Yesterday, Microsoft continued down a path that they’ve been pursuing for awhile by providing even tighter ties between Windows and Linux–including allowing running unmodified Ubuntu binaries directly in Windows. Reactions were, to say the least, varied; many people were preparing for the apocalypse, others were excited about being able to use Unix tools more easily at work, and still others were just fascinated by how this was technically accomplished. These reactions mostly made sense to me.

One did not. Especially on sites like Hacker News, many responses were screaming that people needed to be scared, to remember Embrace, Extend, Extinguish, to run for the exits as quickly as possible.

I find this reaction frustrating and depressing, not because it’s offensive, but because it’s so obviously incorrect and intellectually lazy that it gives me a headache.

I want to do two things in this blog post: convince you that Embrace, Extend, Extinguish is a grossly invalid understanding of Microsoft’s strategy; and convince you that an alternative strategy of Separate, Support, Serve provides a much better lens to view the modern Microsoft.

The Death of the Three Es

I’m not going to try to persuade you that Microsoft isn’t evil–if you believe they are, you’re wrong, but I don’t honestly care–but I am going to explain to you that, even if Microsoft were still evil, they would still not be doing Embrace, Extend, Extinguish.

First, I want to quickly remind you what the computing landscape looked like when Microsoft was using that strategy. Windows ruled everywhere, in a way that’s almost impossible to imagine today. Virtually all desktops everywhere ran some flavor of Windows. Mac OS, while arguably more usable than Windows, was technically inferior, and had such an app shortage (especially in niche spaces) that it was largely irrelevant. This in turn meant that Windows also ruled most of the back office. Paired along with the Office monopoly, Microsoft really and truly had a total lock on the personal computing space. It was basically impossible to use a computer without interacting with at least one Windows device in the process.

In that epoch, Embrace, Extend, Extinguish made a hell of a lot of sense. The idea was simple: if Microsoft saw a technology that threatened Windows, they’d embrace it (make it available on Windows), extend it in such a way that the best way to use the threat was Windows-specific, and, once most uses of the technology were sufficiently tied exclusively to Windows, extinguish it.

When Microsoft was a monopoly, this was a superb strategy to protect that monopoly. If they saw a threat, then bringing the threat in-house and tying it to the Windows platform was a great way to ensure people couldn’t leave, even if they wanted to. In effect, your alternatives had a tendency to evaporate before you had a chance to use them.

But Microsoft is no longer a monopoly. Hell, in many key areas, they’re effectively a non-player. While it maintains a plurality in old-school personal computers, Windows Phone is bascially a failed project, the cloud is all but Linux-only, and even the entire existence of the back office has been threatened by tools like Google Apps and other hosted solutions. They’ve even lost most kiosks to custom Android variants, and most developers to OS X. It’s now surprisingly rare that I do interact with a Microsoft system on a normal day, and I’m hardly unique in that.

This leaves us with two conclusions. First, empirically, Embrace, Extend, Extinguish failed; if it hadn’t, Windows would still be a monopoly. For Microsoft to be continuing this strategy, you have to believe they were not merely evil, but also unrecoverably stupid.

Second, it can’t work in an environment where Microsoft is an underdog. For nearly all shops out there, leaving Windows is honestly pretty trivial at this point; it’s adopting it that’d be an uphill battle. If I pick “Linux”, I can trivially integrate OpenBSD, Illumos, OS X, and any other Unix-like environment into my workflow with few or no issues. I can pick amongst AWS, GCE, Digital Ocean, and others for my hosting. I can pick virtually any language and database I want, use virtually any deployment tool, and migrate amongst all of these options with relative ease.

Windows is the odd one out. Adopting it not only means getting into a single-vendor solution, but also dealing with writing two sets of most deployment pieces, and dealing with licensing, and dealing with training my ops and IT teams with two radically different technology stacks. I’m going to need one hell of a value proposition to even think about it, and I still would likely turn it down to keep my ongoing maintenance costs sane.

Further, it’s a surprisingly hard environment for me to use as a developer these days even if I want to. If I grab a MacBook, I can write apps for iOS, Android, and Unix, all natively. If I grab a Windows laptop, I can’t target iOS at all, and I have to do any Unix development in a VM. This means that at Khan Academy, for example, I’d have to be insane to buy a Surface, even though I love the device; I’d end up spending all day in virtual machine running Ubuntu. It’s not impossible to use Windows, but honestly, if I have to spend all day in a full-screen VMware session, why bother?

