Again, Linux is not an OS
Reading Free Software Magazine's "Google Chrome OS. Or, how KDE and GNOME managed to shoot each other dead" I yet again see the ages old complaining about there being two major desktop environments on Linux and how bad this is. In another article, "10 Things I Hate About Linux" at least two reasons named have to do with there being too many choices, namely, the user having to choose between all the good available distros and the application vendor having to choose between those same distros for full support.
I agree these are issues if you look at Linux as a whole, or rather, all Linux operating systems, as a single thing. If you do that then this favorite "single" operating system of yours has a terrible case of multiple personality disorder and has had it from the beginning. This multiplicity is probably the one thing people most complain about when talking about why Linux hasn't had as much success on the desktop as some had hoped. Yet Linux advocates would readily say that it's also its biggest strength. I no longer am such a Linux advocate. I agree that an operating system which doesn't have a coherent identity is ill poised for winning in the market. This multiplicity is also the major source of all the complexity that people find to be superfluous.
But don't get me wrong. I am not siding with Tony Mobily's call for one desktop environment or any other kind of panning down of choice. Instead I'm calling for a different approach.
This brings me back to something I wrote earlier this year: Linux is not an OS. Besides the typical point that Linux is just the kernel my basic point was that what we typically call "Linux" is not really a single coherent operating system, but rather a framework for developing them or an ecosystem which spawns them. I instead opt to call specific distributions as operating systems rather than all of Linux, whatever that may include.
What other choice do we really have other than sticking to our "Linux with a personality disorder" paradigm? Do we really go and unify KDE and GNOME somehow? What about all of the other window managers and desktop environments such as Xfce, LXDE etc.? Where do you draw the line? I find it rather silly to still be complaining and calling for such unifications. That simply doesn't make any sense and will never happen. It's in the nature of Free Open Source Software. The only way in which that kind of unification would happen is if you had some big corporation somehow buy up all this Free Software and become the de facto "owner" or steward to all that is "Linux". They would then be calling all the shots when it comes to KDE, GNOME, LXDE, XFCE, X.org, Linux kernel, GNU toolchain, you name it.
Of course, that wont happen. So what's the point of calling for unification or complaining about the fact that it hasn't and wont happen? It's like complaining about the electronic components store having too many components doing the same function because you keep insisting on calling the electronics components store a single electronic device of some kind. If you actually want a device and want to sell that device on the market then you better choose your components, build the device and sell the device under its own name that designates only it and not the components store including all the other components which are available in it.
That illustrates quite perfectly the problem I see with the currently prevalent view of "Linux". You have many many software components developed independently among which there are many choices for the same function. These components are taken to build specific kinds of distributions yet there are still so many people who insist that we have to talk about all of these components at once as a single OS instead of talking about these specific distributions as operating systems and this "Linux" software components store as simply a repository of parts.
This shift in thinking removes the "too much choice" problem and all that is associated with it. If you want some "Linux", or Linux based operating system, to succeed in the market and you feel that "this and that" is necessary for it to do that then advocate that and make a distribution that matches those requirements. Then instead of marketing all of Linux while providing that distribution, market that distribution as an OS in its own right. After all, isn't this precisely what Google is doing? They don't market Linux, but Android. They don't market Linux, but Chrome OS. Notice the "OS" in there. They probably couldn't care less about this mythical notion such as the "Linux market share". They care about Chrome and will without a doubt care about Chrome market share.
Talking about "Linux market share" is actually kind of silly because you can't even define Linux as a single coherent thing. It's multiple disparate things at once.
To come back to the GNOME vs. KDE non-issue what I would suggest Tony Mobily advocate is this. Forget about Linux. Think about a perfect operating system which has all of the requirements you feel would win with Google and other major companies or the desktop market as a whole. Forget even about this mutual compatibility between different Linux distributions. It's largely a myth anyway. I'm quite vary of running even Debian deb packages in Ubuntu let alone OpenSUSE or Fedora RPMs. For all intents and purposes these are different operating systems.
If you think a single desktop environment is necessary for an operating system to be marketable then pick one and focus solely on that one. If you pick KDE for instance then forget completely about GNOME, Xfce or any other. You don't even have to ship it in repositories if you even have repositories. If it's not a part of your OS then it's not a part of your OS. Make a fine line between what is supported by your OS and what isn't.
Now I know many people will, again, protest this view because they really for some strange reason seem to like this fuzzy and incoherent idea of what Linux represents. It's supposedly at the same time a single product and many products at once. Yeah, sure.
But you will fail. Ideas at war with themselves seldom win.