Why We Don't Upgrade to Win2K

Get It Try It

16 February 2000

Let's face it: Microsoft's products can be very cool. But sometimes it's best to wait a bit. Here's why we at Radsoft are in doubt.

Who else is upgrading?

An important question. For now, business is hesitant. Maybe business has always been hesitant, but this time around things have happened which we don't remember ever happening before. For example, Microsoft once had a 'Rapid Deployment Program' for Windows 2000, yet the more its members saw of the Windows 2000 betas, the more they dropped out of the program, and in the end the program just fell apart. Business is trying to tell us something here.

And if business has not decided that Windows 2000 is a fait accompli, then it's not going to be one for consultants either. On the contrary, more and more businesses are considering Linux as their next upgrade, because it's cheaper and would appear more stable, and because the need for qualified product support can be largely met anyway.

ROI

The ROI, or 'return on investment', of upgrading to Win2K was supposed to be one of the product's most attractive features. Windows 2000 aspires to what is known as 'zero maintenance'. But there's another factor here: the investment needed. Windows 2000 requires a complete machine park upgrade. Rightly or wrongly, Intel spent US$50 million upgrading their machines so they could comfortably run Windows 2000.

You're going to hear a lot of 'hype' about Windows 2000, and how fast it is, and some of this hype comes from unwitting consumers. One of the best we've ever heard, and this is absolutely true, was the following:

'Win2K is every bit as fast as NT - you just need more RAM and a faster CPU.'

That is an actual quote, and it reminds us a bit of what we're up against. Most of us are just waiting for an excuse to 'take the plunge', because it's fun. But in this case our take is 'you'll be sorry'. Microsoft upgrades and product installs are often woven with grief and mishaps; this one seems to have hit consumers harder than anything before.

Where's the beef?

We keep getting back to this bit over and over again: what does Win2K have that we don't already have? Answer: nothing. There are special interest consumer groups who will see something in Win2K that they like but don't have today, such as advanced support for games drivers and the like. But for ordinary business and home computing, for just getting on the net and surfing around, there's nothing we don't have now.

Who's writing the code?

All evidence points to the fact that the father of Windows NT, David N. Cutler, has basically been out of the loop since 1996. That he's officially head of Microsoft's Win64 project certainly is no help to the Windows 2000 development team. And evidence also points to the fact that most of his initial core team have left the campus as well. Leaving us with the important question: who's writing the code? No offense meant, but we simply don't think the average Microsoft developer is on the same level as Dave and his former team. Windows NT was fundamentally VMS, and VMS was known and respected for being bullet-proof. But whatever has been done to Windows NT to make it Windows 2000, it hasn't been done by the same programmers who made it bullet-proof before.

Linux

A new operating system has 'happened' since last Microsoft released Windows NT, and it runs on a fraction of the hardware needed to get Windows 2000 off the ground, and it's making inroads everywhere. And if the infamous Halloween documents are any indication, then Microsoft's own assessment of the relative merits of Windows 2000 and Linux are the latter's strongest selling point: Microsoft admits and laments in these internal documents that Windows NT/2000 can never approach the stability of Linux because their development method (with closed source) is all wrong and the Linux development method (with open source) is all right.

And the fact that so many major IT companies are not only endorsing Linux but actively working with it and conducting expensive research projects with it should tell us all something. Linux is most likely the wave of the future, and for geeks has been for some time now the only way to go.

Compatibility

We hear just too many stories: Windows 2000 doesn't work well with a lot of hardware 9x and NT support today. And if you don't check the hardware compatibility lists before you buy the product, you're asking for trouble.

If it works, don't fix it

Or in this context, 'if it works, don't upgrade'. NT has 58 million official users out there, and they are not going to disappear overnight. NT is a valid and much-used operating system, and there's no reason to abandon what works today for what might very well not work tomorrow.