What Microsoft’s Developer Initiatives Say About the Future of Windows (Premium)

Build is perhaps my favorite Microsoft event because it highlights two of my biggest tech passions: Windows and software development. But that may be changing, which makes this year’s show was particularly poignant. Now, Microsoft is marking a shift in which it emphasizes cross-platform capabilities over those that work only on Windows 10.

Even writing those words I need to pause and reflect. Folks, this is huge, and it may be months or even years before we can fully appreciate this shift.

But that doesn’t mean that analysis is impossible today. And perhaps the best way to put this past week in perspective is to look back over the previous few Build conferences and see what’s changed compared to this year’s show. We live in interesting times, and the way I see it, everything is changing.

Note: I’m focusing exclusively on client technologies here for the most part. That’s my primary interest, after all, and I have a lot less to say about cloud solutions like Azure.

One billion Windows 10 devices

The biggest news at Build 2015 was that Microsoft promised developers that there would be over one billion Windows 10 devices in the world within two-to-three years. A year later, there were 270 million Windows 10 devices, and today there are 500 million.

Microsoft, of course, reneged on its promise in mid-2016, noting that the original schedule was unrealistic and that it still expected Windows 10 to surpass the one billion devices milestone at some point.

Today, one can argue that 500 million is a big number no matter how you slice it, and that Microsoft has nothing to be embarrassed about. But that number is a red herring: The real issue with Windows 10 isn’t the number of active PCs, it’s user engagement. And as we see on macOS, too, by the way, just because there are lots of users doesn’t mean that they will embrace a mobile app store or its apps. And on both Windows 10 and macOS, engagement is very low.

This is a classic chicken/egg problem. The Windows Store is full of crap, with few professional apps, which turns off users. Meanwhile, because so few users actively browse the store, developers don’t see the point in making new apps, let alone using Bridge technologies (see below) to port their existing apps to the Store.

Let’s say you’re a developer who’s created a hugely successful Windows desktop application. You are already targeting 1.5 billion Windows 10 users, including all Windows 10 users. Why on earth would you port that app to the Windows Store in order to reach a tiny number of engaged users?

Now let’s say you’re a developer who’s going to create the next big software application. Why on earth would you target the Windows Store, with its tiny, non-engaged user base, when you could create a web app or a mobile app, either of which has several billions of potential users? Answer: You wouldn’t. And that’s what’s happening in real life.

Point being, that giant market of Windows 10 users never materialized. And even if it does materialize in the future, it’s not a highly-engaged audience anyway. The number almost doesn’t matter.

Windows Bridges

At Build 2015, Microsoft introduced four Windows Bridge technologies that were designed to help developers port apps and code from other platforms—the web, Windows desktop applications, Android, and iOS—to Windows 10, the Universal Windows Platform (UWP), and Windows Store.

Today, only one of those bridges, the Desktop Bridge (formerly Project Centennial), can be considered successful. And that’s unfortunate for Microsoft because it’s the bridge that is least likely to result in developers creating new UWP apps. (It’s arguably a big win for users, however.) But this project also helped bring Windows 10 S to market, and that OS version is, I think, the future of Windows. So it’s not all bad news.

It’s worth noting, too, that another of the bridges, “Project Astoria,” which sought to bring Android apps to Windows 10, was canceled. This happened because Windows phones ran Android apps too well, believe it or not. The theory at the time was that developers would never adopt UWP if they could simply run Android apps unmodified. But with Windows phone officially killed off and “pure” UWP development at a continued standstill, one has to wonder if bringing back Astoria and/or Android app compatibility doesn’t, in fact, make tons of sense. Just a thought.

UWP and Windows Store

Back in 2015, Microsoft still believed that UWP could be successful, and it was actively courting developers to this platform. At that time, the promise was that UWP apps could target PCs, tablets, phones, IoT devices, HoloLens, Surface Hub, and Xbox One. It also announced the Windows Store for Business at Build 2015, opening up this platform to its biggest customer base.

It didn’t work. In mid-2015, Microsoft formally capitulated the smartphone market to Android and iPhone and began winding down that business. The result? Of the form factors that Windows 10 targets, only the PC is a volume market, and as noted above it’s not an engaged audience. Today, there are only ~33 million Xbox One consoles in the world, compared to 500 million Windows 10 PCs. It’s not even worth discussing.

But Microsoft’s messaging on UWP and Windows Store this year was interesting, and confusing in its vagueness: It repeated the trope that the “Windows Store is a great home for all developers” and then it announced exactly three new apps that would be coming to the Store sometime this year. Only one, iTunes, was in any way interesting, and that app is 14 years old. That’s not momentum, it’s the opposite of momentum, and the lack of any hard adoption numbers indicating success is pretty much all you need to know.

Anyway, it’s clear to me that the future of Store apps, as we should call them, is with non-UWP apps, which is to say other apps packaged in AppX containers: Desktop Bridge apps, Progressive Web Apps (PWAs), and so on. I think this is clear to Microsoft, too, but maybe the firm felt that its Windows 10 strategy shift was a big enough change to reveal this year.

Xamarin

Microsoft purchased Xamarin just ahead of Build 2016, giving it some technical talent with cross-platform chops and, as important, the key to its mobile future. Since then, the software giant has moved quickly to make Xamarin completely free to all developers, integrate this technology into Visual Studio, add new cloud capabilities (on the Mac), and prop up its “mobile first, cloud first” strategy with actual, shipping product.

