I am really surprised this post has caused used such a strong backlash. It honestly leads me to wonder about the long-term stability of this project, specially when an admin seems to be so closed minded.
As an educator, I see a lot of value with Android on Pi. I see it as a low-priced test device. That means everyone in the classroom can have one and even take it home---just like they do with their graphing calculators.
Just to clarify---as someone pointed out, no one programs *on* Android. You do that on a suitable OS (i.e., that can run Eclipse and the Android SDK). Android and ios aren't designed for that.
However, as an Android developer, I can attest that creating apps *outside* of Android is relatively easy, certainly no harder than ios. It's just Java after all.
So what I envision is developing on a regular desktop computer---perhaps even on another Pi running a suitable OS. Then testing your creations on Android Pi.
For $25-$35, I don't see how schools can go wrong...specially when mobile is the future.
OK, as a closed minded admin, who is quite aware of how to make Android apps on a desktop, let me explain in simple terms.
Ice Cream Sandwich won't run on the Raspi.
Earlier versions of Android may run on the Raspi. (700Mhz Arm, 256MB RAM remember)
The Foundation does not plan to port any Android to the Raspi. It's many man months of work, man months that are not available, and Broadcom won't be doing it for the SoC in question.
If someone wants to port Android to the Raspi they are more than welcome, and if we can help we will.
Android is not suitable for writing programs on. Raw Linux is. The Raspi is designed to enable people who cannot afford a desktop to learn to program. If you already have a PC capable of running an Android App dev system, you don't need a Raspi for its primary purpose, so what is the point of Android on a Raspi?
Anything you can program on Android could also be programmed in Linux.
Writing Android apps means Java (or something close!) - no C, no C++, no Python, no Perl, no Basic, no Fortran, no Cobol etc plus...no hardware access (GPIO's etc).
What benefits to education does Android bring over Linux, given the above?
It is simply not with the Foundation's time to do an Android port themselves. It would take till this time next year, and I happen to know that some of the acceleration available on the GPU is an UTTER PITA to implement in Android some of the older versions - it's bad enough in the latest one.