Redmond Review

Life After 7: Windows New

What can Microsoft do to preserve the relevance of Windows -- and the crucial revenue stream that derives from it?

• By Andrew J. Brust
• 02/01/2010

What can Microsoft do to preserve the relevance of Windows -- and the crucial revenue stream that derives from it? As client software moves to the Web and server software moves to the cloud, what's left for a client OS that's designed to run local applications that talk to on-premises servers?

To find the inspiration for what ought to be in Windows 8, 9 and 10, I've been looking very hard at the virtual applications feature found in Windows 7 and its version of Windows Virtual PC (WVPC). Here's what I've found.

Virtual applications execute on WVPC guest instances that are running Windows XP, Windows Vista or Windows 7. Virtual applications actually merge into the host's desktop session, including the taskbar and system tray. They can be run directly from the host's Start menu and can access files on the host's physical disk drives.

The virtual apps feature allows for backward compatibility for software that will only run on an older OS, like XP. But the real impact of the feature is that it makes available an application without requiring it to be locally installed. This is great for beta versions of new software, as you can run them on your "real" PC and avoid the residue that often remains after an uninstall. It also means virtual applications run in a segregated environment, causing minimal impact on the host or on other guests -- available memory permitting. And because virtual apps are made possible by a combination of WVPC and Remote Desktop, the guest image could, one day, be running on another machine.

Virtual Is the New Physical
I would love to run every app as a virtual app. But to do so, I'd need to run multiple virtual machine (VM) images simultaneously, which would use a ton of memory; inhibit intercommunication between apps, including even clipboard sharing; mess up extended graphics modes; and likely cause a host of other issues. Plus, every time I wanted to add a new app, I'd need to build a new image of Windows to install it into. For each image, I'd need to license it, install security patches within it, run anti-malware on it and so on.

But what if I had an operating system built specifically to handle all this? It could, with hardware hypervisor assistance, provision new image instances on the fly; allow them to share memory; be patched and protected in unison; facilitate drive and device sharing between them; manage a global clipboard; and eliminate graphic display shortcomings. And the notion of remote instances, including cloud-based ones, could be a core scenario. In this OS, the accommodation of guest instances wouldn't be a special a trick; it would be the normal way of running an application. The guest instance would be analogous, in many ways, to what a window is today. With that in mind, this special OS could, perhaps, be the future version of Windows. Let's call it Windows New.

Windows New would look a bit like Citrix. But because Citrix's environment today requires a special server -- and its client sits on top of Windows itself -- Windows New would be more integrated, stable, simple and economical. Windows New would also look a bit like the Google Chrome OS. After all, Chrome is a slimmed-down, Web-based operating system that's built to launch applications rather than run them. But Chrome wants to relegate your application experience to the browser and HTML. Windows New, while accommodating browser applications, would instead let you stick with Windows applications, while greatly reducing their deployment and provisioning impact.

Bridging the Generation Gap
Windows New would leverage the Internet without imposing the Web or diminishing the client PC. It would best today's Windows, Android and Mac OSes, as well as tomorrow's Chrome OS. It would leverage the hypervisor, display, processing and memory-management power of today's PCs, rather than demoting them to dumb terminals.

Windows New would embrace virtualization and the cloud, while preserving the client application capabilities that users still want -- and that Microsoft and its OEM partners rely upon. Windows New would still accommodate AJAX and rich Internet application apps too, not to mention capabilities like Windows Media Player and Media Center. Windows New would be a modern, relevant, high-value product, not just another upgrade to an OS that was designed for personal computing in the 1980s. And it would be all this while still drawing on that very important heritage.

I'm convinced Windows needs to transform in a profound way, without abandoning its past and core competency. Windows New is one path to that destination. Would you buy it? Would you develop for it? Let me know what you think.