I installed Android 1.6 on a Google Ion developer phone last week and found it to be a minor update on that hardware, from a user's perspective. But developers will still have to target version 1.6 specifically if they want to support the new features it makes possible on other handset models, such as higher screen resolutions and CDMA support.
Android 2.0, on the other hand, does introduce significant new user features, even on existing phones. More important, it promises to address lingering weaknesses that I hoped would be corrected in Android 1.6, such as the clunky onscreen keyboard.
But the upgrade path for older Android handsets is unclear. Rumor has it that T-Mobile's G1 hardware, in particular, may already be nearing a dead end, leaving early adopters in a lurch just one year after the carrier began shipping the first Android phones.
How much choice is too much?
All of this leaves Android developers in a pretty pickle. On the one hand, they're working with what is arguably the most open mobile OS around. On the other, that high degree of openness means developers can never be sure which version of the OS to target if they want their apps to reach the broadest possible audience.
As I write this, handset manufacturers are reportedly readying some 18 different Android phones. Each will find a home at a different carrier, and each will try to set itself apart in some way, just as Verizon is trying to do by being first to market with Android 2.0. Proprietary add-ons such as HTC's Sense UI will mean one Android phone won't even have the same look and feel as the next.
But more important than how developers will deal with this issue, I suspect, will be how customers react to it. In the mobile world, carriers call the shots. Tech-savvy customers might be able to install Cyanogen, but most consumers will be stuck with the versions of Android that ship with their phones, plus whatever updates their carriers push over the air.
So is this to be the public face of Android? Openness, but lack of choice? An industry-standard OS, but no consistency? A product line in which handsets become obsolete every few months and the next big feature is always right around the corner?
If so, I fear Google's lauded open source OS could become the market equivalent of that old Jedi mind trick: "These aren't the droids you're looking for. Move along."