Patching is too important to be neglected
Patching is too important to be neglected

After the better part of ten years, Microsoft has at last announced that Windows 98 is no longer supported. As  the company's least secure operating system, you might think that most people would be grateful to see the back of it.

But with an estimated 70 million users still running it, the software's demise has been met with howls of despair. Even some security commentators have complained, suggesting that Microsoft is acting irresponsibly and letting down its users.

Personally, I'll be glad to see the back of it. Windows 98 is little more than a cheap facelift stuck on top of DOS, and its “security” features are laughable. Worrying about a lack of security patches for a product that doesn't even have basic memory protection or file access control is somewhat academic.

Of course, as usual, Microsoft gets the flak, while its competitors escape criticism. Windows 98 was released at about the same time as RedHat Linux version 5. Whereas Microsoft had been releasing patches for 98 until recently, RedHat doesn't even list version 5 in its legacy support section. I've picked RedHat as an example as it's my preferred Linux distribution, but other Linux vendors don't do any better. There haven't been any patches for Apple's older systems for years, either.

You could, in theory, apply the patches to the source code of your Linux distribution directly and recompile, but that's hardly a sensible comparison for the user base of Windows 98.

Outside of the software business, it's often assumed that the cost of supporting legacy products through their lifecycle is not a major problem. In reality, this is the most expensive part of the process. Checking that your latest feature doesn't break something written ten years ago is hardly a trivial exercise.

I've even heard people complaining that their customers “have to” run their accounts package under DOS and would have to cease business without it. This is a classic false dichotomy. In reality, there are alternatives; but rather than plan for the future, many businesses simply ignore the upgrade issue and decide to run their software into the ground. Strangely, they seldom apply the same logic to their company cars, mobile phones or other business resources.

Rather than moan about Microsoft's perfectly reasonable decision to drop support after eight years, time would be better spent educating users in forward planning. The software giant publishes a support calendar well in advance; if people choose to ignore it, they shouldn't complain when Microsoft sticks to it.