Microsoft’s software plan for the Duo Android phone is surprisingly realistic I am really into Microsoft’s developer tools for a lot of reasons — especially the various proposals for making the web work better on dual-screen devices, which in theory could help everybody. But the most important thing is the overall context: Microsoft has the horse and cart in the right order. It’s trying to get the software right before it releases the hardware. There have been two big problems with foldable devices thus far: 1. the screens are too fragile and 2. Android is not great on tablets and so the windowing systems have been kind of bad. (And, well, a third big problem is that they have been super expensive.) I have no idea when the fragility thing will be fixed, but I like that Microsoft isn’t bothering with a flexible display. It compromised on whiz-bang hardware to make something more durable and, in many ways, elegant. But the trade-off is that there’s a big ol’ seam between the Duo’s two screens. That’s the cart. The horse, then, is how the software is designed to deal with that trade-off. (This is a bad metaphor because I don’t know what goes in the cart but we’re in too deep to turn back now.) The details of Microsoft’s answer to “how does Android work on a dual-screen device” all seem really smart. Windows Central’s Zac Bowden installed the emulator and made a little video showing how windows move around and it’s refreshingly simple. Apps open on a single screen, you go into the multitasking view and drag them to move them across to the other screen, or you move them over the seam for some kind of split-screen. There are different ways to split-screen: sometimes there’s a list on one side and details on the other, sometimes there’s two pages like on a book, and sometimes the canvas covers the whole thing and you just have to deal with the seam. All that is fine, but it’s not the smart part. Just because Microsoft appears to have created an elegant SDK doesn’t mean that anybody will actually use it. We’ve seen Microsoft try and fail to woo mobile developers before. RIP Windows Phone, we still miss ya. But for the Duo, it’s even worse than that. We’ve watched Google struggle to get Android developers to make better big-screen layouts for their apps for years to disappointing results. Android tablets have gone the way of the dodo and Android apps on Chrome OS are best used in small doses. So the way Microsoft appears to have dealt with that reality is one reason that I’m actually more hopeful today than I was yesterday about the Duo’s chances. That’s because even if literally nobody customizes their Android apps for the Duo, it should still work pretty well. Instead of pinning the Duo’s chances on the nearly impossible task of getting Android developers to invest resources in a completely new and untested phone, Microsoft is working with where the ecosystem is today. The key reason is that Microsoft explicitly says that apps will only open on one screen by default and in fact, apps will not be allowed to open up on both screens — that can only happen if a user drags a window into that state. Your app by default will occupy a single screen, but users can span the app to cover both screens when the device is in a double-portrait or double-landscape layout. You can programmatically enable full-screen mode for your app at any time, but spanning is limited to user activity for now. It has the very practical benefit of working better with existing Android apps by default. Instead of being annoyed that many apps are kind of junky and poorly-designed in a tablet screen context, the entry experience will just be two normal Android apps, side by side. Android apps generally look alright on portrait, phone-style screens — and that’s the way they’ll launch on the Duo. So even in the worst case scenario where only Microsoft’s own apps are aware of the hinge, the Duo will still work. It’s like the theory of progressive enhancement (and graceful degradation) in web design, but applied to dual-screen Android apps. It’s smart because, frankly, the worst-case scenario also happens to be the most likely scenario at launch. Only allowing users to choose when to make apps span two screens adds a level of predictability that will be important for users to built up their intuitions for how things work on the dual-screen device. (Side note: I have a whole rant about how there’s no such thing as “intuitive” design in software, it’s all learned.) Assuming it all works, users won’t be forced to learn a whole series of gestures and layouts and grids and whatever. Instead, they’ll just be able to move stuff around and let the software do the right thing. It is, pardon the alliteration, programmatically pragmatic. None of this guarantees that the Duo will be any good or that my relative optimism will be rewarded. I’m just glad that Microsoft isn’t setting the whole situation up for immediate failure from the jump. There’s simply very little chance that a ton of Android apps will be customized for the Duo’s dual screens for launch, but that hopefully won’t matter. Speaking of things that aren’t guaranteed: Windows 10X. The developer tools for that OS are still forthcoming and the questions about how it will operate are much more numerous than for the Duo. Given how many PC manufacturers are waiting for that OS for their foldables, the stakes for Windows 10X are much higher. As Tom Warren noted yesterday, we should expect to see more at Microsoft’s Build developers’ conference in May. If there were ever a time for Microsoft to be a little less hand-wavy about 10X, that will be it. | 
