Not being a owner, not having a pre-order, and having every intent never to buy one, i can say Android for the Panda, would sure be a asset. If only implemented as a virtual machine, fine, if set up as a multi-boot device with mini menu, full Ångström, and Android options, now thats the real thing. But as always. Space. Namely storage space. And you really need to consider that the current buyers actually want the Linux bit. Be it for the total geek factor of having a very small "laptop" running Linux, or any other purpose, but making it an android device... Well, thats just not geek enough.
Most importantly tho. You cant go with the multi-boot idea, and make one of the boot concepts half assed. What ever options gets picked has to be supported 100%, and made to work perfectly every step of the way.
Personally i think Android is a fuckt up idea. I'm seeing this from a performance, and enthusiast perspective. You'd sacrifice optimization in return for a large software library. On a device expected to run emulators that it just barely will be able to, to start with, forget making it happen with a Android OS as the foundation. No way. But, I'm not making and trying to sell the device. In the search to make the money back, and preferably make a small profit, (heck, a big one if possible) you need lots of buyers. Focus on making the current OS work properly, and fix the SDK so that everything is 100%. Support developers, and they support you. Add an option, Android or Linux, later. Sell pre-configured units, Linux & Android versions. Both identical, 1 flash away from either version, and the buyer has all the options at his or her fingertips. That should take care of the nand space issue. Of course. Make bloody sure the Android version is up to speed b4 adding it as a pre-configured option. Or buyers will bitch about it not working, while *brand X* phones does.
Just my hot air.
B!