Why Linux is a desktop dud

With Windows, the operating system is just a start; you must add applications to make it functional. Many Linux distributions provide a desktop look similar to Windows and include an extensive assortment of applications, programming tools and games.

Installing Windows and sundry applications can take most of a day. Contrast that with Linux, where the process typically takes less than 60 minutes.

Windows and its applications are expensive and require costly upgrades. The cost of a Linux system is nominal, and you can download a vast number of versions for free. If your favorite Windows program hasn't been ported to Linux, software is available to let you run that application from within Linux.

Windows is limited to the offspring of the IBM-PC, while Linux systems can operate on virtually any hardware platform.

Unlike Microsoft products, when you share Linux software licensed under the GPL, you are not committing a crime. And, if a feature is missing or broken, a person with the ability can access source and modify the software.

With Windows, knowledge is hoarded and kept secret. With Linux, the exchange of knowledge is promoted.

With so many positives, why aren't desktop users flocking to the Linux banner?

Once you get beyond the turnkey desktop, the learning curve required to tweak your system can become formidable. Also, GNU/Linux is in the same rut as the rest of the industry; it does not relate to the power user's need for better office tools.

Hackers may brag about how easy and stable Linux is--and they would be right: Once Linux is set up to accommodate a specific computer user, it is easy and safe to operate. But it is a different story if you are a new Linux user trying to customize your system. Without access to those experienced with the operating system, such tasks like changing the screen resolution, setting up a home network or adding a new program often become unnecessarily involved projects.

Providing permanent solutions to the problems people encounter is less a technical obstacle than a commitment to refining the desktop. In addition to the programming effort, this will require a more effective communications loop between hackers, developers and users. In the meantime, a step-by-step guide is needed to help new users of the operating system solve common problems.

Office software remains stuck where it began. Development has cycled twice (mid-1980s and the mid-1990s) between a preference for standalone applications and integrated office suites. Neither is optimum and both are offspring of early applications, which artificially compartmentalize tools into separate functional work areas.

The upshot is that office software has grown more useful for generating elaborate documents than for providing tools that help efficiently organize and manage one's work process.

Professionals want to improve their productivity, but is the industry listening? We can continue upgrading through the process of diminishing returns, or we can actually solve our problems by improving the office toolset. Taking advantage of software progress over the past 25 years, we can develop tools that offer improved capabilities for people.

While several projects are creating components for developers, a project to develop modular tools for people is not on the horizon. The GPL community could take the lead by modularizing the tool capabilities found in office suites for use in a more integrated and process-oriented work environment.

With separate tools, functionality can be clarified and learning curves minimized. People could customize their workshop to include only the tools they need and to perform the same work that's currently done working with office suite applications. Tools could be added or upgraded one at a time. And new open file-standards could also evolve. But most important, a process-oriented environment would enable users to set up systems to manage their work.

GNU/Linux can flourish in the enterprise sector and still be a desktop dud. Though opportunity exists, the market won't abandon Windows except for something significantly better.

Currently, the desktop is more of a low-budget sideshow, with the enterprise sector being the main event. This can change, but it will take an aggressive new approach that provides for both the necessary revenue and a mechanism to establish better relationships with computer users. To refine the user interface and evolve more useful tools, development must move beyond creative cloning. To prevail over proprietary systems, it must take the lead in providing both ease of use and productivity for the desktop user.

Don Soegaard, a graduate of U.C. Berkeley, is an engineer with a specialization in operations research. You can find more of his writings at his website www.linuxwillprevail.com.