By the end of 2004, beta testers and tech enthusiasts were running out of patience with Microsoft: The company hadn't shipped a public Longhorn pre-release build since the previous April and it didn't look like it was ready to do so anytime soon. And while Microsoft had pushed back its schedule, revealing that Longhorn wouldn't ship until 2006, and provided guidance on Longhorn features that would be back-ported to Windows XP and Windows Server 2003, questions remained. Fortunately, 2005 would see the beginning of the turnaround for Longhorn. It would get a final name, Windows Vista. And Microsoft would ship a Beta 1 release and then a rapid-fire series of interim builds that would finally give testers their first taste of the final Vista experience.

2005

In January, I published my fourth and final installment of the Road to Longhorn series, Road to Windows Longhorn 2005. These were an attempt to sift through the mess of information that was out there about Longhorn and provide a sane explanation of what was really going on. In this version, I was able to provide an extensive schedule for Longhorn, which was based on internal Microsoft documentation: At the time, Longhorn Beta 1 was expected in late June 2005, followed by a Beta 2 release in late 2005 or early 2006 and the RTM (release to manufacturing) in mid-2006. I also revealed that Microsoft would disclose information about this next generation operating system in stages. "Microsoft will utilize a disclosure approach it calls 'rolling thunder,' which will build up to a crescendo by the Longhorn launch," I wrote.

The Beta 1 release was then due June 30, 2005. At that time, Microsoft would disclose the "essence of Longhorn" and target businesses and developers. The company would woo consumers with Beta 2, my sources told me.

My "Road" article also walked through the Longhorn product editions as Microsoft then envisioned them and stepped through a variety of Longhorn features, including Aero Glass (now called Windows Aero) and Aero Express (now called Aero Basic), Least Privileged User Account (LUA, now called User Account Protection, or UAP), and a domain-like networking feature called Castles that was eventually dropped from the product. (Jim Allchin said this about Castles at the time: "The idea is to replicate identities across the home machines (if the user wants this to happen). Think about it as a domain in the home. For the scenario across the Internet we are trying to work out a solution so that it is easy to share and publish photos (or other data) with users across the Internet. Our goal is to do this so that standard Windows security IDs can be used.")

Internally, Microsoft was still wrestling to get Longhorn back on schedule. On January 13, I received internal documentation that specified the following schedule:

Windows Client Codename 'Longhorn' Beta 1 code complete: 3/16/2005
Windows Client Codename 'Longhorn' Beta 1 Internal: April 2005
Windows Client Codename 'Longhorn' Beta 2 (and product) code complete: 7/1/2005
Windows Client Codename 'Longhorn' Beta 2 Internal: Q3 CY2005
Windows Client Codename 'Longhorn' RC0 Internal: Q4 CY2005
Windows Client Codename 'Longhorn' RC1 Internal: March 2006
Windows Client Codename 'Longhorn' RTM Internal: May 2006

"Are we serious about May?" one internal email discussion reads. "Or should we plan for August?" The answer: "There has been a lot of talk about whether we are targeting May to hit August or if May is really 'drop dead' for the release. We are going to target the client for May 2006 -- period. This means that you and your teams should schedule for this date, and you are responsible with your teams for tracking your dependencies."

Even the milestones were up for debate. "Should we move to one beta instead of two?" another email reads. "There has been some discussion of whether we will have one beta or two," response came back. 'We are firm on two betas."

Microsoft was also wrestling with product differentiation. Because the company planned to ship so many different Vista product editions, some inside Microsoft were worried that customers would be confused. In a project status email, many decried the addition of Small Business (eventually dropped) and "Uber" (now called Vista Ultimate) editions. The problem, they said, was that Microsoft was dropping features and adding product editions at the same time: How could they possibly justify this? Meanwhile, others argued that certain features should only appear in x64 versions of the product. It was a mess.

