One of the most commonly cited reasons that the business world would stridently resist moving to any non-Windows platform is that of familiarity. Windows may be difficult to learn and frustrating to use (as some assert), but everyone in the business world is already familiar with it, and perhaps more importantly, familiar with the look and features of the commonly used Windows applications, like Office.
It's a source of ongoing contention, but many Linux-on-the-desktop advocates have chosen to respond to this issue by producing Windows-lookalike desktop themes and Office-lookalike applications. This provides an easy transition for corporate desktop users while opening them up to the wonderful world of Free Software. I think it's a perfectly reasonable step, but opponents warn that one of the reasons we all want to use alternative OSes is because we don't want our computers to look and act like Windows machines. Too much effort spent on making Linux work just like Windows takes away valuable effort from making it be better at being different.
I think it's an overblown issue. I've worked for many companies big and small, and it's clear to me that the majority of corporate computer users only have the slightest grasp of computer technology in the first place, and just barely know that you open the word processor by clicking on the big W icon. A company would benefit more my having a simpler standardized desktop that's managed centrally and give the user less ability to screw their machine up by downloading spyware-ridden doohickeys from the internet than it does by having a system that's familiar to the user. People can learn a new word processor or a new file management paradigm pretty easily, because most of them don't know the Windows one all that well anyway, and those that do are probably pretty clever. I won't get into an argument about the ROI of various platforms here, but I think that a little retraining won't kill anyone and can be justified with lower system management costs.
That being said, I think that there's very little chance that any alternative OS will make big inroads into the corporate desktop in the next few years. Corporations are conservative and can be short-sighted. They've made their commitment to Windows, and they'll stick with it.
The Herd Mentality (No, not Hurd)
The corporate mentality brings up an important issue that's central to the alternative OS vs. Windows debate: market share. People hear about Windows have ninety-something percent market share and they wonder how anyone can manage to get by on the table scraps of the few percent that's left over. How can software developers make money developing for Mac or Linux, or (shudder) some other OS with less than 1% market share?
The reason that the corporate desktop is an important consideration here is that it's important to remember that most of the PCs in use today are used by corporations, and a great many of them are in use as single-purpose terminals, such as in vast banks of PCs used in call centers for order entry or banks of screens displaying market data in trading rooms on Wall Street. Why they're using a complicated and hard-to-administer full-featured operating system for such tasks is the subject of another essay, but they are. These computers are single-purpose terminals, sometimes with only one application installed on them. Most of the rest are identically-configured Microsoft Office workstations that have taken the place of the IBM Selectric typewriters that used to grace most every office desk once upon a time. It's bad news for Apple and Red Hat that they're probably from Dell or HP and mostly running Windows, but it's not entirely accurate to include them in all of the marketshare numbers because these computers are not candidates for 3rd party software use.
If I develop an application that allows people to manage their genealogy or make and print-out templates for scrapbooking, is each of those PCs on Wall Street a potential computer that my application will be installed on? No. The percentage of Macs, for example, on the desks in the dens of people who might want to look up ancestors or make a scrapbook may be much higher than the 4% that the analysts report.
Furthermore, most of those numbers are based on easily-collected information, such as how many PCs (with a Windows license) sold that year. Analysts are as lazy as anyone else, and they don't have access to magical data. So you add up all the PCs sold that year, compared to all the Macs, and compared to the handful of computers sold with other OSes pre-installed, and you have your marketshare numbers. Voila! How do they know how many people bought that PC (which you can't really buy without a Windows license) and then installed Linux on it, or even DOS? They don't. And those numbers only tell you how many of the new computers bought today run Windows. Truth is, home users keep their computers longer than companies do, and Mac and Linux computers can be used with current software much longer than Windows ones. So how do you measure how many out of the total number of computers being used today, regardless of when they were purchased, are running a particular operating system? It's really hard.
Let's not forget that the PC market is very big now. Even though at one point Apple was selling, let's say, 8% of the new PCs in the world and is now only selling 4%, it's still selling a lot more computers now than it did then. If it was possible for Apple to have a thriving and prosperous developer community back then, it's still true today, more so. For old OSes that people cling to like Amiga or OS/2, sometimes the developers stick with the platform becauses while the userbase may be small, there's no competition. Competition among Windows developers is fierce, so if you can make a living without it, it's a good life.
So yes, Microsoft indeed has a powerful monopoly on PCs, but there is actually plenty of room for an alternative platform to thrive, but it must attain and maintain a critical mass. If Wal-Mart sells enough Linux PCs over a few years, the desktop Linux market will have grown enough for software vendors and hardware makers in that niche to take notice. If Apple can keep being from screwing things up too bad, then the platform will stay viable.