There is reason today for great celebration throughout computer land. Users, developers and hardware manufacturers should be popping corks and shouting from the rooftops. There should be fireworks and spontaneous parades. And yet, the digital landscape, from the desktop to the far corners of cyberspace, is strangely placid, as if nothing changed last week with the unveiling of Microsofts new operating system, Windows XP.
Thats because the best thing about XP, which is scheduled for release in the fourth quarter, is the death of a beast whose name may not be uttered — at least not in official Microsoft marketing. The lengthy presentation by Bill Gates and Jim Allchin, the companys group vice president for platforms, focused on XPs simplified interface, its expanded networking capability and its vastly improved multimedia capabilities. Yes, all that impressed me, and Im not that easily impressed.
But what truly amazed me was what both men chose not to say. The one word missing from both their spiels, like some cosmic hole in the fabric of reality, was "DOS." They failed to mention that Windows XP is a requiem for DOS. They failed to rejoice, "Dingdong, DOS is dead!"
Dont get me wrong. Im not suggesting Gates and Allchin are dishonest. No, Microsofts worst flaw is that it lacks a sense of humor. This is a company that so venerates its legacy that it cant laugh at its own foibles — and, thus, is unable to connect in a fundamentally honest way with its customers. As I watched the XP unveiling on the Web, my thoughts turned to the introduction of the Macintosh in 1984, and I wondered how Apple Computer Chief Executive Steve Jobs might have presented this groundbreaking product. My guess is that he would have wheeled out a coffin and held a wake for Dear Dead DOS on the stage to symbolize a complete break with the legacy.
The issue isnt showmanship. Its a public acknowledgment that the arcane command-line operating system that built Microsoft has been the cause of a decade of consumer frustration. Instead, Microsoft seems to be treating the demise of DOS like some embarrassing suicide in the family.
Yet, whats really important to consumers is not that XP has a lot more bells and whistles, but that it was built on the DOS-free Windows NT kernel. That leaves the distinct possibility that even when a program crashes, it wont take the whole operating system down. Microsoft is bent on selling features. A weary, frustrated world of consumers and small businesses — the market for XP — would rather hear about stability, reliability and security.
But in order to make clear why XP will be more stable, reliable and secure than Windows 95/98/ME, Microsoft would have to admit two things: First, that despite six years of marketing these products as truly integrated operating systems, they are actually little more than extraordinarily complex graphical shells that rely on — while desperately attempting to hide — a dated and fundamentally flawed platform. And that business users of the more sophisticated and expensive Windows NT/2000 have been living without DOS and its attendant vulnerabilities for five years, while users of consumer versions of the platform have had to put up with incessant crashes, lost data, costly downtime and far greater vulnerability to vandals and thieves.
Granted, there are valid reasons why Microsoft maintained this double standard. On the consumer end, DOS has long remained the bastion of game players — who are among the most passionate power users, early adopters of each new generation of processors, peripherals and software. And the dirty little secret of the business world is that there are still millions of DOS applications running every day, which has kept enormous pressure on Microsoft to maintain a platform that was of absolutely no use to the vast majority of its customers. So there sat DOS, encased in Windows like Vladimir Lenin in his glass box, and just about as useful.
Now DOS is being buried. Its time Microsoft raised a public toast to the demise of its pet beast and to the dawning of a new age in personal computing.