In late January 2005, I received a bonanza of internal Microsoft documentation that lit up my site for at least the next two months. Included in this DVD's-worth of information was a wide range of never-before-seen information, including the first-ever shots of UAP, a UI widget locking feature that was later dropped in lieu of UAP-based icons, and a reliability UI that was also later dropped. I also received an early version of the new Longhorn flag logo.

In February 2005, news began seeping out that Microsoft was considering adding an animated desktop feature, codenamed Aurora, to Longhorn. Aurora would be available in a number of colors, and many later confused it with aborted work on a true vector-based Aero UI that never materialized. Aurora will appear via the Ultimate Extras service in Windows Vita Ultimate in late January 2007.

On February 15, 2005, Microsoft chairman Bill Gates announced that yet another Longhorn feature, Internet Explorer 7, would be back-ported to Windows XP and 2003. "What we've decided to do is a new version of Internet Explorer," he said. "We decided we're going to have the new capabilities even available before [Vista]. It will be another important advance." Just days later, I saw the first mock-up of the IE 7 user interface, which is shown here:

In April, Microsoft finally began talking up Longhorn again in the days leading up to the Windows Hardware Engineering Conference (WinHEC), where the company was preparing to ship its first external build of Longhorn in a year. "At WinHEC we'll give a build out of Longhorn, help developers through the transition of writing graphics drivers," Microsoft co-vice president Jim Allchin said. "You can call it a preview. It's not a beta. But it's dramatically different from the first preview. Nothing we have today has our new user interface on it. But we have some things to show you. There are a large number of people trying to get a jump using new technology already; we've been giving them that. After PDC 2005, we'll have a beta and we'll decide the [shipping] date. We're still on track for shipping by holiday 2006, so we'll be done before then. "

Mr. Allchin said that Microsoft's goals for Longhorn could be encapsulated by five key points:

1. It just works
2. Safe and secure
3. Easy to deploy and manage
4. Client experiences?at work, at home, and on the go
5. OS platform for the next ten years

For the first time, Allchin was showing off Aero on a real running PC as well. It was an early, unfinished version, but many of the animations and effects we now take for granted in Windows Vista were present. Sadly, they would be unavailable in the next build we'd get, which arrived during a tumultuous WinHEC event in late April.

Ah, WinHEC 2005. While I can't easily point to an absolute low point in my relationship with Microsoft, this was one of the contenders. We arrived at the show flush with excitement. Allchin had previewed a working version of the Aero UI and we were ready to get our dirty little mitts on our first Longhorn build in over a year. Microsoft, too, was celebrating an interesting milestone.

"2005 is the 20th anniversary of Windows," Microsoft Lead Product Manager Greg Sullivan told me at the time. "And all the pieces are coming into place for transition to 64-bit computing: the availability of mainstream 64-bit processors and, now, 64-bit versions of Windows." Microsoft announced the immediate availability of the x64 versions of Windows 2003 and XP at the show, not coincidentally, and used WinHEC to promote "the third decade of Windows."

Attendees at WinHEC received Longhorn build 5048, which Microsoft billed as a developer preview. I was disappointed in the new build, because it didn't seem to offer any big advances over build 4074, which we had received a year earlier. Indeed, it was a step back in many ways. "This build doesn't include the advanced Longhorn UI elements--called Aero Glass and Aero Express--although a subset of the Aero UI will debut in Beta 1 this summer," I wrote at the time. "Instead, build 5048 is aimed squarely at developers and visually resembles the WinHEC 2004 build we received a year ago."

Build 5048 featured a Search box in the Start menu because "that's where people start doing things in Windows." It supported virtual folders, though that functionality was scaled so far back over the next year that it eventually became a footnote, not a major new feature as previously planned. Sidebar was gone, apparently for good. Later, however, a new Sidebar project was launched, but it was based around Web-like widgets, not notifications like the original version.

