> I happily run Linux on the desktop and have done so for years.
> market share does not necessarily reflect quality.
Sure, but I'm afraid it can have a some impact in life expectancy. Put another way: the traditional PC/desktop/laptop marked looks more and more closed each year. If _could_ happen that, if we do not manage to position a free OS as a viable option on the desktop, it will close just too much. For a taste of what I mean, read all the articles about secure booting.
But not only that. Hardware is also becoming more and more complex. So complex that developing drivers by reverse engineering will eventually become a non option. For the time they would become usable the targeted devices will be obsolete. The noveau example is a hint of that.
Additionally you have the problem of software. There's not much independent development for an irrelevant platform. Case in point: id software latest game engine (it's not getting ported to Linux), but also the latest Firefox Apps movement. Also, there are no paying jobs in developing software for a OS nobody uses. If we developers want to keep developing software for Linux, somebody has to want it. Someone has to pay for it. Also, life is much more difficult when your bank and your city refuse to serve you unless you run the same proprietary OS everybody else uses.
Finally, there's the question of being useful. Is the goal of a free OS to be useful in general or just for the developer?
To sum it up: being a relevant market on the desktop could be critical for Linux (and free OSes in general) future.