Programming Windows: Fast and Fluid (Premium)

By 2010, Windows’ fate was sealed. The web and the iPhone had disrupted personal computing, forever pushing the user base—and developers—away from the PC. And so Microsoft was left to decide what to do with Windows. It could coast and let the revenues from businesses roll in while their employees engaged in traditional productivity tasks on increasingly uninteresting and anachronistic hardware. Or it could try to escape the innovator’s dilemma and disrupt its own business, as Steve Jobs had done to the iPod when Apple announced the iPhone.

To its credit, Microsoft chose the latter route. Unfortunately, the man running Windows at the time—Steven Sinofsky—was politically divisive and ill-equipped to take on the role of product visionary. And the decisions he made would go on hurt Microsoft internally and, eventually, externally as well.

It started in January 2010, at least according to a highly placed member of Sinofsky’s team who will remain anonymous. In what can only be called an apocryphal story, they recounted how Sinofsky had gone to CES 2010 and witnessed everyone touching all of the screens at the show, many of which were smart TVs and PC monitors. Sinofsky excitedly called the team back in Redmond. “They’re all touching the displays!” Microsoft, he said, would need to really embrace multitouch in the next version of Windows.

In a version of this story that is probably closer to the truth, Sinofsky later recounted how he had been at CES 2010 with his team and was struggling with how Microsoft could shake Windows out of its rut. “The question was, should we go all in and try to build a device+touch-first platform/OS,” he wrote. “We knew Windows 7 had an easy decade of utility ahead, like XP [had]. What was really bugging me was if Firefox/Chrome was where to use the web and iPhone and Android were where to use apps, then the only reason to use a PC was for ‘legacy’ workflows. We’d seen this before: it was what happened to the mainframe and IBM. Plus, mobile was just better/modern.”

And so Sinofsky did what he always did. He wrote a memo. A really long memo.

“Much has been said about the current environment relative to Microsoft,” he wrote to Microsoft’s senior leadership. “Is Microsoft the legacy? Is it ‘game over’ for Microsoft in various aspects of this new state of the world? Many openly wonder if Microsoft is deliberately holding back on innovations because of a fear that these innovations somehow undermine the traditional business model. Many argue that Microsoft’s product development practices are themselves a barrier to participating in the new world. All of these share the almost cliché view of disruptive attacks on the establishment — on Microsoft. To me, this is a naïve view of how things progress and presumes a deer in the headlights view of Microsoft. It also presumes that we do not have the technical or organizational wherewithal to build new cool and exciting technologies.”

“Microsoft has transitioned through so many transitions that it is hard to recount them: from the multi-platform PC world, to the MS-DOS world, to the networked world, to the GUI world, to the client-server world, to the applications world, to the suite world, to the client-server email world, and many more.” (“Transitioned through so many transitions” is as awkward a turn of phrase as Sammy Hagar’s insipid lyrics for “Why Can’t This Be Love,” in which he sings “only time will tell if we stand the test of time.”)

“We have the capability to embark on a new journey,” he boldly asserted. “We need to do so deliberately and with purpose.”

This version of events is also at least somewhat apocryphal, too: aside from positioning himself as the hero, the most important part of this version of the story is that it happened before the iPad. In other words, Sinofsky wants the world to believe that Microsoft had already decided on the mess that Windows 8 would become—and on the Surface PC line, which we’ll examine soon—before Apple launched the iPad. But it’s important to know that the iPad was telegraphed several months in advance, and that the world knew full well Apple was about to launch a tablet that would resemble a large iPhone or iPod touch. Sinofsky’s goal to disrupt Windows would have been a good idea if he had thought of it before the iPhone arrived, and it might have even been the right one in its wake. But it wasn’t exactly an epiphany to make a “device+touch-first platform/OS” after finding out that the iPad was happening. As was so often the case at the time, Steve Jobs had already shown Microsoft the way forward.

Of course, Sinofsky wasn’t Jobs. And Microsoft wasn’t Apple. And so the plan that emerged was less disruption than it was a Frankenplan to graft a new “device+touch-first platform/OS” on top of Windows 7. That is, Microsoft wouldn’t create a new “device+touch-first platform/OS.” It would create a “device+touch-first platform/OS” and put it in on top of Windows. Windows 8 was simply Windows 7 plus this “device+touch-first platform/OS,” all in one. A thing on a thing.

(That Microsoft already had a “device+touch-first platform/OS” in the form of Windows Phone 7 and didn’t use that as the basis for a new generation of tablets is a decision that’s still worth debating today. But Sinofsky’s “not invented here” mentality pitted his Windows team against all the other teams within Microsoft, and his goal was to beat the Windows Phone team, not unify the company around something that they had created. And thus Windows 8 would be completely incompatible with Windows Phone 7 and vice versa.)

