* An operating system upgrade can sometimes render equipment obsolete
Sometimes you just overlook the obvious. A couple of weeks ago, this newsletter responded to a reader’s query about the best way to go about installing Windows Server 2003. One of the questions asked was, “Should we install SP1 or wait till they fix the bugs it’s going to cause?” My reply was that the question was moot at this point, since SP1 has been out long enough – you should install it, and do so before adding any software to the system.
But in concentrating on the bug-fix aspect of the question, I overlooked a very important point. Quite a few of you picked up on that point and reminded me about it.
Sometimes, older applications and devices simply no longer work with the new, or newly patched, system.
I’m a firm advocate of running up new service packs in a lab environment for testing before introducing them to the production network. In this way, you discover any problems which might occur, such as applications (or devices) that no longer work. This is true not only of service packs, but also of upgrades, new operating systems, even changes to a single application – because you can never be sure of the impact of a change until you make it. And you want to inconvenience as few people as possible should a problem occur.
But not all of the problems are caused by bugs in the service pack, upgrade or other new software.
As I’ve said on more than one occasion, the top reason why people don’t upgrade old systems (e.g., your old NT 4 server) is that the applications and devices running on it are no longer supported with newer operating systems.
While it seems over the top to say that the new installation breaks an older application “by design,” that’s figuratively true. I discovered that for myself this past weekend when my wife tried to install a scanner she hadn’t used for a number of years. While the device (a true serial port connection) had worked well with Windows 95 and 98, trying to add it to her Windows XP desktop proved a very frustrating experience. There were no drivers for XP (or, for that matter, for Windows 2000) and the older ones no longer worked.
I don’t believe that Microsoft – which has always been known for including backwards compatibility with every operating system it ever developed – had deliberately left out any needed “hooks” in Windows XP. My theory relies more on a maturity lifecycle. What’s that mean? Come back next issue and I’ll elaborate.
