The core idea would be to offer some kind of virtualisation layer / Dalvik compatible virtual machine that would allow Android apps to b run on Windows 8 and Windows Phone devices. This is similar to the approach taken by BlackBerry in its own support of Android apps, and is also being pursued by other vendors including Finland's Jolla.
The underlining motivation is to attempt to co-opt the Android app ecosystem, which is generally regarded as being much healthier than Microsoft's platforms, especially in the case of Windows Phone (both in terms of app numbers and when they are released).
Tom Warren, writing on The Verge, explains that there are opposing views internally at Microsoft, indicative of the controversial nature of the proposal.
We’re told that some inside Microsoft favor the idea of simply enabling Android apps inside its Windows and Windows Phone Stores, while others believe it could lead to the death of the Windows platform altogether. The mixed (and strong) feelings internally highlight that Microsoft will need to be careful with any radical move.
Mary Jo Foley, writing on ZDNet, points out that it is already possible to run Android app on Windows though a number of third party solutions, though there would still be significant hoops for Microsoft to jump through to make Android apps on Windows a mass market consumer reality:
It's not that it would be technically impossible. BlueStacks already offers users a way to run Android apps on Windows. In fact, it was a year ago today that BlueStacks released its Surface Pro-optimized Windows 8 version of its App Player software that allows users to download and run any Android App on their Windows devices. SweetLabs also has offered Windows users a way to load Android and other non-native Windows apps on their Windows machines.
It's clear that there would be significant obstacles to overcome from both a technical and business point of view. Android app compatibility would not guarantee that all Android apps would automatically become available on the Windows platform since it would be necessary to create a distribution channel (i.e. an app store, since the Google Play Store could not be co-opted since access to it is controlled by Google).
The experience of both Amazon (Kindle Fire) and BlackBerry (Android on BlackBerry OS 10) suggests this collection of apps for distribution / setting up of an app store is the biggest stumbling block to making an Android app compatibility feature truly useful for consumers. Android app compatibility would lose much of its value if apps did not arrive in a Microsoft platform Android app store within a few days of arriving in the main Google Play app store. There are also plenty of technical questions and limitations around how the service layer API compatibility could be maintained (i.e. those Google APIs and services that are not part of AOSP), something that would absolutely be necessary for many of the higher profile apps (perhaps less so for games though).
There is, of course, also a political and strategic dimension to do this. It's hard to see how Microsoft's adoption of Android app compatibility could do anything but annoying its own developer base and the motivation to create Windows Store on the WinRuntime APIs would surely be negatively impacted? Or put more simply, why bother writing apps for Windows? The reality would likely be more complex than this, but either way it's a debate that is likely to draw strong emotions on either side.
Ultimately, in our view, the difficulties in replicating the content of the Google Play Store, plus the likely poor experience (in visual design and UX terms if nothing else), and the wider impact of Microsoft platforms make it unlikely that this will happen. However, we would not rule out some of kind of ease of porting initiative underpinned by some kind of virtualisation technology and/or a Microsoft sponsored improvement in the current third party run-Android-apps-on-Windows solutions (i.e. so some kind of option is there for those that need it).
Regardless, the topic raises some interesting questions about what Microsoft app strategy should be? Is the current course sustainable, with just patience and device volume needed, or will there need to be a strategic shift?