Programming Windows: Frankenstein (Premium)

On February 29, 2012, Steven Sinofsky stepped onto a small stage at the Hotel Miramar and introduced the Windows 8 Consumer Preview. This was a make-or-break moment for Sinofsky and his controversial new operating system. And despite growing internal misgivings about the direction Windows 8 was taking at Microsoft, he seemed to pull it off.

Well, mostly.

“Wow, we are really excited to be here, and you know I’m supposed to say super-excited to be here, since that’s our Microsoft thing,” he started off, reusing the joke he opened Build 2011 with. This time, it landed with silence from the confused audience. He then referenced getting a picture from the operations center that was gearing up for all the Windows 8 Consumer Preview downloads … and didn’t show it to the assembled members of the press. Yep, another awkward Sinofsky presentation was underway.

The narrative hadn’t changed. Sinofsky and his team had begun the Windows 8 project before shipping Windows 7, he noted, but that wasn’t at all unusual. He was clearly trying to reinforce the notion that Windows 8 wasn’t being rushed to market, that its many, many changes when compared to previous Windows versions had, in fact, been in the works for years and had thus passed some form of collective muster. But it’s important to note that Sinofsky had had no more time to build Windows 8 than he had for Windows 7, a minor upgrade. Or as Sinofsky described it, “the most successful release of Windows, the most successful OS of all time.”

For the umpteenth time, we were told that Windows 8 was “nothing less than a bold reimagination [sic] of Windows, from the chipset through the experience.” It was a “no-compromises” release that would combine “the best of the PC with the best of mobility.” This positioning was designed to undercut Apple, which had released the iPad in 2010, creating a third product category between the smartphone and the PC. Doing so made sense to Apple, which earned the vast majority of its revenues from selling hardware. But it was an anathema to Microsoft, which saw the iPad as yet another device that, if successful enough, would further deemphasize Windows and the PC in the minds of consumers.

“In our day-to-day lives with all of these devices, we face too many choices where we have to choose between things, this or that, we’re choosing between consumption or productivity,” he said. “We’re choosing between more battery life or more functionality. We’re choosing between form factors: a tablet or a laptop. Even something very simple, we have to choose: do you want a touch interface or a keyboard and mouse?”

The premise here was enticing. With Windows 8, users who stuck with the PC could do it all. They could enjoy iPad-like consumption activities—reading, browsing the web, reading email, enjoying movies, and so on—using the same PC they used for work. And that PC could be anything from a traditional form factor laptop to a modern new tablet PC, including some models that ran on ARM chipsets—just like iPad—and presumably delivered better battery life too. Choice was good.

That this system offered Microsoft an answer to the iPad was likely what put the radical Windows 8 design over the top for the software giant’s senior leadership team (SLT), which had been stung badly by Apple’s refutation of Bill Gates’ vision for the Tablet PC. If Windows 8 succeeded, after all, it would lift the Tablet PC with it and, maybe, put a ding in Apple’s unparalleled success.

The problem, of course, is that the reality didn’t meet the promise. Contorting Windows to accommodate a new touch-first, mobile-focused app environment created complexity and alienated the entire Windows user base, which reasonably expected a system that looked natural and worked well on the traditional PCs everyone was using. And contorting the PC to accommodate new form factors was at best a mixed bag, with pure tablet form factor PCs—running on ARM or x86—being too limiting for most. In both cases, Microsoft promised no compromises when all it was really offering were compromises.

That wasn’t clear in February 2012, however. And at the Consumer Preview event, Sinofsky and his chief lieutenants made an impassioned plea to the contrary.

To make such a radical change with Windows 8, Microsoft needed to address three core areas. It would evolve the core operating system with a new user experience, invent a new app ecosystem with an app store and some interesting information-sharing capabilities, and get buy-in from its key PC partners that would create new form factor PCs with tablet and touch capabilities.

Given Sinofsky’s methodical approach to, well, everything, it is perhaps not surprising that the event then covered each of these areas in turn, and in that order. But first, he had laid out some whoppers.

It started with the literally incredible claim that Microsoft had somehow made “over 100,000 code changes to Windows 8” since the Developer Preview in just four months. “For many of you, the Consumer Preview is like a whole new product,” he claimed, one that was now “complete,” enabling a “generational change” for Windows.

“The last time we made a generational change in Windows … was with Windows 95,” he said, drawing a comparison to the most important Windows version of all time. “Back then, we had to actually label the Start [button] with an arrow,” he added, mimicking the UI’s animation with his hands. “You know, ‘Click here to start.’ And that was because, at that time, computing was new to most of the customers of Windows. But today, interfaces, computing, touching devices, interacting with mice and keyboards, a huge number of distinct user models and user interfaces are an everyday experience. Just browsing the web. And so the work that we’ve done to make Windows 8 easy to use, we think is going to be a super fun experience and will come naturally to people.”

