With an estimated 850 million PCs installed worldwide, it's not surprising the mantra continues for displacing Windows with a Linux desktop. But is this a commendable objective anymore? Microsoft undertook a lot of dismal software engineering and even legal run-ins with the antitrust police to establish 90% desktop penetration. Linux application developers have better things to do than build a desktop that personal computer OEMs would never adopt anyway. Furthermore, even Windows is arguably not viable as a future desktop platform for Microsoft. Microsoft cannot afford another five-year release cycle like the one it took to launch Vista.

The late Peter Drucker, in his book entitled The Effective Executive, outlined advice fit for any decision maker, whether business executive or open source advocate. Drucker stated "Effective executives build on strengths - their own strengths, the strengths of their superiors, colleagues, and subordinates; and on the strengths in the situation..." It raises one question in particular: What are the strengths of Linux and the open source community, and is a Linux desktop merely an illusory solution to an interminable problem?

All industries change over time and the software industry is no exception. As change occurs, some old ideas remain. FM didn't kill AM radio. Satellite didn't kill cable television. Polyester didn't kill denim. PC's didn't kill mainframes. Likewise the Windows desktop won't be replaced for a long time, maybe never.

Taking a strategic approach to the desktop, what do people dislike about Windows? What is there to fix? They don't like security flaws. They don't like re-inserting their legitimate CD just to help Microsoft fight piracy. They don't like broken registries. They don't like duplicate inboxes and duplicate emails. They don't like incomprehensible workgroups, shared folders, or circular inline help. In other words, people don't want to be annoyed by Microsoft oversights or bad product marketing. But the Linux desktop does not eliminate annoyances; it just creates a different set of them, including user interfaces from the Stone Age.

Moreover, the benefits of a Linux desktop just aren't substantial enough. The savings aren't large enough. The reliability isn't demonstrably better. The functionality is not material or even acceptable sometimes. Half the time, even a browser won't work correctly on a Linux desktop. The only real advantage of a Linux desktop is, in fact, the psychological and economic satisfaction of denying Microsoft your money. Today, a Linux desktop is a misguided investment of time and effort. It will not win mainstream acceptance any more than a replacement for the QWERTY keyboard. It's just not a priority for the vast majority of Windows desktop PC users or administrators.

Rob Enderle of the Enderle Group believes that Windows in the PC world probably has an unbeatable advantage because the current world was designed around Windows, and Linux is simply too different to easily fit into that world. "Much like Windows wouldn't work in a mainframe world, Linux needs a world that favors its design elements to be successful on the desktop," says Enderle. "Its best play for the desktop would be in an embedded product that replaced PCs with something even more attractive."

So instead of a politically popular Linux desktop, where should open source proponents aim today? What should basically matter, and where greenfield opportunity lies, is in leveraging the freedom open source brings to application development, especially Web applications and embedded appliances. Open source will never make a difference by providing a productivity suite. Why? Because it will never be more productive. It will likely be less productive.

On the other hand, the world is moving strongly towards hosted applications in many desktop application markets (excluding graphic, audio and video editing applications). Software as a Service (SaaS) offers new features transparently and incrementally. Google just introduced a suite of e-mail, calendar, and phone services it will host for small businesses. Word processing and spreadsheet applications for businesses will follow. Other SaaS companies successfully deliver CRM, accounting, payroll, and more through the on-demand model.

Many online service providers like Google have wrung inefficiency out of server management. In the future, virtualization of on premise or hosted software appliances will spread further, making corporate IT environments more efficient too. For instance, rPath makes it easy for application vendors to create VMware images of open source applications. Regarding Linux as a zero-cost resource, wasteful efficiency is gained by merging the free operating system with each application and running them on a virtualized server. This is far more efficient than the administrative cost of managing application dependencies across one or more versions of an operating system and on multiple servers.

Microsoft sees the handwriting on the wall. Hosted and appliance-based applications running on Linux servers are popping up like daisies. For text-based applications, there is a strong bias towards SaaS by virtue of the commoditization of hardware and software and improving bandwidth availability. SaaS applications can be developed quickly on relatively small budgets. The future of desktop productivity is the SaaS model of Ajax-type application delivery in a self-configurable browser where a world of data coexists.

Unlike today's Windows or Apple desktop, SaaS is not one size fits all. Instead it allows quick configuration of the application and your browser as you desire. There will always be a need for client software to manipulate bits - games, photos, movies, music. But numbers and text strings may be ultimately better pushed to you by more secure, backed-up, SaaS application providers.

Browser-based applications are a field of dreams for developers of applications and plug-ins. New services rich with data and APIs continue to flourish and perpetuate. Consumers are not just promised, but are being inundated with incredible new choices for buying, selling, messaging, researching, and so on. "Ajax and rich browser technologies are unlocking the applications from the operating system," says John Robb, VP of Marketing and Product Management at Zimbra. "Users on Linux, Mac, and Windows desktops are now seeing a democratization of the desktop where everyone has equal access to the best applications."

Interoperability is key. Users want to configure their browser and desktop with application tabs and icons, and easy shortcuts to online data services. In fact, SaaS-related content business opportunity will outpace the significance of SaaS productivity applications. With a click of a mouse, you will log into a favorite online database, add or extract data, and get right to the task of analysis. The traditional productivity suite looks like an empty house compared to the teeming content from constantly updated SaaS applications. Great SaaS applications encourage content contributed from the outside. Think of Amazon and eBay where every buyer is potentially a seller. Maybe ticket vendors will someday truly adopt a free marketplace for frictionless transactions like the day-trading features of E*Trade or auction environment on eBay. What an exciting way to rent a car or reserve a hotel room that could become!

SaaS also solves a lot of intellectual property issues, chiefly piracy. Microsoft has stated that Microsoft Live must succeed in Asia. It must succeed there because Microsoft has little alternative. Windows and the Office Suite are largely pirated in Asia, so Microsoft has a problem with the traditional licensing model. Hosted software by comparison has absolutely no piracy issues. Intellectual property protection of hosted software is immediate; the SaaS provider simply terminates the subscription if payment is not made.

For open source to make a difference on the new SaaS desktop, think in terms of open APIs, repositories, players, and plug-ins. Offer online tools for migrating away from the Microsoft client and help users get onto the SaaS paradigm that runs predominantly on open source servers and appliances. Concentrate more on best-practices for online service inoperability, and grow consensus around open standards. Create and embrace online data - even the simple dictionary or thesaurus, because no one wants to reinstall the thesaurus feature on a Windows desktop. We all have deadlines to meet.