Metro? What Is Behind It All?

This is my view on what Microsoft’s actual strategy is, though it’s not being executed that well because I think they want to keep it secret and they didn’t come up with a good alternate explanation. I’ll note that my opinion is shaped by me being an Android fan, and that I now use an Asus Transformer Prime for all of my content creation and consumption (it’s a tablet with a dock to turn it into a netbook with keyboard touchpad, mouse support, and 10-15 hour battery life — I RDP or VNC into Windows or Linux machines for the few tasks that don’t have native Android apps, yet [surprisingly few, already]). Microsoft wants something *like* the Asus Transfor Prime, but not quite the same.

Consider for a moment what Android 4.0 is actually *capable* of: Your applications’ interfaces are defined in an XML-based format (like XAML for .NET) that *also* implements a functionality similar to CSS media queries: You specify different layouts depending on whether the application is being viewed in portrait or landscape mode, but also depending on the resolution of the device, so the same application can run well on tablets or smartphones.

Combine this with WiDi and a sprinking of imagination: rather than have a smartphone, tablet, laptop, and desktop, why not take your smartphone and "dock" it with a WiDi-enabled display and a bluetooth mouse and keyboard. Android 4.0 supports both touch-based and mouse-and-keyboard based input (I use the former for games, movies, books, web pages, the latter for productivity apps like IDEs, email, Google Docs for Android, etc on my Transformer).

Not only do you get the "full" desktop experience when docked, but when you undock and go to lunch, your apps are still there, but with a phone-optimized display, and you can show your progress, say, to a colleague over lunch in portrait mode, then you can go to a presentation after lunch, WiDi "dock" your phone with the projector to present a presentation, and change slides with a finger swipe (remember, no cables in the "dock", so you can walk around with it) or highlight key portions with a tap and see the relevant portion light up, then you go home and "dock" your phone to your TV and watch netflix, using the smartphone as a remote, then dock it to your personal "dumb terminal" of a touchscreen display, mouse and keyboard to check email, browse the web, play a first-person shooter, or whatnot.

In Android, you don’t worry about application management — the application adjusts to the orientation and screen size, available input devices, and they all have an auto-save functionality since an Android application can be killed at any time since Android’s task scheduler frees out running applications not in use when another one needs more resources ("background" processes like listening to Pandora are considered to still be running, web browsers are not and have to restore all of your previously opened pages when restarted).

Since Android and iOS can’t count on the mouse-and-keyboard, their interfaces have to be designed for the lowest-common-denominator, but work under a "progressive enhancement" situation where having a physical bluetooth keyboard means the on-screen keyboard doesn’t appear when you click/touch a textbox, for instance. Microsoft Office for Metro follows this principle: the interface now has ways of doing pretty much *anything* with the touchscreen, albeit not as well, but still has the keyboard shortcuts when one is available to make things faster for that case, but the interface can’t *assume* it’s there, or will remain there.

And that starts to delve into Microsoft’s plan: assuming Microsoft also understands this user-friendly future I described where application state is preserved across application restarts, and applications automatically adapt to changing input and output formats as you go from home to work to wherever and can at least *access* your data whenever, and using the same applications (albeit with adapted interfaces) wherever.

Windows currently can’t do that, so they took Metro from Windows Phone 7 and adapted it to handle this scenario. But Microsoft understands the network effects of application marketshare — currently their Windows Phone marketshare is abysmal because they don’t have the apps users have on Android or iOS.

Intel’s experiments with Atom in smartphones could bring back the "love" in the Wintel love-hate relationship: A smartphone that can do everything Android 4.0 can do as described above (adaptive interface, never lose your data by application state tracking, etc), *and* when you "dock" with a display, mouse, and keyboard, you can use the "classic" tile to run Windows 7 and earlier applications and have access to the huge catalog current Windows applications.

But why is Microsoft so "pushy" with Metro? Because they know that if they don’t start growing their catalog of Metro apps that have all of these features, and only have the Windows 7-style applications that can’t be used on the smartphone screen, TV, or projector without a mouse and keyboard also docked, then even if they gain marketshare with Windows 8 phones temporarily, they will lose it to Android as new development ceases for Windows and legacy enterprise apps are accessed by RDP (like I do with my Android tablet).

Android 4.0 shows how a unified experience could work fairly well across all interface dimensions and types, but it requires a complete rethinking of how applications are designed and developed. Windows 8’s Metro (different from Windows Phone 7’s Metro in technology and some key pieces, but designed to look similar) is Microsoft’s answer to that, and the classic Wintel marriage just may survive the recent bruising if Microsoft’s plan is to have a Windows 8 smartphone that can run classic Windows 7 apps when docked, while pushing for the new era of computing where applications don’t lose state and work on any screen size to maintain Windows’ dominant marketshare in traditional, productivity-oriented computing.

The alternative is a slow death to Android as it grows up and takes over more and more of what Windows could once do while not sacrificing the new things it can do, itself.

There is one last sliver of evidence towards this narrative that I couldn’t figure out how to fit above: Windows 8 is going to refuse Metro apps from being able to generate executable code — which means the death of non-IE browsers in Metro mode. Why is this significant? The wildcard in all of this that could still sway things in Android’s favor even if Microsoft plays all of its other cards right is WebGL. WebGL promises no-install-needed gaming through the browser on any platform of your choice. There is a significant contingent of gamers who stick with Windows for games, normally written to Microsoft’s DirectX API on desktops. But on phones, Android and iOS both use OpenGL ES (which is what WebGL is based on). If WebGL can run in a Metro browser, then it could run at full screen on the smartphone screen or TV without fuss, and a major lure for a segment of Windows consumer users would disappear, as developers would decide to use an OpenGL-based game engine that compiles to iOS, Android, and WebGL.

Not only would the games no longer be Windows-exclusive, they would have the performance penalty of being interpretted versus native. By elminating program-generated code, web browser JITs aren’t possible, so the game interpretation would be too slow to tolerate, even if it could run (it couldn’t because the browser couldn’t compile the OpenGL code to pass to the GPU, since each video card has a different assembly and must be compiled for each, so the WebGL standard, just like the Javascript standard, assume the browsers will properly interpret the raw source code provided).