Microsoft will deemphasize the Windows desktop in future releases of its operating system as usage of traditional Windows applications falls to just 10% of users' time by 2020, analysts said this week.
A quartet of Gartner analysts, led by Michael Silver, released a report Tuesday that spelled out the market research firm's prediction for what it described as a Microsoft-initiated "technology shift," the first in nearly two decades.
"Sometimes a new version of Windows represents more than a major upgrade," said Silver in the report Windows 8 Changes Windows as We Know It. "We call these 'technology shifts.' We don't see technology shifts too often."
In fact, for Gartner's money the only other example in Microsoft's 37-year history was the shift from DOS to Windows NT technology that kicked off in 1993 and wrapped up with the 2001 launch of Windows XP.
"Windows XP was the first Windows version based on NT technology to be targeted at businesses and consumers, and signaled the end of the DOS technology era," Gartner said.
Likewise, the emergence of WinRT -- the name for the new runtime and programming model linked to the Metro user interface (UI) and its touch-first apps -- is a technology shift, one away from Windows NT, the core of every edition prior to Windows 8, to WinRT and Metro.
That shift will probably run much like the one from DOS to Windows NT: A long process (eight years for the first shift to run its course) that for some time features both technologies (the DOS prompt remained key to Windows for years) and a gradual diminishing of the older technology's importance.
"Microsoft is not forcing anyone to eliminate Win32 applications or preventing developers from writing them, but Gartner believes that Win32 and the Windows Desktop will become less strategic over time," said the firm.
Most enterprises that adopt Windows 8 will continue to use Win32 applications and a traditional desktop browser through 2015, said Silver and his colleagues, but by 2020, users will spend less than 10% of their time on that platform, and thus on the desktop. By then, the browser and most applications, including what Gartner called "OS-neutral" apps, will run in Metro using the WinRT runtime.
That means Microsoft will likely spend little energy from this point forward improving the Windows desktop, the older, WinNT starting point of its technology shift.
"Gartner expects Microsoft to include the Windows Desktop in future releases, but improvements will be relatively minor," the research firm said, comparing future support for the desktop to the longtime inclusion of the DOS command line in Windows.
Microsoft's bet on WinRT and Metro is, of course, huge, but it was necessary.
"Gartner believes Microsoft needs to make such a change, an idea we first suggested in 2006," said Silver's report. "Our hypothesis [then] was that having to support legacy products was becoming a drag on Microsoft, slowing innovation, and that Windows had become too large for any systems designer to implement radical changes."
In 2008, Silver and Gartner analyst Neil McDonald, who also contributed to this week's report, called Microsoft's then-current position "untenable," and said Windows was "collapsing" under its own weight of legacy code and the need to support old software, both its own and that from third parties.
Microsoft, Gartner said then, had to address the problem by building a Windows system able to run on low-powered and low-priced hardware -- this was years before Apple introduced its iPad -- by crafting an operating system that was "thinner, smaller and more modular."
The result, Windows 8's and Windows RT's adoption of the WinRT runtime and the corresponding Metro UI and apps, proved Gartner prescient.
But although Microsoft had to take the plunge to avoid eventual irrelevance, it's fraught with pitfalls, some of which critics have already identified.
The attempt to shoehorn two operating systems into one package -- a knock almost everyone has leveled at Windows 8 -- is "jarring," said Gartner, which expects that Microsoft will polish the rough UI edges and deal with other issues relatively quickly, perhaps with a 2014 release.
Gartner called that future version "Windows v. next," but said it could be marketed as either Windows 9 or Windows 8 R2.
"[But] Microsoft's problem is that Windows 8 will need to be good enough when it ships, because Microsoft is already late and doesn't have a few more years," Gartner said.
While Gartner seemed to think that Microsoft's ability to influence developers and the sheer size of the Windows market -- Microsoft has itself touted the latter -- would eventually push its user base to WinRT, another analyst wasn't as sure that the company would pull it off.
"Part of the problem is that to target such a wide spectrum of processor architectures and devices, with different fundamentals, such as screen sizes and resolution, there just have to be some trade-offs," said Michael Cherry of Directions on Microsoft in an email. "[And writing Metro-style applications] is not as easy as Microsoft would like you to believe. To a large extent, [WinRT's and Metro's advantages] are all theoretical -- it looks good on paper -- but it will take time to determine if it can really be done."
Best case, said Cherry, would be that WinRT follows the course of .Net framework, a technology and development model that debuted in 2002.
"When it first came out there was a lot of confusion as to what [.Net] really was. This appears to be true of WinRT and Metro," said Cherry. "Then as people started to understand .Net, applications that began to use it appeared. .Net evolved. Better applications appeared. WinRT will likely follow a similar trajectory."
Gartner believes the first two years of WinRT will be crucial.
"Microsoft will fail if it cannot convince developers to adopt WinRT for future development quickly, especially for consumer applications, and stem the tide of OS-neutral application development," the research company argued.
But enterprises shouldn't overlook the long game Microsoft is playing: It's going to take a decade or longer for businesses and other organizations to complete the transition to WinRT.
"If successful, [WinRT] will represent the next 20 to 30 years of Windows," Gartner said.
"The question appears to be: Will enterprises adopt WinRT and switch to writing Metro-style applications?" said Cherry. "Short answer? It's way too soon to tell."
Gregg Keizer covers Microsoft, security issues, Apple, Web browsers and general technology breaking news for Computerworld. Follow Gregg on Twitter at @gkeizer, on Google+ or subscribe to Gregg's RSS feed. His email address is firstname.lastname@example.org.
Read more about windows in Computerworld's Windows Topic Center.