Smartphone and PC? Microsoft just might pull it off

The technology approach is proven, and the ambition has long been there, so there's a chance this time for the merger of PC and mobile to actually happen

Microsoft has long desired to create universal Windows—Windows on every device imaginable. After several failed attempts, it’s trying again—this time with pocketable PCs that will be able to run standard desktop Windows apps, not just Windows Mobile apps.

Microsoft’s technology approach to universal Windows—in fact, to a universal PC where a smartphone transforms into a full PC when connected to large monitors and other periperhals at your desk—looks like it could actually work this time, unlike the failed attempt several years ago known as Windows RT whose limitations are too many to recap.

The new horizons of the universal PC

But say the universal PC ambitions work out this time. Does it matter?

I think it does, because both the PC and mobile worlds are simultaneously intertwining and seeking to blast off their current innovation plateaus. You can see that desire to morph in the hybrid tablet/laptop devices (aka 2-in-1s), keyboard-equipped tablets (aka tabtops, like the Surface Pro and iPad Pro), Chromebooks running Android apps, Windows Continuum-based pocketable PCs like HP’s Elite x3, and other experiments that have shown up in the last few years.

None of those experiments have taken the market by storm. 2-in-1s and tabtops have inspired niche audiences, but no more than that. Chromebooks remain small-time after a decade of sales. A few years ago, tablets like the iPad seemed poised to replace the laptop for most users, but their sales stalled, and they remain a companion device instead. And Continuum-based pocketable PCs suffer the same problem as Windows RT did: They can’t run real Windows apps, so why bother with them?

Microsoft hopes to change that with its Windows-on-ARM effort, porting Windows 10 to the ARM chip used almost universally on smartphones, tablets, and other mobile devices. This time, it’ll be the real Windows 10, not a limited subset, and using emulation it will be able to run most standard Windows apps (those .msi and .exe files we all know and love), not just the handful that use Microsoft’s new Universal Windows Platform’s .appx format.

In other words, it would be Windows RT done right. And it would be Windows RT taken to the next level, running on smartphones and automatically adjusting its user interface (and those of its apps) based on the screen being used, similar to how the universal apps in iOS auto-adjust between the iPhone and iPad. But the universal Windows would be able to adapt from smartphone to tablet to big-screen desktop—a big contrast to Apple, which so far is keeping the Mac out of its universal ambitions.

That flexibility and adaptability would be as major a change in the computing ecosystem as the smartphone was nearly a decade ago. And it could lead to the same degree of seismic shift in user behavior and application capabilities that we saw with the iPhone and later with Android. Imagine that you no longer have multiple devices with different capabilities, but one device that adapts itself and its apps based on the peripherals you have connected.

Of course, that’s very much what Microsoft originally promised for Windows RT, but didn’t deliver. So healthy doubt is a reasonable reaction to the latest promises.

How Microsoft can deliver a universal PC experience

What Microsoft is doing this time on the technology front is to take a page from Apple, which has successfully migrated MacOS and Mac apps from one processor to another: Motorola 680x0 to IBM PowerPC to Intel x86. Apple used a combination of operating-system ports and emulation to support multiple processors, so both old and new gear could run MacOS and Mac apps. It also encouraged developers to create apps compiled for both, running whatever part was compatible with the current computer. After a few years, Apple would retire support for the old processor in MacOS, and apps could jettison the unneeded code.

So, developers and users have already been down this road successfully. If Microsoft’s engineers are as good as Apple’s, it can take us down this road again.

This shift is major for Microsoft. For seemingly ever, it has followed the strategy of separate operating system all named Windows for each type of device, universal in branding only. Nearly two decades ago, it had Windows CE (Compact Edition) in addition to the Windows everyone knows from PCs. WinCE (whose name has changed several times since, and now has been rebranded for the internet of things) was meant for embedded devices (now called IoT devices), such as package scanners. Microsoft also had Windows Mobile, for personal digital assistants, messaging devices, and other smartphone precursors; it then replaced it with an incompatible operating system for smartphones called Windows Phone (now renamed Windows Mobile and basically moribund until Microsoft figures out the universal PC). 

Microsoft is also taking another page from Apple: the adaptable app. Apple’s Xcode development platform lets you create apps that can run on both an iPhone and iPad, choosing the right user interface for the device at hand (that’s what Microsoft’s recent Continuum technology does as well), as well as use capabilities on devices that support them (like cellular connections and the fingerprint scanner) but not on the (usually older) devices that don’t. That’s what Microsoft’s Universal Windows Platform also does.

Apple has chosen to keep iOS and MacOS separate, so even though you can develop iOS and MacOS apps from the same Xcode code base, they’re separate, incompatible apps. UWP can run modern Windows apps on both Windows Mobile and Windows 10, bridging smartphone, tablet, and computer. But there are very few UWP apps, and standard Windows apps can’t run on mobile devices. Microsoft’s failure in the smartphone market kept developers from using UWP because, well, why bother?

That’s why Microsoft needed to find a way to have both the standard Windows and standard Windows apps run on mobile devices, meaning ARM-based devices. And that way is supposed to be the coming Windows 10-on-ARM effort, with the first products expected next fall.

It won’t be seamless, however. ARM-based devices won’t be able to run 64-bit standard Windows apps, just 32-bit ones. Microsoft has had a terrible time getting 32-bit and 64-bit apps to work well on 64-bit versions of Windows, and the fact that Windows came in both 32-bit and 64-bit versions only ensured that users avoided the 64-bit version that initially required special hardware and developers avoided 64-bit apps because so few users had 64-bit Windows. Apple handled its 64-bit transition much better, making MacOS “fat,” so it ran 32-bit on 32-bit Macs and 64-bit on 64-bit Macs—and applications were likewise encouraged to be fat to run optimally on either environment.

Microsoft’s bungling of the 64-bit conversion and its suicidal self-restrictions in Windows RT are why I am not fully confident Microsoft will do it right this time. Even though both were obvious failures off the bat to everyone else, Microsoft for years stubbornly pretended everything was fine. Although CEO Satya Nadella has purged much of that attitude from Microsoft, it still exists throughout the company and could pop up here again.

Still, Microsoft is taking proven paths on the technology front to universality. That in and of itself is a giant leap forward we should all eagerly watch.

Related: