It's not costs, lack of understanding, technical difficulty or ignorance of Microsoft's behavior. It is simply People. Unix, be it SCO, Linux, BSD, Darwin, HP-UX, AIX, Solaris, Tru-64 or OS X, has advantages that are clear to anyone who reads even a minimal amount of what has been published. Once installed and implemented, unix class systems are more secure and reliable than Microsoft systems, and handle large loads better.
Why then are so many companies failing to take advantage of this alternative? Resistance to change comes from the people involved, and this cannot be overcome by technical or cost arguments alone. You must address the core of the resistance, and like psychological counseling, bring that resistance to the surface where it can be seen and understood for what it is.
The way humans have evolved has created a built in resistance to change. Survival in the early days involved reducing risks by sticking to known and tested ideas such as how to hunt. Folklore was passed down generations to guide the young in life, and was generally successful. Had such a process not been useful, those people who used it would have died out, along with the ineffective knowledge.
This process works well except when big changes occur, then resistance to change becomes anti survival. This is where we are in the IT industry today. The current dominant force in IT is Microsoft. Never mind now how they got there, the big question is why they continue to dominate when clearly superior alternatives exist and are well known.
As the first paragraph says, the resistance is from people. They have bought in to the 'Microsoft Way' and have become accustomed to and comfortable with that environment. An objective look at the facts says Microsoft systems are not best in several areas. Analysis by cost, performance, stability or reliability clearly list the unix class of systems as superior.
The other old story, that unix has few applications, has been untrue for some time. Today, that roadblock to unix adoption has shrunk to a small percentage of applications. But companies continue to ride the Microsoft Way train. They choose to avoid the risk of change and unknowingly, accept the risk of being left behind.
Microsoft is the wrong train to be on. One or more of your competitors is already moving to unix class systems, reaping rewards in critical areas such as resistance to viruses, efficiency and operational flexibility. Open Source can provide tools for most of the functional needs in an enterprise operation, and many of these tools are already efficient and stable.
When a CEO asks the CIO about this, the response is usually "Yes, but..." where the reasons involve understanding the technology, which the CEO doesn't have time to do. The real reason at this level is that the CIO knows about unix, but has built his reputation by promising results from investing fully in the Microsoft Way. Admitting that Microsoft may no longer be 'The Way' is a loss of face that many executives can't handle.
Similar mental commitments have been made by every MSCE and most end users. Since these people don't pay for the systems, the only problems they see are reliability and stability. Based on experience with only MS software, they have come to see this as 'normal'. Users often blame themselves for failures of the software and believe that all software is like what they see. When faced with change, they remember earlier IT promises of how it will be better, and don't believe them. Thus everyone in the system, with rare exceptions, doesn't like change, doesn't want change and will resist it either openly or through subterfuge.
User's perceptions of what is 'normal' is one of Microsoft's hidden weapons against change. Since the user believes that change only brings new problems to be solved, they resist change. Given what they understand, right or wrong, their reaction is reasonable and predictable. In fact, one of Microsoft's biggest offenses against everyone is that they have reduced expectations of software by so much that executives and users alike no longer expect significant improvement, only a change of problems.
Given this mass of internal resistance, mentioning unix as an alternative is about as welcome as a bad cold at a fancy dinner. Why would anyone exchange known problems, which can be handled, for unknown problems? Why spend money for something new that may bring (will bring) unknown problems which may be harder to handle? No wonder unix gets less respect than Rodney Dangerfield.
How can this resistance to change be reduced? How can unix be implemented without having to mandate change over strong resistance? How can people's expectations be changed? The short answer is 'Not quickly'. Real change in a person's outlook takes time. In many cases, some face saving reasons must be provided or invented. The ultimate convincing factor is when change brings real improvement - reliability, cost savings and security. It will take time for users to appreciate reliability and stability in place of eye candy and complexity.
It can be done, and has already in a few companies. It will be happening in more companies. The process isn't trivial, but the results are worth the effort. The "Introduction of Unix to Microsoft Shops" is Part II of this article. It will outline the process and provide a starting point so you can begin the process of getting off the Microsoft Train.