> I have had catastrophic failures caused by premature optimization, because locking a tool into a fancy algorithm before you know where the actual bottleneck is is a recipe for disaster.

EDIT: Or, are you saying that it was from the decisions made before Google acquired Android?