Editor's Note: RIP: Linux on the Desktop
Build on successes; don't lament the past
OK, it's official: Linux on the desktop is dead.
Oh, sure, the corpse is still kicking around a little, and there's some debate about whether the patient is officially dead. But all we're waiting for now is the official death certificate from the coroner.
The recent death of Eazel, as well as Corel's inability to practically give away the desktop-oriented Corel Linux, are two strong indications that there isn't an acceptable rate of return on Linux desktop development. (Plus, a leading Linux distribution that put a lot of resources into desktop Linux will shortly be announcing staff layoffs and budgetary cutbacks.) Without investment, Linux on the desktop will recede to a less frenetic rate of development: the enthusiasts will still develop code because they believe in the code, but the less enthusiastic -- and those with demanding day jobs -- will disappear from the scene. Such is the way of all movements: either the professionals take over and the movement evolves, or the movement recedes.
I don't want to be gloomy about this. Many very talented people have put a lot of effort into Linux on the desktop, and there are some very good desktop environments, like KDE or GNOME, that are worthy of consideration.
But as it stands Linux on the desktop is not an entity that is usable by the average PC user when it comes to accomplishing their daily work. This has nothing to do with the quality of the desktop environment, but has everything to do with how PCs are actually used: end users don't use the environment, they use applications. And while someday we may evolve to a world where everyone's work is done via the Web and a Web browser, we're not there yet. Most computing work is done via third-party software tools independent of the operating system. The lack of usable software tools is really what will doom Linux on the desktop.
There are a lot of half-baked tools out there, and one of the problems in being a Linux user is the need to figure out exactly what tools are usable in their present release and which tools merely show a great amount of promise. Sadly, I'd submit that most Linux desktop tools -- like KOffice, like Evolution, like AbiWord -- aren't really usable in their present release. Linux on the desktop is weak because it just doesn't work well on the desktop yet for the work that the average people needs to accomplish. You cannot ask users to compromise when you want them to switch operating systems, and you need to be realistic about how things work. Let's face it: at the present time there's nothing under Linux that works as well as Microsoft Office. Period.
Compare this to the server, where there are a wide range of Linux tools that are usable in their present release: Apache, sendmail, Squid, Beowulf, perl, PHP, Zend, Zope, a range of high-performance file systems, et al. Linux is strong on the server because it works and works well on the server.
Part of the reason a column like this will cause a ruckus is that early on some on the Linux community set out for World Domination, and many Linux enthusiasts think that this is a doable goal. Guess what: it's not. It's not realistic to have one operating system dominate. This is a goal that Microsoft, backed by billions of dollars in investment, has not been able to accomplish. Yes, Microsoft controls the PC desktop, with some competition from Macintosh and Linux. But by and large Microsoft has been ineffectual on the server side: it has a decent market share of low-end servers, but it does practically nothing when it comes to midrange and high-end servers. Similarly, Solaris rules on the high end and fights it out with Linux and IBM AIX on the midrange. No one can claim world domination when it comes to the operating system.
It's not glamorous or sexy to rule a segment of the server space; just ask Scott McNealy. But it's important that we recognize how important this market niche is. It's a place where Linux can thrive, it's a place where future development is appropriate, and it's a place for Linux businesses like Red Hat or VA Linux to grow. World domination? That's the rallying cry of the past, not the future. For now, let's recognize the amazing accomplishment of Linux's server market share and continue to build on that success.