Looking over the build 5048 screenshots now, it's easy to understand why we were so disappointed. It's ugly, drab, and gray, and utterly lacking in nice new features. But the real drama at the time was that Microsoft actually tried to get me and others to remove our Longhorn build 5048 screenshots from the Web, citing a clause in the EULA that prevented publication of these shots. Finally, I capitulated, but was later told I could put them back up. The general silliness of this maneuver, combined with the sad state of Longhorn at the time, put me in a foul mood.

My review of build 5048 reflects this. "It's been a long year," I wrote, describing the reset that Microsoft underwent in the previous year. "Longhorn build 5048 ... does not reflect some of the advancements Microsoft has recently made. Furthermore, it actually represents a usability back-step from last year's build 4074. That's because some features, like the Sidebar and the new system-wide Contacts utility, are missing in action in 5048. There are reasons for these omissions. None of them are particularly good."

"I have to be honest here. After a year without a single new Longhorn build and very little concrete information about what was going on with the project, I had high expectations for build 5048. And a pre-WinHEC briefing with the software giant did nothing to assuage those hopes. Plus, I've seen advanced Longhorn UI work and I knew how cool this thing was going to be. Build 5048 communicates none of that. And that's a shame, because Microsoft had a chance to ramp up the momentum of a product that, quite frankly, could use a little momentum."

"Anyway, Longhorn build 5048 is pretty boring. That it's boring by design doesn't make me much happier. I do know that the company will add back major new functionality in time for Beta 1. But sitting here in early May 2005, surveying the state of Longhorn, it's not pretty. Longhorn build 5048 is a disappointment."

(Microsoft later apologized for the debacle of WinHEC 2005, admitting that it should have done a better job of setting expectations, especially given the year off. And it admitted, too, that it didn't handle the screenshot stuff properly. By the time the show was half over, I was left with a weird queasy feeling. Longhorn was a "train wreck," I decided. I wasn't sure if they'd ever get it right.)

At WinHEC, Microsoft promised that Longhorn would "run fine" on a 1 GHz computer with 256 MB of RAM. Cue that laugh track. The company said it was then building 32-bit and 64-bit versions of Longhorn every day. "The two versions are co-developed and will be launched simultaneously," I wrote. "There will only be one version: During install, you can choose between the 32-bit and 64-bit version if you have a 64-bit PC."

Gates and others from Microsoft demonstrated what I called a "visually unimpressive" version of Longhorn at the show. As I wrote at the time, "Various Microsoft representatives told me that the Aero Glass UI demonstrated during the keynote wasn't the final user experience and that later beta releases would look dramatically nicer. Still, the Aero Glass interface that Gates demonstrated supported OS X-like transparencies, translucencies, animations, and other effects. Gates also showed off application scaling, which will help legacy applications display correctly on the high-dpi displays of the future, and a media-oriented sample application that takes advantage of Longhorn's display features."

I do give Microsoft credit for a rare burst of competitive zeal. Recall that Microsoft had announced its intentions to deeply integrate instant search into Longhorn back in October 2003 at the PDC show. Recall too that its competitors--Google, Apple, and others--then suddenly ramped up their own desktop instant search products. And since Microsoft was taking the leisurely road with Longhorn, these products had since shipped, or in the case of Apple, would soon ship as Spotlight in Mac OS X.

"Search is great," Sullivan told me at the time, referring to Apple's efforts. "We have desktop search with MSN, and we'll have it in Longhorn. But our contention is, if you're searching, you've lost something. We are building an automatically organized system where you don't lose it in the first place. The system is smart enough to understand the data itself and how different types of data relate to each other. What we're doing is much more impressive than the 'Hail Mary' pass of search, which often returns lots of irrelevant results. Don't get me wrong: Search is important. But it's only part of the story. The system we are delivering won't force you to search for your data."

More disappointments from the show: The Longhorn delivery timeline was adjusted yet again: Beta 1 would ship that summer, followed by a second developer preview in September at PDC 2005. Beta 2 would ship after that, Microsoft said, at which time Longhorn builds will be offered to the public. Longhorn was then due in time for Holiday 2006, and not for mid-2006 as previously planned. Obviously, the company later missed this schedule as well.