In that environment, the old Three Es just don’t apply. They were about locking me into Windows, but we’ve long since passed that point. The problem Microsoft now faces is one of staunching the bleeding, and that requires a radically different strategy.

The Facts on the Ground

So: if you’re Microsoft, and you’re facing a world where you’ve largely lost all the current fights; where you’re losing developers left and right; where the challenge isn’t keeping people from leaving, but getting them to knock on the door in the first place; what do you do?

There are a couple of strategies that Microsoft could take in this environment, but I want to assert two key facts before we get going.

First, it’s very unlikely that Microsoft can stage a meaningful comeback at the OS layer in mobile, cloud, or server rooms at this point. We’re all now at least as entrenched in iOS and Android on mobile, and Linux on servers, as we ever were in Microsoft PCs. So if Microsoft is going to remain relevant, they’re going to have to do it in a way that meaningfully separates going Microsoft from going Windows.

Second, even if somehow they gained a meaningful foothold in those markets, it’s very unlikely they’ll be anywhere near a monopoly player in the space. iOS, Android, and Linux are so firmly established, and so pervasive, that any conceivable world for now is one where Microsoft has to get along with the other players. In other words, Microsoft-specific solutions are going to be punished; they’ll need technologies common to everyone.

If you agree with those two facts, their current strategy falls out pretty cleanly.

A Way Forward

First, Microsoft has to enable me to even use Microsoft technology in the first place. If Microsoft keeps tying all Microsoft technology to Windows, then they lose. If I have to use Windows to use SQL Server, then I’ll go with PostgreSQL. If I have to use Windows to have a sane .NET server environment, then I’ll pick Java. To fix that, Microsoft needs to let me make those decisions separately.

That’s indeed the first phase of their strategy: separating Windows from the rest of their technologies. SQL Server is available on Linux not to encourage lock-in, but because they need you to be able to chose SQL Server even though you’ve got a Docker-based deployment infrastructure running on RedHat. .NET is getting great runtimes and development environments (Visual Studio Code) for Unix so that I can more reasonably look at Azure’s .NET offerings without also forcing my entire dev team to work on Windows. This strategy dramatically increases the chance of me paying Microsoft money, even though it won’t increase the chance I’ll use Windows.

Next, Microsoft needs to do the reverse: make it feasible for me to use Windows as a development environment again. That’s where the dramatically improved Unix support comes from: by building in an entire natively supported Ubuntu environment, by having Visual Studio be able to make native Linux binaries, they’re making it feasible for me to realistically pick Windows even in a typical Unix-centric cloud-focused development shop. Likewise, Visual Studio’s improved support for targeting iOS and Android, and Microsoft’s acquisition of Xamarin, are going to go far to enabling me to do something similar on the mobile front.

In both of these cases, while there may be an “embrace” component, the “extend” part is notably missing from here–and it should be. Microsoft can’t meaningfully extend iOS, Android, or Linux in a way that’d actually matter to anyone at this point; it has to just support them on their own terms. And in that environment, it’s not possible to extinguish things; if Microsoft woke up one day and announced Xamarin was dead and gone, people would grumpily rewrite their stuff in Swift and Java, not suddenly announce that they were Windows-exclusive.

Finally, Microsoft still needs to make money, and they can do that by selling software as a service (Azure, Office365, and so on), rather than off-the-shelf. That not only gives them a steady revenue stream independent of their Windows installed base–after all, a person using Office365 pays the same whether they’re on Windows, OS X, or a Chromebook. It also provides insulation for them from any future platform changes. Does HoloLens take off? PlayStationVR? Oculus Rift? Will Microsoft catch the next wave? Who cares. As long as you’re using Microsoft products somewhere in your stack, they’ll be fine.

Separate, Support, Serve

It’s not as catchy as the original, and it certainly sounds a lot less ominous, but I think this can be summarized as the Three Ss: separate all of Microsoft’s offerings from Windows itself; support the reality of this heterogenous world when on Windows; and be the company that serves as much content as possible from its data centers.

I’m not saying that Microsoft can’t still lock people in some way. Apple definitely tries to lock in its customers with iCloud and iOS, and its developers with Swift, for example. But I do hope that this has convinced you that Embrace, Extend, Extinguish is dead–and, with it, at least some of the FUD about Microsoft’s software.

Jobs once famously said that Microsoft didn’t need to lose for Apple to win. Today, I think it’s worth realizing the reverse: Microsoft doesn’t require you lose for it to succeed.