My first thought as well. Battery life on my Incredible is poor at best as it is. I don't want to pony up for an oversize battery and wreck the profile, I need it to fit in a case/holster/mount/etc.
How fast does battery technology progress? I've replaced a few batteries in iPods and Nintendo DS's and I'm fairly certain they were rated much higher than the stock ones, while being surprisingly cheap.
Does it go fast enough to make a material difference between choosing a phone with or without user accessible batteries and purchasing a newer battery 6 or 12 months later?
Historically, battery tech progress has been depressingly slow. Slow enough to be the primary limiting factor for a lot of technologies. The vast majority of progress has been in reducing power consumption rather than increasing power storage.
Interesting, I've never thought about replacing a phone battery with a higher-end model (can you tell I have an iPhone?) I'm curious as to the actual battery life gains one would see and the effect on weight , etc.
From what I've seen extended life batteries for phones tend to be 100-200mAh larger than stock batteries. So I think that would work out to a ~10% improvement for most phones. It really depends on the capacity of the OEM battery. I remember seeing a 2000mAh+ battery for the EVO which would be nearly a 25% improvement over the stock battery.
Just to provide a counterpoint: N900/Maemo has had the ability to have apps distributed in an arbitrary language since launch, and Nokia seems to especially like Python, sponsoring the developement of Qt bindings for it etc.
Do you have some reason to believe that Google seriously hosed the Java implementation on Android or these guys have massively increased the efficiency of the Python implementation? Otherwise, it seems the answer would have to be yes.. the same algorithm will usually perform (much) worse in Python than in Java.
Note I don't mean to slam Python here.. just being practical. Python's my primary language. Development efficiency usually trumps execution inefficiency.
Python is going to chug even on relatively simple action games. I did write some 2D stuff in Pygame/SDL and Pygame/OpenGL back in the early-mid 2000s that could do 30FPS, but to do it was akin to writing in BASIC on an 8-bit; you had to micro-optimize in a fairly brutal way, one which obscured Python's powers. Things like rendering a tilemap could get huge speed boosts simply by avoiding nested loops in favor of a cache. Even using OpenGL wasn't really enough to secure a comfortable margin of performance.
If I did it again I would have an easier time because I missed a few critical constructs back then(most notably, the array package). It still wouldn't be pretty, especially not on a phone.