Also, while this wasn't discussed publicly at the show, I discovered during WinHEC that Office 12 was being decoupled from Longhorn: All Longhorn-related features were being stripped from Office 12 and the suite would ship before Longhorn. However, Microsoft was looking into shipping a Longhorn compatibility pack for Office 12 that would add back the missing features at a later date.

One side-note from WinHEC: On April 26, 2005, I (now infamously) wrote, "This one's bizarre, but we heard at lunch today that Apple is unhappy with the PowerPC production at IBM and will be switching to Intel-compatible chips this very year. Yeah, seriously." While I was roundly criticized for this by Mac fan sites at the time, I was eventually proven correct and credited with first breaking this story. Today, over a year later, all of Apple's Macs are Intel-based.

With WinHEC over, a curious silence descended over the Longhorn project. It's painful now to remember how badly Microsoft botched this show: This should have been a wonderful coming out party, where Longhorn was resurrected from the ashes of previous mistakes. But instead, Longhorn suddenly seemed to be dogged by indecision. This is a state that would exist for the remainder of the project, though a rapid-fire series of release milestones later in 2005 would help things dramatically.

In the meantime, Microsoft announced in June that Longhorn would support Raw image files. This would later be proven to be a bit of chicanery, however: Windows Vista does not ship with any form of Raw image support at all. Instead, Microsoft has made it easier for camera developers to ship Vista-compatible Raw filters and compatibility.

Microsoft also announced that it would support RSS in Longhorn via a new RSS subscription feature in IE 7 and RSS platform and data store features. These features would all later be ported to XP as well, marking yet another instance of a unique Vista feature not being all that unique after all.

In June, Microsoft also released the first beta of Monad, its .NET-based command line and scripting environment. Monad was originally envisioned as a feature of Longhorn, but, yup, you guessed it, was made available to XP and 2003 users and eventually stripped out of Longhorn all together.

In late June, I discovered that Windows user experience guru Hillel Cooperman was now involved in a skunkworks project at Microsoft codenamed Project M. This small group, which reported directly to Chris Jones, the Microsoft VP in charge of the Windows Shell, was working on mysterious Longhorn shell improvements that would appear in Beta 2 or beyond. To my knowledge, this work never materialized, and Project M morphed into Project Max, a bizarre application for sharing photos.

Meanwhile, Longhorn Beta 1, I was told, would be "by far the best OS Beta 1 release" Microsoft had ever shipped. Given my disappointment with build 5048, that was great news. But as with the Project M stuff, this information would prove to be less insightful than I had hoped.

In early July, Microsoft sent out the first Longhorn beta invites to testers. "YOU ARE INVITED TO JOIN IN THE PRE-RELEASE TESTING OF WINDOWS," the email thundered to excited recipients. The following Q &A was provided:

What is Windows Code-Name "Longhorn"?

The next version of Windows, Code-Name "Longhorn," promises to be the most secure and intuitive Windows release to date. It delivers on the promise of allowing people to use their computers more effectively and confidently to achieve their goals and pursue their passions. It offers new tools to help protect the integrity of your system and your information, easier ways to find, visualize and organize your information, and provides better integration across applications, devices and systems.

Longhorn will provide advancements in the following key areas:

A strong focus on the fundamentals of the operating system, including advancements in reliability, performance, deployment, and ease of use.

Major improvements to help PC users to work smarter and provide exciting new experiences for home users.

The next-generation developer platform to make it easier for developers to create breakthrough applications.

In early July, Longhorn build 5023 also leaked online. This build featured a number of features we've now become very used to in recent Vista builds, including the new file copy animation, the categorized Control Panel view, a near-final IE 7 UI, the new Start Menu layout, and more. There were still weirdisms that would later be dropped: Microsoft at the time was trying to figure out a way to colorize the Details pane at the bottom of each Explorer windows, for example, where specific colors would be applied to windows based on their contents. They eventually dropped this idea, however, and today in Windows Vista, the Details pane is always light blue.

