Last year, after Microsoft released Windows 98, Bill Gates said Win98 would be the last in the long line of Windows OSs. Many pundits solemnly pronounced the end of an era. The news that week must have been slow, because even I got a fair amount of air time commenting on the ostensible end of Windows. My call at the time—which you can read in En Garde, "The Last Windows? Hogwash!" October 1998—was that Windows would not die for a long time and that Gates would munch on his far-fetched phonemes before long. (Not that this episode highlights any great prophetic powers on my part. Microsoft had already signed Windows' death warrant in 1992 and 1995, only to later issue stays of execution in both cases.)
My prediction was confirmed in April at Microsoft's annual Windows Hardware Engineering Conference (WinHEC). This conference was an unusual place for news of a new OS, because WinHEC is where Microsoft usually tries to push the hardware industry toward new developments and standards. For example, the first big push for Universal Serial Bus (USB) took place at WinHEC a few years ago. Cool new hardware suffers from the chicken-and-egg syndrome—no one will buy the hardware until Windows supports it, but Windows won't likely support it until it's popular. Microsoft intends for WinHEC to jumpstart this process by offering a heavyweight perspective on what's new, what's hot, and what hardware future versions of Windows will support.
Although Microsoft exerts a noble effort at these conferences, no single computer hardware manufacturer is large enough to establish new hardware standards by fiat. This situation is a significant problem. To see why, consider transportable storage. From 1981 to 1987, IBM was important enough in the desktop market to create standards unilaterally. During that time, disk storage increased from 160KB to 1.44MB commensurate with hard disk capacity. However, as hard disk capacity continued to increase and IBM's dominance decreased, no medium for transportable storage took the place of 1.44MB disks. Zip drives are an excellent replacement candidate, but they haven't achieved the ubiquitous usage that disk drives attained—largely because no hardware vendor is influential enough to create a hardware standard.
Steve Ballmer took the opportunity at this year's WinHEC to announce that Microsoft will release a new edition of Win98 sometime in 2000. The new version will be the same type of 32/16-bit chimera as Windows for Workgroups (WFW) and Win9x, rather than an adaptation of the more rational NT kernel design.
Ballmer didn't reveal the new OS's name, although Microsoft is hoist with its own petard on that score because the company already allocated the Windows 2000 (Win2K) name to the latest vintage of NT. What, then, to do with this new Windows? First, no matter what Ballmer says, Windows Without A Name (WWAN) can't ship in 2000. Microsoft hasn't released any betas for the new OS and hasn't even released the second edition of Win98 at this writing. Perhaps Microsoft could ship WWAN in 2001, then an NT update as Windows 2002, another Win9x successor as Windows 2003, and so on. Consumers would learn that versions of Windows are like Star Trek movies. Odd-numbered episodes are eminently forgettable.
Although I never really believed that Wintendo was dead, I would be glad if it were true. I dual-boot to play Master of Orion II (I know, it also runs under NT, but very slowly), and I've been looking forward to the day that manufacturers build all new games on NT because no other choice exists. Of course, after Microsoft actually ships WWAN, the company will simultaneously announce that this OS will be the last 32/16-bit Windows. Get ready, you can be the first on your block to predict a recanting.