What's next for Linux? There's no simple answer because Linux isn't a single entity but a galaxy of implementations and possibilities. The Linux kernel--version 184.108.40.206, to be precise--is at the center of it all, with the operating system continuously morphing into new shapes.
Two forces in particular are driving changes to Linux. One is a push to develop Linux into a standard, more predictable platform for developers and users. The other is to leverage Linux, and open source in general, as mediums through which innovation gets delivered to users.
Those who create Linux distributions have a massive task ahead: taking the myriad open source offerings out there and pulling them into a unified whole. That's certainly a goal for Ubuntu Linux, says Mark Shuttleworth, founder of Canonical, which provides commercial support for the popular Linux variant. Ubuntu 8.10 was specifically assembled as the next step toward unification. Among the improvements with this in mind: automation that lets users install unsupported or commercial handlers for multimedia codecs.
Such efforts make Ubuntu the lead contender for Linux on the desktop, but that's not the only place it's making inroads. A Ubuntu spin-off distribution, the "Netbook Remix," targets the emerging market for netbook computers, where Linux still needs hardware integration and user interface polish to be accepted.
The server market--Linux's biggest and most profitable realm--continues to mature and expand. Ubuntu, Red Hat's Fedora, and SUSE are the leading distributions, though dozens of others are actively supported and used, too.
In one key area, market leader Red Hat is pushing to make virtualization easier and cheaper, as a way of giving users an alternative to VMware and Windows Server virtualization technologies.
Shuttleworth's goal: Pull it all together
Photo by Sek Leung
Tying together the ingredients that go into a distribution isn't merely a matter of cosmetic polish. There must be guidelines for how the pieces lock into place, and not just for current distributions and projects, but for future ones as well.
The consortium that handles that most directly is the Linux Foundation, creator of the Linux Standard Base. The LSB isn't simply a guidebook, but a "living" suite of test tools for evaluating whether a Linux app is compatible across distributions. Given Linux's culture of freedom, no one is forcing anyone to follow the LSB, except in the sense that noncompliant work might not be as transportable across distributions.
The theory is that it will be a self-policing process. If your application doesn't work across Linux distributions, there's less chance it will be adopted. This is itself something of a limitation, but probably an inevitable one. An open development culture means more choices but less automatic standardization.
Another area where the LSB needs to make progress is to offer verification for applications written in a broader range of interpreted environments. A good many Linux applications aren't delivered as binaries, but run in Perl, Python, Java, and Ruby. The application-checking tools in the LSB only deal with binaries, Perl, Python, and shell scripts. The scope of applications audited through the LSB should expand with time.