Two Paths, One Destination (Premium)

We can debate the relative merits of Win32 desktop applications vs. newer platforms, but Microsoft is moving very clearly in one direction. The problem? It’s not sure how to get there.

The good news? It’s doesn’t matter how it gets there.

You have presumably seen the two blockbuster news items from today: Microsoft delayed Windows 10X to early 2021 and will not include the backward-compatible Win32 container in that first release, if ever. And Microsoft is working on a subscription version of Microsoft 365—including Windows and Office 365—that will be delivered from the cloud.

These two stories are related, and not just because both were first revealed by my Windows Weekly cohost Mary Jo Foley: Both point to a future that we’ve been not coincidentally discussing a lot in the past few weeks. Many will call this future a realization of the 1990s dream of thin client computing. But I think it’s a bit more nuanced than that. It’s really about making the client more heterogeneous, where Microsoft doesn’t at all care what you use as long as you’re still running Windows applications.

And running Windows applications is a problem.

As we’ve discussed many times, Microsoft has been trying to move past the legacy Win32 foundation of Windows and to a more modern, secure, and reliable future. But each step it’s taken in this direction—Windows RT, Windows 10 S Mode, and Windows on ARM being the most obvious examples—has failed, and for the same reason. In each case, Microsoft made something called Windows that looked and felt like Windows but could not run all Windows apps.

And as I’ve observed many times, a version of Windows that can’t run all Windows applications is pointless. In fact, it’s arguably not even Windows.

To get from here to there, so to speak, something has to change. To date, the possibilities—that Microsoft’s customers have to stop using the legacy Win32 desktop applications on which they rely, or that Microsoft has to simply drop support for legacy technologies and evolve the platform more rapidly—has not worked. Microsoft’s customers do rely on legacy apps. And Microsoft will not abandon them.

But there is another way. Two other ways, really.

The first is the virtualization/container approach that was/is promised by Windows 10X. This platform would still run Win32 desktop applications, but would corral them in a technical container to prevent errant code from infecting the wider system and harming security, reliability, and battery life. It’s a good idea, and because it relies on a rich client—i.e. a real PC—it’s good for the industry. PC makers only make money on expensive premium products.

The second is to deliver those applications via the cloud. This requires two things: A substantial cloud infrastructure, which Microsoft has in Azure, and pervasive high-speed Internet, which I think we can all agree is hit or miss depending on where you are.

That second reality makes cloud delivery of apps—via Microsoft Cloud PC—an iffy proposition for individuals, at least for the short term. But it could emerge as a business-only option at first and then expand to consumers once 5G is more widespread.

But with Windows 10X dropping support for Win32 containers, at least in its first shipping version, that means that Microsoft is once again targeting this platform at low-end and inexpensive Chromebook-like hardware that can only run web and Store apps. In other words, it’s a platform for thin clients and not a true alternative to a rich client like Windows 10 for desktop.

And that second option, of cloud-based application delivery, adds an additional wrench. If Microsoft is going to deliver Windows apps from Azure, then users could conceivably use any client platform, including iPads, Android tablets, Chromebooks, and Linux PCs. Less well understood is that this could support rich Windows clients, too, with legacy apps running from the cloud, where they will be contained and unable to infect client PCs. It’s a nice solution to the compatibility issue that would let Microsoft advance Windows 10 on PCs with less worries about carrying forward technologies that in far too many cases date back to the 1990s.

In fact, it’s not hard to imagine future users combining the two plans, with Microsoft customers deploying Windows 10X-based thin clients to their users and then delivering legacy Windows desktop app from the Microsoft Cloud PC service.

As Brad and I discussed on today’s First Ring Daily, the weird evolution of Windows 10X, which is now back to where it was two years ago as a single screen, Chrome OS-like option, really brings things full circle. But the addition of the Microsoft Cloud PC service puts Microsoft’s plans into better focus. As a standalone OS with no Win32 app support, Windows 10X doesn’t make any more sense than did Windows RT, Windows 10 S Mode, or Windows 10 on ARM. But combined with Microsoft Cloud PC … voila. It’s starting to make sense.

But here’s the thing. The only part of this puzzle that matters to Microsoft in the long term is the second piece, the Microsoft Cloud PC piece. Ultimately, Microsoft doesn’t care what you’re using the client, and in some ways, it would be better for the firm if you weren’t using Windows at all because of the complexity and support costs. But customers paying a monthly fee to access Windows 10 apps, Office 365 apps, and other Microsoft 365 apps and services is exactly what Microsoft wants. If they can do so only on clients that Microsoft doesn’t have to support, all the better.

Yeah. It really is starting to make sense.