Harmattan was supposed to ship in fall of 2010. It's a year late.
Yet, even if the MeeGo mess hadn't happened and Nokia had shipped a Harmattan phone on the original Maemo schedule, would it have made any difference? It would have been hamstrung by Nokia's confusion over Symbian's direction, and steamrolled by iOS and Android just the same.
It's not Elop's fault that Nokia's software is in shambles. The damage was done years earlier.
Put yourself in Elop's shoes when he arrived a year ago. There's Symbian, an OS that has become extremely expensive to develop. It has been open-sourced and modernized at great effort, but the resulting product (Symbian^3) has critical bugs and a lackluster UI. The people in charge tell him that the big leap forward is just around the corner... But apparently that's exactly what they were saying two years earlier.
Then there's MeeGo, a Linux variant that used to be one VP's hobby project. Recently it has been promoted to become the company's new flagship -- except that nobody is quite clear on how it should be positioned alongside Symbian. There's a powerful Symbian lobby inside the company that's determined to see the "eternal upstart" MeeGo fail. The technical state of the OS doesn't inspire much confidence either: the UI framework has been reset several times in the past few years, so although they have a sprawling Linux distribution, there isn't really a phone UI yet. "Don't worry," say the people in charge, "this latest framework is so great that we'll whip up a smartphone experience in no time!"
Faced with these options and all their internal political baggage, is it any wonder that Elop went looking for an operating system outside the company?