I'd like to offer some practical advice regarding controlling Linux on corporate networks (Linux: Renegade or ally?):
What organizations should be doing is finding out why users feel the need to add Linux. It's a good bet they don't set up the box just for kicks. Most likely IT isn't supporting a need, and the users are deciding that they'd be better off supporting themselves.
It's been my experience that when new systems start coming in the back door, it's usually because IT can't or won't do something that users want -- maybe they don't have the experience or resources, or they're wound up in a bureaucratic power trip.
If you don't have the resources to support Linux, then the right approach may be to formally let the Linux users support themselves, at least until you have the resources. It's likely that users installing systems on their own expect little or no support from IT, so they won't be surprised or offended if they don't get it. Unless what they're doing harms other IT services, this laissez-faire approach can allow significant experimentation without requiring IT resources. In certain roles, Linux has proven to be dirt cheap and terrifically reliable -- two things that are not always true of more traditional solutions.
Lastly, IT often views backdoor systems as a challenge to its authority. That is an easy trap to fall into, but it's the wrong way of looking at the situation. These systems are not intended by the users to be a challenge; they're intended to get something done that IT is not efficiently getting done. As such they represent an opportunity to fix problems in IT's services and organization.
Art Technology Group
"Linux: Renegade or ally?" is an interesting and rational story, and I can appreciate the point that sneaking things onto networks, no matter how good, presents serious problems to IT managers. I suggest the real problem for many organizations is either inflexibility or a lack of technical experience in IT management.
Let's examine the issue of breaking the rules a little more closely. It might be helpful to examine a more rigid structure than corporate management hierarchies -- the military. There, you find a concept of "showing initiative," in which an officer extends or reinterprets his orders, sometimes even violating them. Where successful, this behavior is highly encouraged; it is half of what makes a truly good officer, and what frees higher officers from routine decisions.
Of course, if you don't go by the book and fail, your career just got significantly shorter. This is a pretty good way to sort those who can exercise good judgment from those who can't, althoughh it is a little bit hard on the troops at times.
In the less rigid civilian sphere, breaking the rules again has its place in civil disobedience. Social rules are more frequently ignored, and the most successful such experiments often become mainstream, just as in the military. Rule-breaking is a little like mutation: You don't want a lot of it, but you can't evolve or adapt without it.
So when users are sneaking stuff onto IT networks, I have to think at least half the problem is an inflexible policy by IT management, designed to make the manager's job easier rather than the users'. And that, ultimately, is bad business.
If I found a Microsoft machine on my network, it would be removed, and the person who added it to the net would be fired. Microsoft machines are not compatible with our applications, and most of our software is poorly supported on Wintel machines. Therefore, we have no use for Microsoft renegades.