While on vacation in Stowe, Vermont with my family that month, I got a frantic phone call from a friend at Waggener Edstrom (which handles much of Microsoft's PR) about a Longhorn-related briefing. I told them I was on vacation, but he told me I'd want to take this call. So cell phone pressed to head, Dell laptop on lap, I sat out on the balcony of the hotel, overlooking my kids swimming in the pool, surrounded by the forests of Vermont, and was told by Greg Sullivan that Longhorn ... would be named Windows Vista.

I laughed. You're kidding, right? You interrupted my vacation for this?

Anyway, Sullivan explained the key Vista marketing terms: Connected, Clear, and Confident. He noted that Microsoft had considered other names, like Windows Seven, Windows 7.0, and Windows 7, but thought Vista communicated the company's vision of the possibilities of this next Windows version nicely. "We live in a world of more information, more ways to communicate, and more things to do," Sullivan told me. "You want the PC to adapt to you and help you cut through the clutter to focus on what?s important to you. That's what Windows Vista is all about: bringing clarity to your world, so you can focus on what matters to you."

Microsoft group vice president Jim Allchin expressed his feelings for the name as well. "I love this name," He said. "'Vista' creates the right imagery for the new product capabilities and inspires the imagination with all the possibilities of what can be done with Windows ? making people?s passions come alive." I figured I'd get used to it eventually. (And sure enough, like Pentium before it, Vista is now just part of the language.)

Windows Vista (well, Longhorn) Beta 1, or build 5098, shipped later that month. It featured an install routine that was very similar to what appears now in the final version of Vista. There was a neat Network Sharing feature (later dropped of course), Aero Glass all around (and curiously nearly identical to the final Aero version, despite promises that improvements were coming), new notification pop-ups for power management and volume control that are surprisingly similar to the final versions, a new Start Menu with in-place folder expansion (at first reviled, but now accepted), and so on. Beta 1 features some nifty shell features that would later be dropped as WinFS fell by the wayside, like Lists.

My Windows Vista Beta 1 review was less tepid than my reaction to build 5048. "Windows Vista Beta 1 is about what I expected to see in April, when Microsoft released build 5048 at WinHEC 2005," I wrote. "On that note, it's not a horrible disappointment like build 5048. However, because it lacks the end user niceties we'll see in the PDC 2005 build, in Beta 2, and in the final product, it's not something that will excite average users. Seeing how the virtual folders will sort out is somewhat interesting, and I'm eager to use this organizational system full time, as I'm anal retentive about creating specific document folder structures right now anyway. Beta 1 is all about possibility and promises, and that's OK. My only real disappointment is that it took so long to get to this point: I first saw many of these features almost two years ago and now I want more."

To be fair to Microsoft, Windows Vista Beta 1 was much further along than was Windows XP/Whistler Beta 1. The problem is that Whistler Beta 1 was a clean slate at the time: With Longhorn, we'd been getting builds for years by that point. Testers, reviewers, and analysts were looking for progress.

At the annual Microsoft Financial Analysts Meeting in late July 2005, Windows Vista was, as you might expect, the source of much discussion. "I think about Windows Vista as the beginning of the next generation of products for Microsoft," Microsoft CEO Steve Ballmer said. "The PC market, the server market?those are growth hardware markets. The units in marketplace will grow. We have an opportunity, and we see it when we come out with exciting innovations to stimulate even greater growth on the software side than we see on the hardware side." He also revealed that Microsoft would ship an "enterprise version" of Windows Vista. Curious, Vista didn't come up even once during the executive Q & A that day.

As July came to a close, I came into possession of some interesting Vista prototypes. These images portrayed how Microsoft felt the Vista UI would progress and they're reasonably close to the final version.

2005 was proving to be a watershed year for Windows Vista, but the best was yet to come. Next up: The 2005 PDC and a sudden and surprising slew of Vista interim builds.

Continue with Road to Gold: The Long Road to Windows Vista Part 5: August-December 2005...