In the last two issues of Windows & .NET Magazine UPDATE, I discussed the fates of users still implementing Windows NT Server 4.0 and Windows 98 and what they can expect going forward as Microsoft inevitably retires these two systems. To most tech-savvy individuals, the idea of using an OS that's 6 to 8 years old might seem antiquated. However, the reality of corporate spending during an economic downturn, buying cycles, and the necessity of keeping legacy applications working means that many enterprises and even small and midsized businesses still use these and other, older non-Microsoft systems today and likely will for some time to come. This week, I wrap up this topic, present some feedback from readers, and discuss the relevant parts of the most recent 2003 Windows & .NET Magazine Reader Survey. Both the reader feedback and survey results speak volumes about the pains of moving technology forward in the real world.

Like many of you, I'm a technology enthusiast, and I have to temper my desire for new technology for technology's sake with the real-world needs of Return on Investment (ROI) and necessity. With the exception of the short-lived dot-com boom era, there was never a point in modern corporate history when companies implemented technology simply because it was available. Instead, forward-thinking IT decision makers, administrators, and other technology-driven users are regularly forced to sell technologies they want based on sometimes vague cost justifiers. Business- and government-oriented technology purchases (and, generally speaking, educational institution technology purchases) have historically held to a long usage cycle, one that's much longer than Microsoft's OS release cycle. The result is a mismatch of product versions in companies around the world, each with different capabilities, vulnerabilities, and other concerns. This management nightmare gets even worse when you combine it with the inevitable deployment of numerous non-Microsoft technologies, all of which are expected to work together.

As noted earlier, Microsoft's response to this situation has changed dramatically over the years from the "Windows everywhere" theme of the mid-1990s to a more realistic theme of interoperability today. But don't give Microsoft too much credit for bowing to market realities: The company is still far more interested in providing migration technologies than interoperability technologies and does little to support older products, beyond suggesting that newer products might be more secure and thus less expensive to operate. However, for most of the people using suddenly orphaned software, upgrading isn't an option.

Two weeks ago, I noted that the actual cost of software--that is, the price one pays to license a particular software product--is typically one of the least expensive aspects of a system's overall lifetime cost. But several readers, faced with Microsoft's new licensing schemes, are discovering that this is often no longer true. "I'm in the middle of making a technology plan for a \[non-profit organization\] and I'm finding that this is increasingly not the case," one reader noted. My suspicion is that we've reached an unprecedented point in the PC price model: Whereas typical PCs cost $1500 or more just a few years ago, today you can purchase full-featured PCs for less than $500. Thus Microsoft's OS (and, potentially, Microsoft Office) licensing fees represent a much greater percentage of the upfront costs of a PC than they did before. I suspect that's one of the many reasons people are starting to evaluate open-source desktop solutions such as Linux and OpenOffice.org in increasing numbers.

On a related note, a few readers opined that, by extending the support life cycle of Win98, Microsoft was giving Linux a chance to catch up. Systems that are now running Win98 will be unable to support Windows XP or Longhorn in a year or two, but these Pentium II- and Pentium III-era boxes with small amounts of RAM might make good candidates for desktop Linux installations, further stretching out the cost of the hardware. And let's not forget the application software these legacy desktops are running. Most Win98 desktops are probably running Office 2000 or Office 97, or other, older software packages that are either unsupported now or will be soon. I still believe that Linux and OpenOffice.org training costs and a loss of expected features will be problematic, but with another year or two to improve, who can say? Suddenly migrating a wide range of users to new hardware and software systems can be both expensive and painful.

With respect to security, Microsoft correctly notes that only its most recent systems are the most secure, and that aging, nonsupported systems will quickly become liabilities. In my mind, the security concern renders the "it still works" argument for legacy systems somewhat moot, although I should note that I did receive a startling number of email messages that supported the "it still works" argument. "Any migration is painful," one reader correctly noted, especially when your system is still working just fine. And for businesses still working in pre-Active Directory (AD) environments, AD is still daunting.

Let's look at some numbers from the September 2003 Windows & .NET Magazine Reader Survey, which is based on feedback from hundreds of responses. Eighty-five percent of respondents--largely IT management and staff--said that they'd already migrated to Windows 2000 Professional Edition on the desktop, compared to 46 percent for XP. On the server side, Win2K Server migrations are unchanged over 6 months at 78 percent, while Windows Server 2003 migrations are at 15 percent, up from 6 percent 6 months earlier.

Perhaps more relevant, 83 percent of respondents use Win2K as their server platform, compared to 7 percent for Windows 2003 and 10 percent for NT. But just 63 percent of respondents have rolled out AD, compared with 15 percent currently undergoing migration and 22 percent who have no plans to do so. I expected these figures to be skewed somewhat by the size of the companies that the respondents work at, but that wasn't the case: Workstation and server migration plans and time frames were similar regardless of company size.

I know from personal experience that a rift often exists between wanting to write about what's new and needing to support the technologies people actually use. Here in Windows & .NET Magazine UPDATE, I try to walk a line between presenting newer technologies you might not have heard about and tackling the concerns you actually face in your day-to-day jobs. As always, I'm interested in your feedback. The phenomenal responses I received from the previous two editorials were much appreciated, and I apologize for not having the time to respond to everyone personally. Thanks for reading.