them hook into experimental ppa's but assure that Joe Citizen has access to applications that have been tested over and over, and provide him with what he needs.
No overcompensation: While GIMP is a wonderful tool and has a huge toolbox at its disposal, it offers too much to be efficient. So much even that it's not unlikely someone will not understand how to do something and look for something simpler with less options, but a clearer 'clicking-here-does-that' approach.
Agression: Linux isn't being marketed aggressively, at least not in the commercial sense of wooing people to leave their current OS for it. As I see it, it's something techies pick up by reading up on technology, and finally decide to give it a try. But there's no driving force to push for said aggressive market penetration and as it stands, there are still quite a lot of drawbacks that put Linux at a disadvantage vs OS X and Windows. But "more features" is not the answer to everything and doesn't ensure success. Ballmer was laughing at the iPhone when it yet had to be released, and said the WinMobile devices did everything the iPhone did and more. I don't think Ballmer's laughing now, though.
Litigation: While the perception of Linux on the desktop is seen as small and insignificant, litigation isn't likely to occur in the domain the user-oriented Linux distro's. But seeing what Android already had to endure and what kind of shitstorm only recently appeared on the horizon with the forming of Rockstar patent troll (Apple and Microsoft even joining forces!), I'd be very wary of any Linux project ever getting a substantial market share. It'd become the target of frivolous patent suits in no time, and while it may be hard to find something to attack it with, lawyers can get very creative (Exhibit A: the US patent system). Also, many corporations might pull their patents (Mono from Microsoft) once their own products are threatened.
As both Apple and Microsoft already have their own desktop OS, I could potentially see Android bringing in a breath of fresh air. If Google really wants to, they have the power to address the above points. While I don't see where they were going with Chrom OS, I could see Android OS becoming something viable, as the worlds of mobile and desktop seem to be headed on a collision course (seeing recent Windows 8 and IOS developments).