We should pause here momentarily to retroactively see this as the warning that it was.

In 1995, Microsoft had added a “Click here to Start” animation next to the Start button in Windows 95 so that users unfamiliar with its new interface would know, literally, where to Start. But Windows 8 would controversially go on to ship with several undiscoverable new user interfaces, none of which were present in other computing interfaces with which its users were familiar. Windows 8 would include no proactive help in any form, and Sinofsky would literally ignore the feedback and telemetry data that would prove during the testing of the Consumer Preview that users did, in fact, need this help. The result, as we’ll see was a disaster: key system features were hidden from users, and those that did stumble upon them by mistake wouldn’t know how to repeat the action that had triggered a hidden UI.

Perhaps fittingly, it is at this point that Sinofsky then brought out key lieutenant Julie Larson-Green to deliver yet another familiar Windows 8 user experience demo. It was she, after all, who led the development of the Windows 8 user experience, and it was her responsibility to not only OK something that worked, but to push back against Sinofsky when he refused to mar the simplicity of the UX with user help.

She did neither, and nothing substantial had changed since the Developer Preview: Microsoft had made only small refinements to the Windows 8 user experience, and Larson-Green’s delivery, as always, was lackluster. Sitting in a chair to demonstrate Windows 8 on a tablet—shades of Steve Jobs demonstrating the iPad, but without the presentation skills—she stuttered haltingly through a tour of Windows 8 that included the lock screen with its non-descriptive notification icons, the touch-friendly picture password feature, and the full-screen Start screen, with its bizarre panoramic UI that made it hard to find apps that you couldn’t see in the first screen full of information.

There was a lot of misinformation.

Windows 8 would not “adapt” to users as she claimed, at all. Instead, it would just dumbly add tile-based shortcuts for newly installed apps way at the end of that huge panoramic experience, all but hiding them from the user, who would then need to manually edit the Start screen by moving app tiles to the left of the Start screen or into other app groups. Microsoft also used this panoramic UI style in many of its own apps for Windows 8—and on the Xbox 360—and its landscape orientation was vexing for traditional PC users who were used to the more traditional portrait scrolling used by desktop apps.

Green also tried to position Windows 8’s weak multi-app capabilities as a continuation of the desktop multitasking capabilities of the past. “Windows has always been really good at letting you do more than one thing at a time,” she noted, without mentioning that Windows 8 would severely impede this capability by forcing Metro-style apps to run full-screen or, in those rare times when an app supported it, in a tiny docked area on one side of the screen. She was so proud of this incredibly limited feature that she demoed it repeatedly back-to-back. And she even made lemonade from its limitations by noting that it was nice she didn’t need to “manage” these app windows because they couldn’t overlap. Metro simply wasn’t sophisticated enough to allow it.

And Microsoft’s plan to support popular social networking services like Facebook in Windows 8 would quickly fail just as it would in Windows Phone: the firm built a People app that could “aggregate” content from multiple services and let users post from a central location. But Facebook and other services quickly soured on this functionality, preferring for their users to use native or web-based apps. People would go on to become a functional vestige, ignored by users, as would be a related Messaging app.

Confusingly, Windows 8 would ship with two versions of Internet Explorer, one for the desktop and one a barebones Metro-style version.

Even more confusingly, tablet users would need to learn a bewildering series of gestures that would require swiping in various ways on the screen or from one of the four edges of the display. And some were complicated.

For example, swiping from the left edge would switch to the previously-used app, but swiping from the left edge and then continuing the swipe downward would display the Switcher, a Metro-style way to switch between running apps. Swiping up from the bottom would display app commands. Swiping from the right would display Charms, a toolbar of sorts with system functions. And so on.

Next up, Antoine Leblond appeared to show off the new Windows Store and its mobile apps, which were now available to external testers for the first time. But he would do his demo on a traditional laptop, not a tablet, and show how Microsoft had adapted its Metro user experiences for the keyboard and a mouse or touchpad.

There were some nice touches for keyboard users. Microsoft had removed the Ctrl + Alt + Del requirement to get to the sign-in screen, for example. And you could now use a 4-digit PIN, instead of your password, to sign in to Windows, similar to a smartphone.

As for the Start screen, Microsoft allowed users to navigate around as expected, with the arrow keys or PgUp and PgDn, and you could use a mouse naturally, including its scroll wheel if present. Launching apps was likewise straightforward. Other navigation would require non-discoverable actions, and Microsoft had decided to leverage the four corners of the screen so that mousers could perform specific actions.

To get to Start, for example, you would mouse into the lower-left corner—where the Start button would have been had Microsoft not removed it—and click the Start screen thumbnail that appeared. (This worked as a toggle too; if you performed this action from Start, a thumbnail would appear for the most previously-used app.) The only thing that might have made this easier was a little animate arrow noting that you could “Click here to Start.”