The result was first shown off publicly in mid-2011 and seeing the mistake Microsoft was making was likely one of the last true joys of Steve Jobs’ life, as the visionary would sadly pass away later that year.

Windows 8 was an aesthetic nightmare of poorly designed, ugly, and non-discoverable functionality that grafted an immature “touch-first” mobile platform on top of the classic and powerful Windows desktop, replacing the Start menu with a full-screen Start screen with large, flat, colorful, and animated live tiles. Sinofsky, Julie Larson-Green, and others on the Windows team would insist again and again that it was “just Windows,” and not additional layers on top of something else. But that was untrue: Windows 8’s new UIs were simply grafted on top of Windows 7.

Windows 8 also provided new full-screen mobile apps, called Metro apps, and users could switch between them and the classic desktop applications that everyone relied on. But it was impossible to use these different types of apps side-by-side, as Metro apps could only run in full-screen and not in a floating window on the desktop. Here again, the team insisted again and again that it was all “just Windows,” but what it had really created was two very different environments in a single product, and nothing highlighted that fact more clearly than the boundaries between Metro and traditional desktop apps.

Yes, Windows 8 also included lots of innovations under the covers: new features like Client Hyper-V, early load anti-malware, Refresh Your PC and Reset Your PC, Secured Boot, SmartScreen, and Windows To Go were impressive platform upgrades. And even the new Metro app environment—called WinRT for Windows Runtime—had its own useful innovations, despite not being based on .NET, an objectively terrible decision that sent shock waves throughout Microsoft’s Developer Division. But all of that got swept away by the Windows team’s decision that all users, even those on classic desktop PCs and laptops, would have to deal with the new full-screen experiences before they could get to the desktop that they understood, expected, and preferred. They were moving forward, and they would take the Windows user base with them, kicking and screaming if necessary.

This wasn’t at all obvious in mid-2011. Inside of Microsoft and without, Sinofsky still carried an aura of invincibility about him thanks to the tremendous success of Windows 7, which was racking up a suspiciously consistent 20 million new licenses each and every month by that point. Heading into the newly renamed Microsoft Build conference that September—Sinofsky had killed PDC, which was owned by his perceived rivals at DevDiv, and replaced it with his own show, just as he had taken control of the Windows 8 developer platform from them—most were simply willing to wait and learn how this would all work.

The warning signs were there, are in fact obvious in retrospect. But at the time, there was a belief that Sinofsky might actually pull it off. And no one wanted it more than him: if he succeeded again, he might rise ever higher in the Microsoft corporate hierarchy and perhaps even replace Steve Ballmer as CEO.

But that would be a battle for another day.

Sinofsky opened his Build 2011 keynote in September by revealing that Microsoft was “approaching over 450 million copies of Windows 7 sold,” and that, as important, Windows 7 consumer usage [was] finally greater than Windows XP usage.” He then made the curious claim that his team had somehow made “1,500 product changes to Windows 7 since we released it to manufacturing … not security updates, those are just things that added small things, that fixed features that you said weren’t quite right, and also addressed some of the challenges in the product.”

But he moved on quickly to the main event, Windows 8, and his rationale for why Windows needed to be reimagined in the first place.

“Things are a lot different than they were three years ago with computing, and they’re a lot different than they were in say 1995 or the last time Windows underwent a pretty significant and bold overhaul,” he said, drawing a comparison to Windows 95, which was still held up as the best-ever Windows launch. “Form factors and user interaction models create a whole new set of scenarios and opportunities for you [developers]. I mean, it’s incredible to think about the evolution of desktops to laptops, convertibles, to now small slate computers or, as we call them, Windows tablets.”

“I think touch is going to become a huge part of interaction, and what we’re going to see is something that I don’t think a lot of people are expecting, which is as soon as you’ve used touch on a PC, you want touch on all your PCs,” he continued. “So, people who say touch is only for small devices or it’s only for lightweight things, I promise you, the minute you use a touch device with Windows 8, by the time you go back to your laptop, your desktop, you’re going to be hitting that screen, and I promise you’ll have fingerprints all over your monitor if it doesn’t support touch.” That you would have fingerprints all over your touch-enabled monitors too was left unsaid. But Sinofsky’s vision of the future has never materialized. Some people like multitouch, some don’t. Do people miss it when it’s not available? No, most do not.

“Mobility is a whole new dimension to computing these days,” he continued. “It used to be that it was enough to be able to just carry your laptop around, put it down, plug it in and use it. But now you want devices that you can use while you’re carrying them around or just seated kind of uncomfortably or reclined. A whole new way of using computing has really arisen, and we want Windows to respond to that.” Sinofsky’s use of the term “has arisen” here is interesting as it echoed Steve Jobs’ “the question has arisen” assertion at the iPad launch the previous year.

