In-Depth

Primary Reasons to Upgrade to Windows Vista

Peter O'Kelly summarizes the daunting amount of Windows Vista information presented at PDC 2005 last week. Find out what''s changing between Longhorn and Vista, what aspects of Vista are being ported and more.

• By Peter O'Kelly
• 09/19/2005

The Windows client operating system now known as Windows Vista, previously code-named the Windows "Longhorn" client and first referenced more than five years ago at PDC 2000, is now close to its beta 2 release. For Microsoft product code name trivial pursuit enthusiasts, incidentally, the Windows client code names at PDC 2000 included "Whistler" for what would become Windows XP and "Blackcomb" for the post-XP release. Longhorn was introduced at PDC 2003, as a product family code name, and Blackcomb was implicitly either revised as a longer-term project or retired as a code name. The next major release of Windows Server still carries the Longhorn code name.

I'll summarize the conceptually daunting amount of Windows Vista information Microsoft presented at PDC 2005. I'll address three key themes in this article: changes en route from Longhorn client to Windows Vista, aspects of Vista that are slated to be ported to "down-level" releases of Windows (Windows XP and Windows Server 2003), and an overview of the reasons why individuals and organizations are likely to find Vista appealing even though many of its key features will also be available on earlier releases of Windows.

The many PDC 2005 keynotes, breakout sessions, and labs included a wide variety of Vista-related topics. A couple Vista details that were not addressed at PDC 2005, however, included pricing/licensing details and precise target schedules. Microsoft continues to follow its "We'll ship it when developers tell us it's ready" policy, but its overall target for Windows Vista is the second half of 2006. Given the stability and scope of the PDC 2005 Vista build, it's likely that won't mean late December 2006. Indeed, to align its timing with the 2006 back-to-school and holiday season, Microsoft may well be targeting a date closer to September than December, but we won't have a good sense of the likely timing until Vista beta 2 is available and analyzed.

From Longhorn Pillars to Vista Foundations
Before getting into Vista details, it's useful to review the key features originally planned for the Windows Longhorn client. Revisiting my July 2004 article, "Microsoft's Platform Strategies for 2006 and Beyond" (see Resources), Longhorn was intended to represent, from a developer's perspective, the full .NET-ification of the Windows platform. Microsoft explained at PDC 2003 that Longhorn would include four key areas. All the areas were to be presented to developers through WinFX, a new, unified, and radically simplified Windows API (based entirely on managed code). The four areas included:

• Presentation: Avalon was the code name for the Longhorn presentation subsystem, a complete upgrade for the traditional Windows presentation stack. While Microsoft explained that Windows Forms, Web Forms, and traditional browser-based applications would still run on Avalon, Microsoft believed HTML was peaking in terms of its functional evolution. Avalon was to be Microsoft's attempt to unify the best of Windows and Web Forms into a single architectural model.
• Data: WinFS was introduced as the code name for a new data management subsystem designed to complement the widely deployed Windows NT File System (NTFS) with richer, schematized, automatically synchronized, and agent-oriented data services.
• Communications: Indigo was the code name for an entirely new Microsoft implementation of the WS-* stack of Web services standards (WS-* is shorthand for the myriad standards initiatives designed to facilitate secure, reliable, and transacted Web services).
• Fundamentals: WinFX was (and still is) the name for the full Longhorn programming model, encapsulating lower-level (Win32) operating system services in an expanded .NET Framework, with fundamentals focused on security, manageability, and deployment.

Avalon, Indigo, and WinFS were introduced as the "pillars" of Windows Longhorn.

Fast-forwarding to PDC 2005, Microsoft's plans and naming conventions have changed quite a bit. Avalon is now the Windows Presentation Foundation (WPF) and Indigo is the Windows Communication Foundation (WCF). Both WPF and WCF have evolved considerably since PDC 2003, and the early versions made available to developers at PDC 2005 are considered fairly stable. WPF and WCF were originally expected to be exclusively available on the Windows Longhorn client, but Microsoft pragmatically concluded, after PDC 2003, that developers would benefit more if the technologies were also available on Windows XP (SP2) and Windows Server 2003, a topic we'll return to momentarily.