The genius of Xamarin cannot be overstated, and this year’s announcements—Visual Studio for Mac (formerly Xamarin Studio), Xamarin Live Player, and so on—only bolster that contention. So do the Microsoft Graph features in the Windows 10 Creators Update (see below) that work cross-platform with Android and iOS, as well as Project “Rome” on iOS and other related initiatives.

With Xamarin, Microsoft has the makings of a true cross-platform developer solution that leverages the skills that Microsoft/.NET-focused developers have grown over the years while providing them with access to popular mobile platforms in addition to Windows 10. Yes, that’s huge.

Microsoft Graph

We’re going to be discussing the Microsoft Graph a lot in the months and years ahead, I think, because this is the backroom technology that drives everything I’m writing about in this article.

A week ago, I never would have guessed at this future, and when Microsoft Graph first appeared as an Office 365 feature for commercial customers, I wasn’t particularly interested. But today, it is the underpinning of Microsoft’s cross-platform dreams. And it will work hand-in-hand with the Xamarin mobile client toolset to help developers deliver modern, AI-driven solutions to customers across multiple platforms.

Put simply, Microsoft Graph is a set of cloud services that brings in data from multiple sources, or what Microsoft calls inputs, and then uses AI and machine learning to make sense of it all. In an Office 365 environment, this is simple enough, and there Microsoft Graph works with a relatively limited set of Microsoft-controlled inputs—like email, schedules, documents, and people—inside of a single organization only.

That humble beginning in no way prepared anyone for what Microsoft announced this week. Now, Microsoft Graph is nothing less than the back-end AI that will drive everything we do within the Microsoft ecosystem and beyond, on PCs, mobile devices, and other devices, and for both consumers and business users. It’s the glue, if you will, that will bring Microsoft’s ambitions to life.

You may have seen the “Windows PCs love your devices” imagery. Each of the Windows 10 features that relate to this connectivity—Timeline, Pick Up Where You Left Off, Clipboard, and OneDrive Files on Demand—is an example of functionality that is driven by the Microsoft Graph up in the cloud. Yes, each relates to a single Microsoft app or service, essentially—Clipboard utilizes SwiftKey to provide cross-platform copy and paste functionality, for example—but the “glue” between your PC and your devices is the Microsoft Graph. These aren’t individual, isolated services. They’re part of a family of services. The Microsoft Graph.

As with its Office 365 implementation, Microsoft Graph is more powerful than simple one-to-one connections, is in fact at its best when it is working with multiple inputs. And those are the situations where AI and machine learning really come into play. I’m going to plumb Microsoft’s Build resources to come up with some more complex examples. It’s still early days but this is real Skynet stuff we’re dealing with here.

Cortana Skills

Over the previous two Build conferences, Cortana played a sort of ancillary role. It was, of course, expanded to Android and iOS over time, and to Xbox One. These efforts may have seemed pointless at the time—after all, you can’t fully replace Siri on an iPhone anyway, and even if you could, few Apple users would ever do so—but today we have a better picture of where this is all going.

First, Microsoft is finally and belatedly moving Cortana towards true ambient computing by working with partners to deliver Cortana-powered smart home devices like the Harmon Kardon Invoke. Second, and as important, it also finally released its Cortana Skill Kit in public preview form, providing developers with the tools they need to expand Cortana’s capabilities.

As with UWP/Windows Store, Cortana is kind of a chicken/egg problem, but we’ll need to wait and see whether Microsoft can attract both developers and users to this platform. In my opinion, they’re not doing a great job: Cortana only works properly in the United States, and its poor mobile penetration doesn’t help much.

Worse, Apple will almost certainly announce a dramatic expansion of Siri at next month’s WWDC, including probably a Siri appliance of some kind for the home. Say what you will about Apple, but it’s users love to buy anything it sells. A Siri device could immediately be very popular, leaving Cortana—like Windows phone before it—as an also-ran.

I won’t write off Cortana just yet, of course. But this Skills Kit should have come out a year ago. Microsoft can’t move quickly enough on this.

Windows Mixed Reality

Windows Mixed Reality is an interesting example of Microsoft being very honest with me privately and then getting up on a stage in public and declaring it the “next personal computing wave.” It is nothing of the kind: Mixed Reality is on a slow boil, and while I don’t ever feel it will be a mainstream computing interface, even if that happens, we’re still years away. (Cortana-type ambient computing is clearly the next personal computing wave.)

Worse, the very notion of “mixed reality” is, at best, an exaggeration when it comes to Microsoft’s new platform. Those Mixed Reality headsets are just VR headsets; there’s no “mixing” of realities there, no holograms, and certainly no real world.

That’s not to say that Windows Mixed Reality doesn’t have some advantages over previous VR solutions; the built-in sensors negate the need for external sensors found on other VR solutions like HTC Vive and Oculus Rift. But that’s just an evolution, not something new.

HoloLens, for all its viewport issues, is truly revolutionary. But the Windows Mixed Reality headsets, even the most expensive and powerful ones, will not deliver any HoloLens-style experiences. And as with other niche products, the potential audience here is small. I bet it remains small.

What’s next?

I have a lot of work to do around researching Microsoft Graph—I’m curious if Microsoft ever hinted at its amazing future, for example—and I need to get a better handle on some of the new cross-platform scenarios it enables.

But the big change at Build this year, I think, was a shift away from Windows at the center of Microsoft’s client story. It’s not that Windows doesn’t matter anymore, of course it does. But Windows’ role has changed to meet the way Microsoft’s customers really work.

As with many of you, it will take me a while to come to terms with that one, even though it’s really just an implementation of something the software giant has been communicating now for a few years. Like any other major change, it’s one thing to know something is going to happen and another thing entirely to actually experience that change.

Here we go.