The Biggest Problem with Windows Today? The Past

Like previous versions, Windows 10 is an amalgamation of the old and the new. But today, more than ever, it is the old that is holding Windows back. And I think we’re overdue for a change.

This mix of old and new has been the hallmark of Windows for decades, it’s the secret sauce that has powered the success of this product line for both businesses and individuals: Through the simple act of buying a new PC, we were able to take advantage of new functionality in both the hardware and the software while retaining compatibility with all of the applications and peripherals we already owned. This has made Windows a safe bet, and combined with the choice of PCs types and price points, it has made Windows the right bet.

But there have always been competing forces pulling at the Windows ecosystem from opposite ends. And as time has moved on, those forces have pulled ever stronger. The first, of course, is the Mac, which sits at the high end of the market and has attracted ever more users steadily over the years thanks to its “it just works” reputation and a halo effect from Apple’s super-successful iPhone handset and other devices. And then, from the low-end of the market, we’ve had Linux-based netbooks in the past and, today, Google Chromebooks. These devices have succeeded (albeit in a very limited way) almost solely through low-ball pricing to date. But with the advent of Android app compatibility, we see the very real possibility of a range of PC-like devices that are much less complex than PCs while offering true PC-like functionality.

Meanwhile, Windows, macOS and Chrome OS have all felt the impact of the far more popular mobile platforms that have made the PC market almost an also-ran. And each platform maker, in turn, has responded by adding mobile functionality to these desktop products with varying results. As noted, Google is adding Android app support to Chrome OS. Apple has added many iOS features to macOS, though it has done so slowly, almost prudishly. And Microsoft? Today, Windows 10 is the ultimate hybrid platform: Not only does the desktop version of the OS utilize many mobile features, it offers unique transforming capabilities that the others can’t match.

But this transition, this revolution, has come at a cost. Windows 10 is also an unwieldy mess, a legacy software product that is serviced like a mobile platform, often to disastrous results. It’s a grand idea, but the implementation is held back by the technical realities of having to support the legacy systems that still beat at the center of this platform.

I’ve written a lot about the servicing issues with Windows 10, but I’ll point you to the two articles that are perhaps the most relevant to this discussion: Windows as a Service Isn’t Working and Microsoft, It’s Time for a Reliable Computing Initiative. Those titles should provide all the information you need about my stance, but I’ll boil it down to something even more general: Microsoft isn’t very good at “finishing.” That is, it ships … and then it neglects. And it’s the users who suffer when Windows isn’t updated reliably.

But it’s not just servicing. As I noted in Microsoft May Never Fix the High DPI Issues in Windows 10, the legacy Win32 desktop applications that most of us still rely on need to be updated to support modern PC features like high DPI displays. But most never are, or are only updated in more recent versions for which you need to pay. For example, my copy of Photoshop Elements 11 still works perfectly fine, except that it does not support high DPI displays, rendering its tiny icons and menus as unreadable and unusable on modern devices like Surface Book. But Photoshop Elements 14—a roughly $100 upgrade—supports high DPI. While offering literally no functional upgrades that I really need.

Legacy desktop applications fall apart in other irritating ways for users on modern PCs. If you load up your desktop with applications on a tablet or 2-in-1 PC and then rotate the screen, all of the on-screen windows flip into an unholy jumble. And when you rotate it back, they don’t return to their previous sizes and locations; you just screwed up your entire desktop.

There is only one solution to these problems: Get rid of Win32.

That is, Microsoft needs to be more aggressive about ridding Windows of the legacy Win32 technologies that still form the very heart of Windows. It needs to establish a timeframe for this transition. And over time, it needs to continue supporting legacy technologies only in long-term servicing branches of the OS aimed at enterprises going forward.

And that, of course, is never going to happen. Yes, Win32 is the Achilles Heel that holds Windows back, but it is also the only thing making Windows viable as a platform today. Windows Store apps and the Universal Windows Platform (UWP) are going nowhere fast, and we as users rely on legacy Win32 applications because they still work and still work better than the Windows Store equivalents when they’re even available.

The kicker here is that Microsoft already tried to do this, too. What I’ve basically described is Windows RT. Or at least a more modern, Windows 10-based version of Windows RT that runs on Intel (and ARM) hardware.

I once described Windows RT as “the new Windows NT” because it represented a clean break with the past. Windows RT offered almost all of the in-box Win32 apps one expected from Windows—WordPad, Paint, Calculator, and so on—while offering full compatibility with a Windows Store-based app platform that never took off. It looked and worked just like “real” Windows. But it didn’t run third-party Win32 applications like Apple iTunes, Google Chrome, and Adobe Photoshop.

We still need Windows RT. But we also need a more common sense way of getting there.

The way I see this is that the core part of Windows, the thing I think of as Windows RT, is the default. And if you want to run Win32 applications on that platform, they will need to packaged into Centennial-like containers to ensure that they don’t impact the rest of the system, and can be removed cleanly when you’re done with them.

Over time, the goal should be to decrease our reliance on these applications, and to make it harder and harder to even use them. Win32 applications—yes, even the ones I rely on every single day—need to go away, for good. They’re holding all of us back.

This is the sort of decision making that Apple excels at, and while it has come under fire for what some perceived as a too-fast removal of legacy technologies, Microsoft is in a position to do something more rational. But also something that is more rational than forever supporting the technologies that are screwing everything up. A middle ground.

In hardware form, that middle ground is something like the 2016 HP Spectre x360, which utilizes two USB-C/Thunderbolt 3 ports in a nod to the future, but also provides a single USB 3.0 port for the current. This is rational, and customer-focused, and it’s not hard to image the next x360 dropping the legacy port. Microsoft needs to do the same, with Windows, in software. It requires, yes, the courage to do the right thing, but also the smarts to communicate what it’s doing effectively. And then the doggedness to just do it.

We spend far too much time worrying about the surface-level baloney in Windows, about the color of a File Explorer icon or the position of the Action Center icon in the system tray. There is more important work to be done, however, and while Microsoft is indeed doing some of it—look to the Unified Update Platform (UUP) technologies its rolling out over the next year as an example—I feel like there are even more fundamental changes that need to occur.

Microsoft needs to start the long goodbye to Win32 and step boldly into the future. The transition will be painful, as all transitions are. But the result will be a smaller, faster, simpler, and superior platform that meets the needs of users in this mobile first, cloud first era.

Let’s do this thing.