At the Desktop: Whistle While You Work
At longlast, the end of the Windows 9x-dominated desktop computing era is coming intoview. Not soon enough for the millions of users that suffered through desktop instability,frequent crashes, corrupt files, and mismatched Dynamic Link Libraries (DLLs).
With themid-September release of Windows ME, the much-recycled Windows 9x code base --including the underlying DOS components and the 16-bit subsystems -- is havingits last hurrah. Looking ahead, Microsoft is marching into a world where it hasa single foundation for all of its non-embedded operating systems, ranging fromclient systems to 32-way SMP-and-beyond servers. For Microsoft, this is thefulfillment of an objective that has dogged the company for the better part ofa decade.
Microsofthas talked for some time about a consumer product based on the Windows 2000code base, but that is not meant to be. We have to look beyond Windows 2000 forthe answer.
In fact,Windows ME is nothing more than a stop-gap measure added into product plansabout 18 months ago to bridge the delay between Windows 98 and the availabilityof this next-generation client operating system.
It is clearthat the replacement for Windows 9x technology, particularly for consumer uses,will be derived not from the Windows technology currently in development underthe code name Whistler. When it becomes a set of shipping products,Whistler-based operating systems will exist for consumer PCs, business PCs, andin three versions for servers. Further complicating this roll-out will be theexpected availability of 64-bit versions of the business PC operating systemand the server products.
Microsofthas not yet released a public beta for the new client product, but it willlikely offer a user experience that borrows from both the Windows 9x/ME familyand Windows 2000 Professional. It is likely that some of thestability-enhancing features that Windows 2000 Professional offers, such as thesystem file protection and insistence that applications install DLLs and otherfiles in appropriate locations, will be vigorously enforced.
But it isalso possible that consumer products built on Whistler technology might nothave the ability to fully participate in a Windows 2000 Active Directoryenvironment, as can a system running Windows 2000 Professional. Anotherpotential problem is legacy applications. Microsoft is already encouraging ISVsthat market Windows 9x applications to upgrade them to be compatible withinstall requirements and Windows 2000 APIs. Supporting applications that arenot upgraded or can’t be upgraded will remain a difficult challenge forMicrosoft. The harsh reality is that if the Whistler-based replacement forWindows 9x fails to support such applications, this could create a substantialimpediment for customers considering widespread client system upgrades. Inother words, Microsoft still won’t be able to say goodbye to Windows 9x.
It islikely that the introduction of a Whistler-based consumer product in parallelwith a replacement for Windows 2000 Professional will result in a truedivergence between consumer and business uses of Windows client operatingenvironments. Finally, there will be an operating system that is positioned foruse by business users, and a different operating system geared up forconsumers.
Unfortunately,this won’t stop business users from using the consumer-grade Whistler productin business environments. Microsoft has two radically different pricing modelsfor its client products. It is highly unlikely that Microsoft will drop thepricing of Windows 2000 Professional -- and follow-on products -- so they’recompetitive with the Windows 9x pricing model. Likewise, consumers won’t be tooreceptive to paying significantly more for a Whistler-based consumer operatingsystem than they are accustomed to with Windows 9x.
Even withthat long-term divergence a given and the limited life expectancies of desktopPCs, change never occurs instantly. Windows 95, a 6-year-old product that hasbeen replaced by three successive product offerings, continues to sell incalendar year 2000. It is equally likely that in 2005, Windows ME will continueto sell in small quantities. At the same time, the installed base ofdiscontinued client products tends to shrink by 20 percent to 40 percent peryear, leading to a minimum five-year wind-down for products that are completelyunavailable. --Al Gillen is researchmanager for system software at IDC (www.idc.com) and former editor in chief ofENT. Contact him at agillen@idc.com.