“You know, you guys [developers] want much richer connectivity and sharing between applications,” he forged on. “The idea that applications are silos and don’t talk to each other or talk to each other through very narrow interfaces, that’s just not a rich enough interaction model for customers. They want much more. They don’t want apps to stand alone, they want a web of applications.”

“And we also learned that the world is different because services are an intrinsic part of the software that everybody uses,” he said, completing his explanation of the “whys” of Windows 8. “You just don’t write an app anymore if it doesn’t connect to some web backend, doesn’t share information or consume infrastructure from a service. So, it’s very important that we recognize this as we go and evolve Windows 8.”

And with that out of the way, Sinofsky explained that Windows 8 was “everything great about Windows 7” but better, something that built off the success of Windows 7. “We did something that we took a step back and we said, what’s the boldest thing we could say?” he asked rhetorically. “And what we said is we’re going to reimagine Windows. From the chipset to the experience, Windows 8 reimagines what Windows can be. And we mean so many things by that, and you’re going to hear that word a lot.”

For the Developer Preview reveal, Sinofsky would once again bring Julie Larson-Green onstage to demonstrate new features, many of which had been demonstrated previously at D9 and Computex. Larson-Green, like Sinofsky, was still awkward onstage—I still laugh out loud each time I hear the microphone capture her hissing to Sinofsky, “shhh… don’t look at me” mid-demo—but we did at least learn more about Windows 8 and its new features.

Windows 8 would incorporate a number of touch-based gestures, including so-called “edge gestures” that would trigger new UIs or features depending on which display edge was used. For example, one could swipe in from the right edge of the display to access the “Charms,” a set of icons representing system functions like Search, Share, Start, Devices, and Settings. Or you could swipe in from the left to display the Switcher, a column of thumbnails representing the running apps and other open windows.

You could personalize the new Start screen in a variety of ways, including dragging tiles to new locations—an action Green flubbed—using groups, and accessing a new “semantic zoom” feature that displayed the entire Start screen layout in miniaturized form.

There was a touch-first Control Panel—which would later be renamed Settings—and several new Metro-style apps (though it was unclear which would be included in the product and which were only for demonstration purposes).

One of the more limiting aspects of the new Metro environment was that these new apps had to run in full-screen unless they explicitly supported a thin, columnar “snapped” view that would let two of them run side-by-side at the same time. As originally conceived, the snapped view was a fixed width, so there was no way to resize the two apps, and not all apps supported it. As Green showed this off, Sinofsky interrupted her to explain that “of course, Windows is multitasking from the kernel on up, and so all we’re showing here is the fact that two applications can always be running in Windows. And this ability to dock them is a really cool opportunity for you to sort of create like a heads-up view of your application.” Calling a limitation like this “really cool” was a bit of a stretch.

Windows 8 would feature a cross-application communications system called Share, which was typically accessed via the Share charm. The idea here was that each Metro-style app could register itself as a Share source or target for specific types of data—like text or images—using a new system of Share contracts. “You can kind of think of sharing as a very semantically rich clipboard almost that when all of your applications can share with each other, even with applications that they didn’t know about when you write your application,” Sinofsky noted. “So, you don’t have to try to figure out how to get connected to everywhere else.” (This was true of the Clipboard, too, of course.)

Windows 8 would also feature a new full-screen Search experience that would let users search across apps, settings, and files. Metro-style apps could be written in such a way that their data could be searchable from this interface, which seemed useful. For example, you could search here to find information in an email exposed by the built-in Mail app.

“What you saw was fast and fluid,” Sinofsky said after Green had concluded her demo. “So, when we set out to design Windows 8, the whole idea is you should be able to move through things really quickly, touch-centric [or] keyboard and mouse if that’s what you want. It’s got to be responsive, alive, and beautiful. I mean, just from the minute you sign on and you see that Start screen, you saw everything’s moving, everything’s animated. Tiles are live. You can control that as a developer, and you can control that as an end-user of the system. But the rich animations, the Metro-style apps, it really feels like a different system. It’s fast, it’s fluid, it’s snappy, it’s responsive, it’s alive.”

Of course, what the developers in the audience really wanted to know about was the new developer frameworks and tools: to date, all we had heard officially was that Microsoft would let developers create Metro-style apps using web technologies (HTML, JavaScript, and CSS). But the assumption was that it would also offer .NET-based tools and frameworks, similar to Windows Presentation Foundation (WPF). On that note, developers would be disappointed: Windows 8’s developer story was even less well-conceived than its user interface.

We’ll look at that next.