Whither WinFS?
The recent history of WinFS has been more problematic for Microsoft, and WinFS is no longer part of the Vista picture, at least not for Vista's initial release. The relational file system vision WinFS introduced is still under development, and a beta build was made available to developers at PDC 2005, but WinFS is no longer part of the initial Vista release picture.

Somewhat confusingly, many of the key features initially introduced as part of WinFS will still be part of the initial release of Vista, but not as facets of WinFS. Vista will include virtual folders (the ability to have a file, or other item type, appear in multiple logical folders, in contrast to the hierarchical model used in today's NTFS), for example, but as an extension of Vista's update to NTFS. There will also be a new Synchronization Manager in Windows Vista, but it's not part of the more ambitious synchronization services still planned for WinFS.

WinFS is expected to be included in an interim release of Vista at some point during (late) 2006 or 2007, and Microsoft plans to significantly expand WinFS with server-side WinFS capabilities in Windows Server Longhorn (during 2007). In the meantime, WinFS, unlike Avalon and Indigo, has not been promoted from a Longhorn pillar to a Windows Vista foundation. Microsoft has hinted that WinFS might ultimately be offered for "down-level" versions of Windows as well, but the virtual folder services included in Vista NTFS will not be "back-ported."

An Expanding Foundation
PDC 2005 included several surprises, along with updates covering Windows, Visual Studio, SQL Server, and other products. One of the biggest surprises was the introduction of a third Windows foundation, Windows Workflow Foundation (which Microsoft prefers to abbreviate as WinWF to avoid an unfortunate namespace collision with the wrestling-related WWF). WinWF, derived in part from the orchestration services pioneered in Microsoft BizTalk Server, is conceptually at a peer level with WPF and WCF (although it resides within the Communications part of the Vista WinFX diagram), and it will be included in the Vista version of the .NET Framework. In keeping with Microsoft's commitment to make the Windows Foundation services widely available to developers, it's reasonable to assume that WinWF will be available on Windows XP and Windows Server 2003, along with WPF and WCF.

Why Windows Vista is Still a Big Deal
Many people have asked if Vista is still a must-have Windows update, because WPF and WCF (and probably the new WinWF) will be available on earlier releases of Windows, and because WinFS is no longer part of the plan for the initial Vista release. The short answer is yes; the long answer will take many articles, over the next few months, to fully explore.

Perhaps somewhat counter-intuitively, the primary reasons Vista will be attractive to organizations include security enhancements, cost effectiveness, and operational efficiencies. Vista will include new data protection and application virtualization capabilities, for example, and will also be much simpler to manage because of significantly simplified image features (Windows Vista and Windows Server 2003 share a common code base and global binary images).

From an information worker (end-user) perspective, attractive Vista upgrade considerations include improved performance, a considerably streamlined user experience, and new productivity-impacting capabilities such as virtual folders and consistent, system-wide search services. Vista will also start (and resume from standby or hibernation) considerably faster than XP, a feature that will by itself warrant an upgrade for many mobile information workers.

The upgrade value proposition for developers is less clear at this point, because Microsoft has committed to release pivotal Vista foundation features on Windows XP, but it's important to note that the Windows XP versions WPF and WCF will not be functionally equivalent to their Vista counterparts. WPF on Vista will fully exploit Vista's (combined hardware- and software-based) graphics hardware optimization capabilities, for example. Vista's WCF, as another example, will be able to fully enlist the new transaction management features in Vista, simplifying the development of applications that require secure, transacted, and reliable Web services.

Windows is likely to become something of a successful chameleon. Many organizations will find Vista attractive primarily for security and operational advances, while information workers and individual consumers are likely to be swayed by Vista's more powerful and effective user experience. Many developers will consider Vista, especially when eventually used in conjunction with Windows Server Longhorn, to be a new distributed computing platform with an advanced and consistent, end-to-end (and deeply standards-based) programming model. That's going to make Vista marketing a rather tricky challenge for Microsoft, but ultimately there should be no doubt that Windows Vista is poised to rapidly become a major milestone in Windows product family history, certainly the most significant release since the introduction of Windows 95 a decade ago.