From Start, you could access the semantic zoom view of Start by mousing into the lower-right corner. Otherwise, mousing into this corner, or the upper-right corner, would display Charms. To switch between two apps, you could move into the upper-left corner to display a thumbnail. Or you could access the Switcher interface by mousing into the upper-left corner and then mousing down the left side of the display.

Leblond also highlighted the few but meaningful updates that Microsoft had made to the desktop environment in Windows 8. Among these features was much-improved file copy functionality in Windows Explorer that supported multiple concurrent copies, the ability to pause a copy operation, and better resiliency if a laptop went to sleep in the middle of a copy operation.

But the big changes since the Developer Preview were all about the apps: while still behind the system itself, the apps in the Consumer Preview were both more voluminous than before and a bit more refined, though each still clearly communicated the limitations of the underlying mobile app platform: Some content-forward apps like Bing News and Bing Finance were pretty, but Metro-style apps were clearly never going to be as sophisticated and full-featured as their desktop-based predecessors.

As for the Windows Store, it featured the same inefficient panoramic interface as Windows 8’s Start screen and the initial app selection was unsurprisingly light and lacking in the professionally-made apps that users were then getting from Apple’s App Store and Google Play. Highlighting this problem, Microsoft highlighted the winners of a Windows 8 app contest that it had kicked off when it first launched the Store back in December, and they were predictably amateurish in quality.

There were some good ideas in Windows 8. It’s just that most of them were just implemented poorly, and the system would have benefitted from another year of testing and the changes that would have resulted from actually paying attention to the feedback.

It featured “at a glance” information on its Start screen, similar to what Microsoft had made available previously on smartphones with Windows Phone 7, allowing its users to get updates from apps without opening them, in turn, in a sequence of events I called “whack-a-mole” at the time. It supported app-to-app sharing capabilities that went well beyond the Clipboard copy and paste features that Windows had used for decades. And it was cloud-connected, supporting signing in to the system with a Windows Live ID and using that ID across OneDrive storage, Xbox games, music, and videos, Bing content, Internet Explorer, Windows Store apps, and more.

One of the more intriguing additions to Windows 8 was the notion that apps and services could be accessed via the new File Picker interface—a Metro-style File Open dialog, basically—instead of just local and network-based file sources. For example, you could point a File Picker at OneDrive and access your files in the cloud too. Or you could even mix and match, choosing one or more files from two or more locations.

Where Windows 8 fell apart was in how it interacted with the user.

Contrary to the claims, its Metro-style user interface—first created for the TV-based Media Center interface—did not scale well across different screen sizes and form factors, and it was useless if not aggravating on desktop PCs with large displays accessed with keyboards and mice. Likewise, it was unacceptable to almost all user types: power users felt constrained by its weird limitations, and beginners couldn’t even find the new experiences. Sinofsky also introduced the beginning of an ill-fated attempt to bring this user experience across different types of devices—PCs, phones, and the Xbox—an effort that would culminate in another disaster when his successor tried to use the same technical underpinnings on these and other device types. (That is a story for another day.)

In short, where Sinofsky sought to eliminate what he called the “hard stops” between different device types, all he and his team did was create a system that seemed to prove that each should, in fact, have its own unique user experience.

To wrap up the Consumer Preview event, Sinofsky discussed the work that Microsoft’s PC maker partners would do to address Windows 8’s unique new capabilities.

He claimed, correctly, that the Windows team engaged earlier than ever before with PC makers. The reasons for this were two-fold and related. First, the PC makers were upset by how long they had been kept in the dark about Windows 7, and the resulting first generation of Windows 7 PCs offered little new or specific to Windows 7 as a result. And second, the PC makers—with rare exceptions—had never shown much interest in following Microsoft down every rabbit hole it invented with each new Windows version anyway. For example, when Microsoft asked PC makers to create stereo component-like Media Center PCs several years earlier, only one HP model was available at launch and it was a traditional tower PC design that looked out of place in the living room.

But Microsoft needed the PC makers on board with Windows 8, and it needed to present a concerted effort to customers. And so Sinofsky had allowed his team to open up about some of the plans for Windows 8 much earlier than had been the case with Windows 7. In return, it asked them to strongly consider creating both tablets and touch-based Ultrabooks, and to release them in both ARM and Intel/AMD variants.

The lack of trust between the Windows team and the PC makers was so bad, however, that Sinofsky also hatched a divisive and controversial plan during the development of Windows 7 for Microsoft to create its own PCs. This idea was rightfully shut down by Microsoft CEO Steve Ballmer, the SLT, and the company’s board of directors repeatedly. But with the plans for Windows 8 mostly set in early January 2010 and Sinofsky’s star still on the rise thanks to strong Windows 7 sales, he got the go-ahead. And within a few months, the top-secret plans for what would become Microsoft Surface took form inside of the software giant’s hardware labs. If the PC makers couldn’t be trusted to fully embrace Windows 8 and its unique features, then Microsoft would simply do so itself.

