Linux's Desktop Growing Pains

It's long past time to stop talking about Linux as the hotshot new upstart, and to demand the same things from it as any other environment. That means no more excuses about what's to come, but results right now -- especially on the desktop.

This is about more than just Windows 7 coming along and eating desktop Linux's (free) lunch. It's about the fact that the Linux desktop has become more about an arena for defending various conceits about software rather than a place where present and future work can be done. It also means that in order for Linux to really advance its game, it may need to shed a lot of its current software baggage and start anew.

I'm not suggesting that it's impossible to do real-world work in Linux. Rather, I'm saying the way Linux is developed works against those needs rather than for it. Randall Kennedy at InfoWorld felt the same way:

And because there is no "Linux", but simply various implementations of the Linux kernel with packages of userland programs (i.e., distributions), it's all the more difficult to point to any one entity and say "Clean up your act!" Once, when I was in a particularly testy mood about the whole thing, I groused something like this at my computer: "Not only is no one at the wheel, there is no wheel."

This sort of broken-record issue, as Randall put it, is not exclusively a Linux problem. Far from it. It's not even a software engineering problem, but a planning and coordination problem. No software project is immune from it. It's just that Linux exhibits it more than most, possibly due to the nature of the attention drawn to it.

For any particular version of Linux to make headway on the desktop, it needs to be directed and planned and executed by someone, or a group of someones, who can take final responsibility for everything. Case in point: Ubuntu/Canonical. They're trying to do just that, but they haven't done it by starting anew. They've inherited both the legacy software and legacy development behavior of existing Linux projects, which are steered more often by engineering conceits than user needs. Their work has not been radical enough by an order of magnitude.

The most creative and exciting work done with end-user-facing Linux right now is being done by people who take its essence -- the kernel -- and leave behind as much of the legacy end-user material (X.org, GNOME, KDE) as possible. Moblin and Android come to mind. But I don't yet see anyone being bold enough to step up to the plate and start anew that aggressively on the Linux desktop.

InformationWeek Analytics has published an independent analysis of the next-generation Web applications. Download the report here (registration required).

Follow me and the rest of InformationWeek on Twitter.