Specialization and the Linux Desktop
Our benevolent dictator for life recently claimed that he was still aiming at Linux being as prevalent on the desktop as it is in the datacenter or in the cloud. The statement was meant with roaring applause from the crowd, and a few healthy, and a few not so healthy, doses of skepticism from the press. Recently, IT World asked “Does it still make sense for Linus to want the desktop for Linux?”, and Matt Asay from Tech Repubic asked “Can we please stop talking about the Linux desktop?”. Both publishers are critical of the claim that there is still room for Linux on Personal Computers, and point to Android as a Linux success story. What both articles miss though is that the flexibility of Linux, and the permissiveness of its open source license may be the thing that saves Linux on the desktop, just not in the way we were expecting.
Computers are fascinating tools, they can be so many things to so many different people. In fact, there are few occupations today that are not enhanced in some way by use of a computer. Because of this, the two biggest companies in the industry, Microsoft and Apple, take a generalist approach and build operating systems aimed at the majority of the people. As we’ve seen, for many use cases this approach works well for them, but it’s also the approach Linux has mimicked for the past twenty years and we’ve not seen the same success. Clearly we need to take a different approach. Most suggestions on how to make the Linux desktop more appealing focus on making the user interface “better,” or making it easier to adopt for new users. I suggest we take a different approach and leverage the strengths of Linux to build specialized, highly focused computers that appeal to specific user groups.
My favorite example for this is scientific computing, because all of the building blocks are already there, and many scientists are already using a significant amount of open source software for their research. While much of the software is geared towards OS X, since Apple takes a generalist approach to the design of their software, they are not intentionally catering towards scientists in the same way that a dedicated Linux computing platform could. Imagine a machine designed from the ground up for data analysis and report generation, with built in hooks for open research sharing and publication.
Or, consider a system designed with professional grade, Hollywood level film production in mind. Linux could be the base of a reliable, open, dedicated system for the incredible amount of computational power that creating a modern blockbuster requires. Even if the application for actually dong the editing comes from Adobe, perhaps they build on Linux and release their own dedicated system, much like Steam is doing with their Steam Machines.
Big businesses that are tired of riding the complicated Microsoft desktop train could get off at any time with a Linux desktop distribution built for virtual desktop infrastructure and long term stability and support in mind. Assuming that most business functions and internal applications move to web based systems anyway, pushing out a Linux desktop running a modern browser would make perfect sense for many companies.
In each of the examples given above, Linux is bent and shaped exactly to fit the task at hand. Instead of presuming to be a general, consumer computing platform, it assumes that it will need to be modular from the start, and provides the internal building blocks to make this possible. The Linux desktop doesn’t need us to stop talking about it, it simply needs us to, ever so slightly, change the conversation.