I'd argue they haven't. The success would've been the adoption of the platform - but as far as I can see none of the major Android phone vendors actually support the platform - they see it more as a way to skimp on software development costs, and offload it onto Google instead. There's little to no interest in even updating the phones, or supporting the platform as a whole.
I feel that Google has been taken for a ride by Motorola, HTC, et al - there's a lot of lip service paid to Android, but none of their actions have helped solidify the platform, instead being treated simply as free code. The fact that there are a lot of units of hardware moving off shelves means little, IMHO, for the long-term success of the platform.
That's fair play tho' - Google sought to commoditize handsets.
This is true for all mobile platforms. People change their phones often and have few true dependencies on the underlying OS. RIM might be the "dominant smartphone" at the moment but I doubt they're feeling super comfortable.
On the other hand Ruboto[1] looks very promising, it's a project that brings JRuby to Android application development.
Apple have done a great job of outdating the old sdks, mainly through the method of dropping simulator support for the previous sdks and no legitimate route of running previous firmwares on an iPhone other than holding them back (obvious flaw in this plan, how does a new developer acquire an ipod running 2.2 or 3.0)
Android on the other hand has had no where near this rate of attrition of the previous sdks with only really the initial release not being heavily supported.
However Android suffers from the problem you are not just waiting for Google to update the firmware so you can't make your app 2.0+ only in the thoughts that if users want my app they can update their firmware, this step would remove a lot of htc, sony and other phones that do not currently have the ability to run a 2.0 firmware whilst still being a very comptent phone.
As for the carrier "enhanced" firmwares....
Remember those graphs of user's current phones related to the next phone they bought? Apple users were happy to continue with Apple, and even BB users wanted iphones.
Google has had great success, which is good to see.
"Which just means they can do even better if they can address those obvious problems."
Not necessarily. If all these people are just buying android phones because they can not get an iPhone, then nothing Google does will change that. The real test will be when the iPhone is available on more carriers, and when people need to upgrade there Android devices, and stay with Android. It is just too soon to tell, so far.
All true, and still they've had great success. (I'm actually surprised, given how bad the Market is and how blatantly the carriers are screwing up the phones). Which just means they can do even better if they can address those obvious problems.
On the other hand, even though I generally like Java I'm finding the move from iPhone development to Android development to be a depressing step backwards
It was the opposite for me. Between getting rid of header files and manual memory management, and having the app run right away on my phone rather than futzing with certificates, I'm enjoying Android development much more.