I'm a huge Linux enthusiast, bigot even (self admittedly), and I've been using Linux since 1997 - even contributed to major projects like XFree86, KDE and helped with some things in the Kernel - waay back when. I've also been an IT executive for years dealing in industry and am now a CTO. (just background)
Every year since IDK KDE pre 1.0 was released, we've had "Is this the year of Linux on the Desktop" articles, hopes and dreams. The reality and good part is that Linux is already on the desktop - so we're here.
Also the reality, and the bad part, is that until the Linux community deals with the extreme level of fragmentation out there, and fixes some long standing issues like app packaging and distribution for Linux - with 1 solution everyone can agree on, Linux on the Desktop will never be more than a fraction of a percentage of the marketplace. It's a horrible shame that the LSB failed - that's what they tried to do, but unfortunately with a community driven by engineers, you have NIH syndrome to the extreme. We need one base OS layer, with a common way of doing things, common fs layout, common init system, common security mechanism, etc. i.e. Pick a distro, pick a package format (for the underlying OS), and - to a lesser extent - pick a desktop. When that happens, then Linux will start to make more serious inroads - but it will take 2-5 years to gather momentum.
I know I'm going to get flamed out of existence for the above, "you don't need to do that, or this, or whatever" but the reality is that software and hardware vendors will never spend the time and money to invest in supporting the ridiculous number of permutations of Linux on the desktop. Its just not economically feasible. Anyone who's run an open source project knows this. Packaging, distribution and then user troubleshooting and support is horrific. When I developed and ran LinEAK (lineak.sourceforge.net) in the late 90's and early 2000's I refused to touch packaging. Ain't nobody got time for that. You see that today with Game Devs - Linux is 1% of our revenue and 99% of our support tickets. Did anyone watch LTT try to go Linux? - hugely problematic for a guy that most mainstreamers / i.e. non-techy - would consider extremely technical. And don't tell me your grandma uses Linux, good for her - I'm sure all she does is use a web browser. She's not the mainstream - she's grandma, and I don't care how easy it is to install. How many non-techies you know install their own OS? None.
Until the economics work out for Linux on the Desktop (from a corporate perspective across the industry) and the customers are demanding (and paying for it.) - This is a Catch-22 btw - it's just not happening. Until you can fix some of Linux Desktops' major problems, it will *never* be the year of the Linux Desktop (i.e. when Linux becomes common or mainstream.)
Right now, Linux is the same as Java in 1998 - Write once, debug and support everywhere. The only way I see *any* of this changing is if Linus Torvalds himself gets involved, cracks the whip and starts a concerted effort to bring the distros together to fix their problems. That just ain't gonna happen though - Linus is an engineer and that's all he cares about. Unfortunately he's the only one with the respect and personal gravitas to make it happen at this point.