Longhorn was the codename for the successor to Windows XP, which begun development in 2001, before the initial release of Windows XP. Originally, it was envisioned as a minor upgrade to Windows XP. It was planned to release in October of 2003.
At the same time, Microsoft was also internally working on Blackcomb, the next major version of Windows. However, as Longhorn began to grow in scope within Microsoft, it included many new features and enhancements originally meant for Blackcomb, such as a new management system called "WinFS", or Windows Future Storage, a relational file system which would allow users to easily manage, organize, and search for files and their other data in large folders.
Microsoft added features such as the Sidebar, the Windows Presentation Foundation (also known as Avalon), and Aero Wizards, most of which can actually be found in some early builds, such as 3683 and 3706. Many of these features were built on the new .NET framework at the time. From 2003 to 2004, Microsoft released many builds of Longhorn, and later builds were eventually considered slow and riddled with bugs. 4039, for example, is a build most known for its extreme bugginess.
With no real end goal for the project, .NET framework integration which caused instability and memory leaks, starting componentization much later than it should have, and many, many features that were added haphazardly on top of one another, one could argue that Longhorn as it was originally pictured was doomed to fail from the start; rapidly changing code, a plethora of merely work-in-progress ideas and concepts, a new graphics engine--all of this was poorly managed and handled. Slowly, but surely, over time, the situation grew worse. All of these new features: the Sidebar, WinFS, and others caused Longhorn to become bloated and beyond fixing. It seemed as though something needed to be done.
A screenshot of an early build of Longhorn (6.0.3683)
A screenshot of the new "Display Properties" Window (6.0.3683)
A picture of build 6.0.4081. This was the first build to begin the componentization period.
More than halfway during the development of Longhorn, the process of componentization had begun. This is exactly how what it sounds like; The operating system is stripped of all the extra components, and what remains is the system itself and the shell--the minimum required for it to function. Microsoft wanted to componentize Longhorn to make different versions of it. Let's think of a Home and Professional edition: A Home edition would have basic applications for the average end-user, while the Professional edition would contain extra tools, such as a Remote Desktop application. This design philosophy was carried over into the final version of Windows Vista, with many different versions targeted at different users, such as the Home Basic, Home Premium, Business and Ultimate versions, each with their own features enabled.
As previously stated, most of the componentized builds lack major features, however, 4081 and 4082, the first componentized builds, still retain Aero if you wish to enable it.
The main issue with componentization of Longhorn isn't the fact that it happened at all, it was instead the fact that it happened far later into development than it should have--componentization should have been one of the first steps to developing Longhorn, not as a last-ditch effort to try and fix what was unfixable at this point. Many builds from this period are nigh unusable, crash or hang almost constantly and are full of memory leaking issues and bugs.
By 2004, it was unanimously accepted among the teams tasked with completing the project that development of Longhorn was too far gone, with the decision at Microsoft to reset development completely, and worked on this time with a clear direction, scaling down the large scope of the project, starting fresh by building onto a newer, more stable codebase (Windows Server 2003 SP1) as well as removing features such as WinFS. The last build of Longhorn created under the original codebase, build 4093, was created merely hours before the infamous reset occurred and features a surplus of hidden system modules, new dialog interfaces, applications and components, and gives us the closest look at what Longhorn could have looked like had it been completed.
Let's fast-forward to 2005. Build 5048 was a pre-Beta 1 build, and the first build to show the progress Microsoft had made at reintegrating the features from the pre-reset period into the system. It was showcased to attendees of the Windows Hardware Engineering Conference 2005. Judgements of what was showcased at the event were described by Paul Thurrot as the "makings of a train wreck" "...sitting here in early May 2005, surveying the state of Longhorn, it's not pretty. Longhorn build 5048 is a disappointment." Build 5048 was lacking in features and seemed very similar to Windows XP at this point in development.
Features such as the Sidebar, and the fancy graphical enhancements such as Aero were eventually reimplemented and were included in the final product of Vista, but had features cut or did not have all the functionality as they did pre-reset. Namely, WinFS as it once was envisioned was nowhere to be seen in Vista, but search functionality via the Start menu and Search tool comparable to that of WinFS was included in Windows Vista.
Windows Vista was eventually released in 2006, 3 years after its slated release. And it was the most hated piece of software Microsoft had ever released up to that point. It's easy to see why--some felt lied to as a number of promised technologies and enhancements that would not have only made the usage of Windows easier, but development for applications through the integration of said promised technologies were plainly nowhere to be seen in the final release of Windows Vista. Windows Vista also required specifications much higher than that of its predecessors, especially if you wanted to enable to new Aero interface. On release it was infamous for its instabilities, lack of technical polish, crashes, and a sea of other problems. OEMs did not supply machines with Windows Vista the necessary hardware to have a smooth experience. And that's just scratching the surface of the problems that Microsoft had faced in the mid-2000's when Vista released.
4 new versions of Windows and almost two decades years later, what else is there to say about Windows Vista that hasn't already been said? In spite of it's rocky launch, Windows Vista did manage to eventually get its footing, especially with the release of Service Pack 2, but by that point, Windows 7 was on its way and the damage had been done. I think as a piece of technology, it is a prime example of what can go wrong when you have too much ambition for a single project. Despite that, I'm sure most do have an appreciation for it now for what it was trying to accomplish especially when comparing it to releases such as Windows 8. Microsoft has been, for over the last decade, been chasing industry trends which each new release of Windows as opposed to making Windows as a platform better to use and develop for. They chase trends in the hopes that they will strike gold in a certain computing market. However, I think with the rapidly advancing technology of the modern computing world, maybe it's time to take a look back at the plans slated for Longhorn and make something more efficient for everyone.
A picture taken of build 6.0.5048 taken during WinHEC 2005. An explorer window is open.