But in February 2012, few outside of a small team at Microsoft even knew about these plans, and the firm’s unwitting PC maker partners were in fact prepping for the Windows 8 launch in a more studious nature than ever before. The annual Consumer Electronics Show (CES) had just occurred the month before the Windows 8 Consumer Preview event, and these hardware makers had shown off the final generation of Windows 7-based PCs at that time. Each would run Windows 8 well enough, of course, given the identical system requirements. But what Microsoft was most interested in was those PCs that featured multitouch capabilities and, ideally, tablet or tablet-like convertible form factors.

At the Consumer Preview event, Sinofsky’s ecosystem chief, Mike Angiulo, showed off a variety of third-party hardware designs, most of which were traditional laptops and Ultrabooks, a new form of Intel-based thin and light laptop. And he especially highlighted a handful of prototype ARM tablets that were perhaps indicative of the PC-like devices consumers would be offered that coming holiday season. But this segment of the show was a mostly monotonous half-hour that did little more than highlight how badly the Windows 8 Start screen and its apps scaled to larger displays.

With that out of the way, Sinofsky provided another pedantic update about where Windows 8 was in its development cycle and remind viewers that his team had continued to meet the schedule set out years earlier. The Windows 8 Consumer Preview would be the last major milestone build released to the public, he said, and it was for the most part feature complete, though the bundled apps would be iterated many times, both before and after the release of the product. Microsoft would provide an update about the business-focused features in Windows 8 at CeBIT that year, but would not release a milestone build at that time. And then it would ship a Release Candidate, followed by the final RTM (release to manufacturing) build. Windows 8 would ship in 2012 as originally scheduled.

“So that’s going to be our pattern, and we’re going to sort of commit to executing on that pattern today,” he said, noting that the Consumer Preview downloads had been uncorked while they had been talking. He claimed that he “was pretty sure the network hiccupped” right when that download went live—alluding to a brief and perhaps invented incident Larson-Green experienced—but Microsoft’s download servers were in no way related to the connectivity at the event in Barcelona, so it’s unclear what to make of that statement.

Of course, in the context of the lies Sinofsky had already told and would continue to tell, this was a minor incident. But it is perhaps interesting that he would continue to try to frame what was happening with Windows 8 as if it were somehow as momentous as Windows 7 and would thus automatically be another success. And that these kinds of slips would occur again and again as Windows 8 barreled towards an uncertain future.

The Windows 8 Consumer Preview got off to a great start. Microsoft reported the next day—via Twitter, no less—that it had seen over one million downloads in the first 24 hours. And the reviews were surprisingly positive.

“Windows 8 is … a huge radical rethinking of Windows — and one that’s beautiful, logical, and simple,” noted Apple fan and Microsoft critic David Pogue wrote in The New York Times. “In essence, it brings the attractive, useful concept of Start-screen tiles, currently available on Windows Phone 7 phones, to laptops, desktop PCs and tablets. I’ve been using Windows 8 for about a week on a prototype Samsung tablet. And I have got to tell you, I’m excited.”

“I’m impressed by what I see,” USA Today’s Edward C. Baig wrote in his own review. “We really are on the threshold of a whole new era of personal computing.”

I was surprisingly positive about the Consumer Preview in my own reporting and writing, at first. But I did offer the following warning as well.

“The big question with Windows 8 is whether users will embrace a system that basically has two UIs: the new Metro-style Start screen and the legacy desktop,” I wrote at the time. “Interaction between the two isn’t particularly seamless, but it does provide a way for users to utilize their stable of current applications while using new Metro apps as well.”

And there were the first rumblings of discontent from Microsoft’s userbase as well. The issue, I responded, was whether one believed that Windows 8 offered dual or dueling interfaces.

“The enthusiast side of me is ecstatic that Microsoft is at least going for it—really, really going for the gold—with this one, and doing something truly different, something that retains the past advantages of Windows while potentially positioning it to thwart the tablet threat and do to that market what the company previously did to Linux in the netbook market,” I noted. “The realist, or pragmatist, in me worries that Windows 8 is disjointed, that these two separate environments will never be crossed, and that Microsoft is essentially creating two different products that will coexist, weirdly, together. And that which you use more will depend on which type of product you buy: tablet users will exist primarily in touch-friendly Metro, while traditional PC users will exist primarily in the desktop.”

It wasn’t clear at the time whether Microsoft, its partners, and its customers would embrace or reject Windows 8. But either way, it was barreling forward on Sinofsky’s preordained schedule. And it would do so largely unaffected by the reaction, good or bad.